1
0
mirror of https://github.com/fluencelabs/wasm-bindgen synced 2025-04-08 21:28:07 +00:00

3385 lines
123 KiB
Rust
Raw Normal View History

First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
use crate::descriptor::VectorKind;
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
use crate::intrinsic::Intrinsic;
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
use crate::wit::{Adapter, AdapterId, AdapterJsImportKind, AuxValue};
use crate::wit::{AdapterKind, Instruction, InstructionData};
use crate::wit::{AuxEnum, AuxExport, AuxExportKind, AuxImport, AuxStruct};
use crate::wit::{JsImport, JsImportName, NonstandardWitSection, WasmBindgenAux};
Dramatically improving the build time of web-sys (#2012) * Pre-generating web-sys * Fixing build errors * Minor refactor for the unit tests * Changing to generate #[wasm_bindgen} annotations * Fixing code generation * Adding in main bin to wasm-bindgen-webidl * Fixing more problems * Adding in support for unstable APIs * Fixing bug with code generation * More code generation fixes * Improving the webidl program * Removing unnecessary cfg from the generated code * Splitting doc comments onto separate lines * Improving the generation for unstable features * Adding in support for string values in enums * Now runs rustfmt on the mod.rs file * Fixing codegen for constructors * Fixing webidl-tests * Fixing build errors * Another fix for build errors * Renaming typescript_name to typescript_type * Adding in docs for typescript_type * Adding in CI script to verify that web-sys is up to date * Fixing CI script * Fixing CI script * Don't suppress git diff output * Remove duplicate definitions of `Location` Looks to be a preexisting bug in wasm-bindgen? * Regenerate webidl * Try to get the git diff command right * Handle named constructors in WebIDL * Remove stray rustfmt.toml * Add back NamedConstructorBar definition in tests * Run stable rustfmt over everything * Don't run Cargo in a build script Instead refactor things so webidl-tests can use the Rust-code-generation as a library in a build script. Also fixes `cargo fmt` in the repository. * Fixup generated code * Running web-sys checks on stable * Improving the code generation a little * Running rustfmt Co-authored-by: Alex Crichton <alex@alexcrichton.com>
2020-03-03 00:39:36 +01:00
use crate::{reset_indentation, Bindgen, EncodeInto, OutputMode, PLACEHOLDER_MODULE};
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
use anyhow::{anyhow, bail, Context as _, Error};
use std::borrow::Cow;
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
use std::collections::{BTreeMap, BTreeSet, HashMap, HashSet};
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
use std::fmt;
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
use std::fs;
use std::path::{Path, PathBuf};
use walrus::{FunctionId, ImportId, MemoryId, Module, TableId};
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
mod binding;
pub struct Context<'a> {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
globals: String,
imports_post: String,
typescript: String,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
exposed_globals: Option<HashSet<Cow<'static, str>>>,
next_export_idx: usize,
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
config: &'a Bindgen,
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
pub module: &'a mut Module,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
aux: &'a WasmBindgenAux,
wit: &'a NonstandardWitSection,
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
/// A map representing the `import` statements we'll be generating in the JS
/// glue. The key is the module we're importing from and the value is the
/// list of identifier we're importing from the module, with optional
/// renames for each identifier.
js_imports: HashMap<String, Vec<(String, Option<String>)>>,
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
/// A map of each wasm import and what JS to hook up to it.
wasm_import_definitions: HashMap<ImportId, String>,
/// A map from an import to the name we've locally imported it as.
imported_names: HashMap<JsImportName, String>,
/// A set of all defined identifiers through either exports or imports to
/// the number of times they've been used, used to generate new
/// identifiers.
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
defined_identifiers: HashMap<String, usize>,
exported_classes: Option<BTreeMap<String, ExportedClass>>,
/// A map of the name of npm dependencies we've loaded so far to the path
/// they're defined in as well as their version specification.
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
pub npm_dependencies: HashMap<String, (PathBuf, String)>,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
/// A mapping of a index for memories as we see them. Used in function
/// names.
memory_indices: HashMap<MemoryId, usize>,
table_indices: HashMap<TableId, usize>,
}
#[derive(Default)]
pub struct ExportedClass {
comments: String,
contents: String,
typescript: String,
has_constructor: bool,
wrap_needed: bool,
/// Whether to generate helper methods for inspecting the class
is_inspectable: bool,
/// All readable properties of the class
readable_properties: Vec<String>,
/// Map from field name to type as a string, docs plus whether it has a setter
/// and it is optional
typescript_fields: HashMap<String, (String, String, bool, bool)>,
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
const INITIAL_HEAP_VALUES: &[&str] = &["undefined", "null", "true", "false"];
// Must be kept in sync with `src/lib.rs` of the `wasm-bindgen` crate
const INITIAL_HEAP_OFFSET: usize = 32;
Add support for optional slice types (#507) * Shard the `convert.rs` module into sub-modules Hopefully this'll make the organization a little nicer over time! * Start adding support for optional types This commit starts adding support for optional types to wasm-bindgen as arguments/return values to functions. The strategy here is to add two new traits, `OptionIntoWasmAbi` and `OptionFromWasmAbi`. These two traits are used as a blanket impl to implement `IntoWasmAbi` and `FromWasmAbi` for `Option<T>`. Some consequences of this design: * It should be possible to ensure `Option<SomeForeignType>` implements to/from wasm traits. This is because the option-based traits can be implemented for foreign types. * A specialized implementation is possible for all types, so there's no need for `Option<T>` to introduce unnecessary overhead. * Two new traits is a bit unforutnate but I can't currently think of an alternative design that works for the above two constraints, although it doesn't mean one doesn't exist! * The error messages for "can't use this type here" is actually halfway decent because it says these new traits need to be implemented, which provides a good place to document and talk about what's going on here! * Nested references like `Option<&T>` can't implement `FromWasmAbi`. This means that you can't define a function in Rust which takes `Option<&str>`. It may be possible to do this one day but it'll likely require more trait trickery than I'm capable of right now. * Add support for optional slices This commit adds support for optional slice types, things like strings and arrays. The null representation of these has a pointer value of 0, which should never happen in normal Rust. Otherwise the various plumbing is done throughout the tooling to enable these types in all locations. * Fix `takeObject` on global sentinels These don't have a reference count as they're always expected to work, so avoid actually dropping a reference on them. * Remove some no longer needed bindings * Add support for optional anyref types This commit adds support for optional imported class types. Each type imported with `#[wasm_bindgen]` automatically implements the relevant traits and now supports `Option<Foo>` in various argument/return positions. * Fix building without the `std` feature * Actually fix the build... * Add support for optional types to WebIDL Closes #502
2018-07-19 14:44:23 -05:00
impl<'a> Context<'a> {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
pub fn new(
module: &'a mut Module,
config: &'a Bindgen,
wit: &'a NonstandardWitSection,
aux: &'a WasmBindgenAux,
) -> Result<Context<'a>, Error> {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
Ok(Context {
globals: String::new(),
imports_post: String::new(),
typescript: "/* tslint:disable */\n/* eslint-disable */\n".to_string(),
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
exposed_globals: Some(Default::default()),
imported_names: Default::default(),
js_imports: Default::default(),
defined_identifiers: Default::default(),
wasm_import_definitions: Default::default(),
exported_classes: Some(Default::default()),
config,
module,
npm_dependencies: Default::default(),
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
next_export_idx: 0,
wit,
aux,
memory_indices: Default::default(),
table_indices: Default::default(),
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
})
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn should_write_global(&mut self, name: impl Into<Cow<'static, str>>) -> bool {
self.exposed_globals.as_mut().unwrap().insert(name.into())
}
fn export(
&mut self,
export_name: &str,
contents: &str,
comments: Option<&str>,
) -> Result<(), Error> {
let definition_name = self.generate_identifier(export_name);
if contents.starts_with("class") && definition_name != export_name {
bail!("cannot shadow already defined class `{}`", export_name);
}
let contents = contents.trim();
if let Some(c) = comments {
self.globals.push_str(c);
}
let global = match self.config.mode {
OutputMode::Node {
experimental_modules: false,
} => {
if contents.starts_with("class") {
format!("{}\nmodule.exports.{1} = {1};\n", contents, export_name)
} else {
format!("module.exports.{} = {};\n", export_name, contents)
}
}
OutputMode::NoModules { .. } => {
if contents.starts_with("class") {
format!("{}\n__exports.{1} = {1};\n", contents, export_name)
} else {
format!("__exports.{} = {};\n", export_name, contents)
}
}
OutputMode::Bundler { .. }
| OutputMode::Node {
experimental_modules: true,
}
| OutputMode::Web => {
if contents.starts_with("function") {
let body = &contents[8..];
if export_name == definition_name {
format!("export function {}{}\n", export_name, body)
} else {
format!(
"function {}{}\nexport {{ {} as {} }};\n",
definition_name, body, definition_name, export_name,
)
}
} else if contents.starts_with("class") {
assert_eq!(export_name, definition_name);
format!("export {}\n", contents)
} else {
assert_eq!(export_name, definition_name);
format!("export const {} = {};\n", export_name, contents)
}
}
};
2018-04-14 09:37:49 -07:00
self.global(&global);
Ok(())
}
pub fn finalize(
&mut self,
module_name: &str,
) -> Result<(String, String, Option<String>), Error> {
// Finalize all bindings for JS classes. This is where we'll generate JS
// glue for all classes as well as finish up a few final imports like
// `__wrap` and such.
self.write_classes()?;
2019-06-04 09:18:48 -07:00
// Initialization is just flat out tricky and not something we
// understand super well. To try to handle various issues that have come
// up we always remove the `start` function if one is present. The JS
// bindings glue then manually calls the start function (if it was
// previously present).
let needs_manual_start = self.unstart_start_function();
// Cause any future calls to `should_write_global` to panic, making sure
// we don't ask for items which we can no longer emit.
drop(self.exposed_globals.take().unwrap());
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.finalize_js(module_name, needs_manual_start)
}
Dramatically improving the build time of web-sys (#2012) * Pre-generating web-sys * Fixing build errors * Minor refactor for the unit tests * Changing to generate #[wasm_bindgen} annotations * Fixing code generation * Adding in main bin to wasm-bindgen-webidl * Fixing more problems * Adding in support for unstable APIs * Fixing bug with code generation * More code generation fixes * Improving the webidl program * Removing unnecessary cfg from the generated code * Splitting doc comments onto separate lines * Improving the generation for unstable features * Adding in support for string values in enums * Now runs rustfmt on the mod.rs file * Fixing codegen for constructors * Fixing webidl-tests * Fixing build errors * Another fix for build errors * Renaming typescript_name to typescript_type * Adding in docs for typescript_type * Adding in CI script to verify that web-sys is up to date * Fixing CI script * Fixing CI script * Don't suppress git diff output * Remove duplicate definitions of `Location` Looks to be a preexisting bug in wasm-bindgen? * Regenerate webidl * Try to get the git diff command right * Handle named constructors in WebIDL * Remove stray rustfmt.toml * Add back NamedConstructorBar definition in tests * Run stable rustfmt over everything * Don't run Cargo in a build script Instead refactor things so webidl-tests can use the Rust-code-generation as a library in a build script. Also fixes `cargo fmt` in the repository. * Fixup generated code * Running web-sys checks on stable * Improving the code generation a little * Running rustfmt Co-authored-by: Alex Crichton <alex@alexcrichton.com>
2020-03-03 00:39:36 +01:00
fn generate_node_imports(&self) -> String {
let mut imports = BTreeSet::new();
for import in self.module.imports.iter() {
imports.insert(&import.module);
}
let mut shim = String::new();
shim.push_str("let imports = {};\n");
if self.config.mode.nodejs_experimental_modules() {
for (i, module) in imports.iter().enumerate() {
if module.as_str() != PLACEHOLDER_MODULE {
shim.push_str(&format!("import * as import{} from '{}';\n", i, module));
}
}
}
for (i, module) in imports.iter().enumerate() {
if module.as_str() == PLACEHOLDER_MODULE {
shim.push_str(&format!(
"imports['{0}'] = module.exports;\n",
PLACEHOLDER_MODULE
));
} else {
if self.config.mode.nodejs_experimental_modules() {
shim.push_str(&format!("imports['{}'] = import{};\n", module, i));
} else {
shim.push_str(&format!("imports['{0}'] = require('{0}');\n", module));
}
}
}
reset_indentation(&shim)
}
fn generate_node_wasm_loading(&self, path: &Path) -> String {
let mut shim = String::new();
if self.config.mode.nodejs_experimental_modules() {
// On windows skip the leading `/` which comes out when we parse a
// url to use `C:\...` instead of `\C:\...`
shim.push_str(&format!(
"
import * as path from 'path';
import * as fs from 'fs';
import * as url from 'url';
import * as process from 'process';
let file = path.dirname(url.parse(import.meta.url).pathname);
if (process.platform === 'win32') {{
file = file.substring(1);
}}
const bytes = fs.readFileSync(path.join(file, '{}'));
",
path.file_name().unwrap().to_str().unwrap()
));
} else {
shim.push_str(&format!(
"
const path = require('path').join(__dirname, '{}');
const bytes = require('fs').readFileSync(path);
",
path.file_name().unwrap().to_str().unwrap()
));
}
shim.push_str(
"
const wasmModule = new WebAssembly.Module(bytes);
const wasmInstance = new WebAssembly.Instance(wasmModule, imports);
wasm = wasmInstance.exports;
module.exports.__wasm = wasm;
",
);
reset_indentation(&shim)
}
/// Performs the task of actually generating the final JS module, be it
/// `--target no-modules`, `--target web`, or for bundlers. This is the very
/// last step performed in `finalize`.
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn finalize_js(
&mut self,
module_name: &str,
needs_manual_start: bool,
) -> Result<(String, String, Option<String>), Error> {
2019-04-01 19:45:53 -03:00
let mut ts = self.typescript.clone();
let mut js = String::new();
let mut start = None;
if let OutputMode::NoModules { global } = &self.config.mode {
js.push_str(&format!("let {};\n(function() {{\n", global));
}
// Depending on the output mode, generate necessary glue to actually
// import the wasm file in one way or another.
2019-04-01 19:45:53 -03:00
let mut init = (String::new(), String::new());
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
let mut footer = String::new();
let mut imports = self.js_import_header()?;
match &self.config.mode {
// In `--target no-modules` mode we need to both expose a name on
// the global object as well as generate our own custom start
// function.
OutputMode::NoModules { global } => {
js.push_str("const __exports = {};\n");
js.push_str("let wasm;\n");
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
init = self.gen_init(needs_manual_start, None)?;
footer.push_str(&format!("{} = Object.assign(init, __exports);\n", global));
}
// With normal CommonJS node we need to defer requiring the wasm
// until the end so most of our own exports are hooked up
OutputMode::Node {
experimental_modules: false,
} => {
Dramatically improving the build time of web-sys (#2012) * Pre-generating web-sys * Fixing build errors * Minor refactor for the unit tests * Changing to generate #[wasm_bindgen} annotations * Fixing code generation * Adding in main bin to wasm-bindgen-webidl * Fixing more problems * Adding in support for unstable APIs * Fixing bug with code generation * More code generation fixes * Improving the webidl program * Removing unnecessary cfg from the generated code * Splitting doc comments onto separate lines * Improving the generation for unstable features * Adding in support for string values in enums * Now runs rustfmt on the mod.rs file * Fixing codegen for constructors * Fixing webidl-tests * Fixing build errors * Another fix for build errors * Renaming typescript_name to typescript_type * Adding in docs for typescript_type * Adding in CI script to verify that web-sys is up to date * Fixing CI script * Fixing CI script * Don't suppress git diff output * Remove duplicate definitions of `Location` Looks to be a preexisting bug in wasm-bindgen? * Regenerate webidl * Try to get the git diff command right * Handle named constructors in WebIDL * Remove stray rustfmt.toml * Add back NamedConstructorBar definition in tests * Run stable rustfmt over everything * Don't run Cargo in a build script Instead refactor things so webidl-tests can use the Rust-code-generation as a library in a build script. Also fixes `cargo fmt` in the repository. * Fixup generated code * Running web-sys checks on stable * Improving the code generation a little * Running rustfmt Co-authored-by: Alex Crichton <alex@alexcrichton.com>
2020-03-03 00:39:36 +01:00
js.push_str(&self.generate_node_imports());
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
js.push_str("let wasm;\n");
Add tests for the interface types output of wasm-bindgen (#1898) * Add tests for the interface types output of wasm-bindgen This commit expands the test suite with assertions about the output of the interface types pass in wasm-bindgen. The goal here is to actually assert that we produce the right output and have a suite of reference files to show how the interface types output is changing over time. The `reference` test suite added in the previous PR has been updated to work for interface types as well, generating `*.wit` file assertions which are printed via the `wit-printer` crate on crates.io. Along the way a number of bugs were fixed with the interface types output, such as: * Non-determinism in output caused by iteration of a `HashMap` * Avoiding JS generation entirely in interface types mode, ensuring that we don't export extraneous intrinsics that aren't otherwise needed. * Fixing location of the stack pointer for modules where it's GC'd out. It's now rooted in the aux section of wasm-bindgen so it's available to later passes, like the multi-value pass. * Interface types emission now works in debug mode, meaning the `--release` flag is no longer required. This previously did not work because the `__wbindgen_throw` intrinsic was required in debug mode. This comes about because of the `malloc_failure` and `internal_error` functions in the anyref pass. The purpose of these functions is to signal fatal runtime errors, if any, in a way that's usable to the user. For wasm interface types though we can replace calls to these functions with `unreachable` to avoid needing to import the intrinsic. This has the accidental side effect of making `wasm_bindgen::throw_str` "just work" with wasm interface types by aborting the program, but that's not actually entirely intended. It's hoped that a split of a `wasm-bindgen-core` crate would solve this issue for the future. * Run the wasm interface types validator in tests * Add more gc roots for adapter gc * Improve stack pointer detection The stack pointer is never initialized to zero, but some other mutable globals are (TLS, thread ID, etc), so let's filter those out.
2019-12-04 15:19:48 -06:00
for (id, js) in crate::sorted_iter(&self.wasm_import_definitions) {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
let import = self.module.imports.get_mut(*id);
footer.push_str("\nmodule.exports.");
footer.push_str(&import.name);
footer.push_str(" = ");
footer.push_str(js.trim());
footer.push_str(";\n");
}
Dramatically improving the build time of web-sys (#2012) * Pre-generating web-sys * Fixing build errors * Minor refactor for the unit tests * Changing to generate #[wasm_bindgen} annotations * Fixing code generation * Adding in main bin to wasm-bindgen-webidl * Fixing more problems * Adding in support for unstable APIs * Fixing bug with code generation * More code generation fixes * Improving the webidl program * Removing unnecessary cfg from the generated code * Splitting doc comments onto separate lines * Improving the generation for unstable features * Adding in support for string values in enums * Now runs rustfmt on the mod.rs file * Fixing codegen for constructors * Fixing webidl-tests * Fixing build errors * Another fix for build errors * Renaming typescript_name to typescript_type * Adding in docs for typescript_type * Adding in CI script to verify that web-sys is up to date * Fixing CI script * Fixing CI script * Don't suppress git diff output * Remove duplicate definitions of `Location` Looks to be a preexisting bug in wasm-bindgen? * Regenerate webidl * Try to get the git diff command right * Handle named constructors in WebIDL * Remove stray rustfmt.toml * Add back NamedConstructorBar definition in tests * Run stable rustfmt over everything * Don't run Cargo in a build script Instead refactor things so webidl-tests can use the Rust-code-generation as a library in a build script. Also fixes `cargo fmt` in the repository. * Fixup generated code * Running web-sys checks on stable * Improving the code generation a little * Running rustfmt Co-authored-by: Alex Crichton <alex@alexcrichton.com>
2020-03-03 00:39:36 +01:00
footer.push_str(
&self.generate_node_wasm_loading(&Path::new(&format!(
"./{}_bg.wasm",
module_name
))),
);
if needs_manual_start {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
footer.push_str("wasm.__wbindgen_start();\n");
}
}
// With Bundlers and modern ES6 support in Node we can simply import
// the wasm file as if it were an ES module and let the
// bundler/runtime take care of it.
OutputMode::Bundler { .. }
| OutputMode::Node {
experimental_modules: true,
} => {
2019-07-10 14:45:33 -07:00
imports.push_str(&format!(
"import * as wasm from './{}_bg.wasm';\n",
module_name
));
Add tests for the interface types output of wasm-bindgen (#1898) * Add tests for the interface types output of wasm-bindgen This commit expands the test suite with assertions about the output of the interface types pass in wasm-bindgen. The goal here is to actually assert that we produce the right output and have a suite of reference files to show how the interface types output is changing over time. The `reference` test suite added in the previous PR has been updated to work for interface types as well, generating `*.wit` file assertions which are printed via the `wit-printer` crate on crates.io. Along the way a number of bugs were fixed with the interface types output, such as: * Non-determinism in output caused by iteration of a `HashMap` * Avoiding JS generation entirely in interface types mode, ensuring that we don't export extraneous intrinsics that aren't otherwise needed. * Fixing location of the stack pointer for modules where it's GC'd out. It's now rooted in the aux section of wasm-bindgen so it's available to later passes, like the multi-value pass. * Interface types emission now works in debug mode, meaning the `--release` flag is no longer required. This previously did not work because the `__wbindgen_throw` intrinsic was required in debug mode. This comes about because of the `malloc_failure` and `internal_error` functions in the anyref pass. The purpose of these functions is to signal fatal runtime errors, if any, in a way that's usable to the user. For wasm interface types though we can replace calls to these functions with `unreachable` to avoid needing to import the intrinsic. This has the accidental side effect of making `wasm_bindgen::throw_str` "just work" with wasm interface types by aborting the program, but that's not actually entirely intended. It's hoped that a split of a `wasm-bindgen-core` crate would solve this issue for the future. * Run the wasm interface types validator in tests * Add more gc roots for adapter gc * Improve stack pointer detection The stack pointer is never initialized to zero, but some other mutable globals are (TLS, thread ID, etc), so let's filter those out.
2019-12-04 15:19:48 -06:00
for (id, js) in crate::sorted_iter(&self.wasm_import_definitions) {
let import = self.module.imports.get_mut(*id);
import.module = format!("./{}_bg.js", module_name);
footer.push_str("\nexport const ");
footer.push_str(&import.name);
footer.push_str(" = ");
footer.push_str(js.trim());
footer.push_str(";\n");
}
if needs_manual_start {
start = Some("\nwasm.__wbindgen_start();\n".to_string());
}
}
// With a browser-native output we're generating an ES module, but
// browsers don't support natively importing wasm right now so we
// expose the same initialization function as `--target no-modules`
// as the default export of the module.
OutputMode::Web => {
self.imports_post.push_str("let wasm;\n");
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
init = self.gen_init(needs_manual_start, Some(&mut imports))?;
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
footer.push_str("export default init;\n");
}
}
2019-04-01 19:45:53 -03:00
let (init_js, init_ts) = init;
ts.push_str(&init_ts);
// Emit all the JS for importing all our functionality
assert!(
!self.config.mode.uses_es_modules() || js.is_empty(),
"ES modules require imports to be at the start of the file"
);
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
js.push_str(&imports);
js.push_str("\n");
js.push_str(&self.imports_post);
js.push_str("\n");
// Emit all our exports from this module
js.push_str(&self.globals);
js.push_str("\n");
// Generate the initialization glue, if there was any
2019-04-01 19:45:53 -03:00
js.push_str(&init_js);
js.push_str("\n");
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
js.push_str(&footer);
js.push_str("\n");
if self.config.mode.no_modules() {
js.push_str("})();\n");
}
while js.contains("\n\n\n") {
js = js.replace("\n\n\n", "\n\n");
}
Ok((js, ts, start))
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn js_import_header(&self) -> Result<String, Error> {
let mut imports = String::new();
if self.config.omit_imports {
return Ok(imports);
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
match &self.config.mode {
OutputMode::NoModules { .. } => {
for (module, _items) in self.js_imports.iter() {
bail!(
"importing from `{}` isn't supported with `--target no-modules`",
module
);
}
}
OutputMode::Node {
experimental_modules: false,
} => {
Add tests for the interface types output of wasm-bindgen (#1898) * Add tests for the interface types output of wasm-bindgen This commit expands the test suite with assertions about the output of the interface types pass in wasm-bindgen. The goal here is to actually assert that we produce the right output and have a suite of reference files to show how the interface types output is changing over time. The `reference` test suite added in the previous PR has been updated to work for interface types as well, generating `*.wit` file assertions which are printed via the `wit-printer` crate on crates.io. Along the way a number of bugs were fixed with the interface types output, such as: * Non-determinism in output caused by iteration of a `HashMap` * Avoiding JS generation entirely in interface types mode, ensuring that we don't export extraneous intrinsics that aren't otherwise needed. * Fixing location of the stack pointer for modules where it's GC'd out. It's now rooted in the aux section of wasm-bindgen so it's available to later passes, like the multi-value pass. * Interface types emission now works in debug mode, meaning the `--release` flag is no longer required. This previously did not work because the `__wbindgen_throw` intrinsic was required in debug mode. This comes about because of the `malloc_failure` and `internal_error` functions in the anyref pass. The purpose of these functions is to signal fatal runtime errors, if any, in a way that's usable to the user. For wasm interface types though we can replace calls to these functions with `unreachable` to avoid needing to import the intrinsic. This has the accidental side effect of making `wasm_bindgen::throw_str` "just work" with wasm interface types by aborting the program, but that's not actually entirely intended. It's hoped that a split of a `wasm-bindgen-core` crate would solve this issue for the future. * Run the wasm interface types validator in tests * Add more gc roots for adapter gc * Improve stack pointer detection The stack pointer is never initialized to zero, but some other mutable globals are (TLS, thread ID, etc), so let's filter those out.
2019-12-04 15:19:48 -06:00
for (module, items) in crate::sorted_iter(&self.js_imports) {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
imports.push_str("const { ");
for (i, (item, rename)) in items.iter().enumerate() {
if i > 0 {
imports.push_str(", ");
2019-01-10 23:37:12 +00:00
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
imports.push_str(item);
if let Some(other) = rename {
imports.push_str(": ");
imports.push_str(other)
2019-01-10 23:37:12 +00:00
}
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
imports.push_str(" } = require(String.raw`");
imports.push_str(module);
imports.push_str("`);\n");
}
}
OutputMode::Bundler { .. }
| OutputMode::Node {
experimental_modules: true,
}
| OutputMode::Web => {
Add tests for the interface types output of wasm-bindgen (#1898) * Add tests for the interface types output of wasm-bindgen This commit expands the test suite with assertions about the output of the interface types pass in wasm-bindgen. The goal here is to actually assert that we produce the right output and have a suite of reference files to show how the interface types output is changing over time. The `reference` test suite added in the previous PR has been updated to work for interface types as well, generating `*.wit` file assertions which are printed via the `wit-printer` crate on crates.io. Along the way a number of bugs were fixed with the interface types output, such as: * Non-determinism in output caused by iteration of a `HashMap` * Avoiding JS generation entirely in interface types mode, ensuring that we don't export extraneous intrinsics that aren't otherwise needed. * Fixing location of the stack pointer for modules where it's GC'd out. It's now rooted in the aux section of wasm-bindgen so it's available to later passes, like the multi-value pass. * Interface types emission now works in debug mode, meaning the `--release` flag is no longer required. This previously did not work because the `__wbindgen_throw` intrinsic was required in debug mode. This comes about because of the `malloc_failure` and `internal_error` functions in the anyref pass. The purpose of these functions is to signal fatal runtime errors, if any, in a way that's usable to the user. For wasm interface types though we can replace calls to these functions with `unreachable` to avoid needing to import the intrinsic. This has the accidental side effect of making `wasm_bindgen::throw_str` "just work" with wasm interface types by aborting the program, but that's not actually entirely intended. It's hoped that a split of a `wasm-bindgen-core` crate would solve this issue for the future. * Run the wasm interface types validator in tests * Add more gc roots for adapter gc * Improve stack pointer detection The stack pointer is never initialized to zero, but some other mutable globals are (TLS, thread ID, etc), so let's filter those out.
2019-12-04 15:19:48 -06:00
for (module, items) in crate::sorted_iter(&self.js_imports) {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
imports.push_str("import { ");
for (i, (item, rename)) in items.iter().enumerate() {
if i > 0 {
imports.push_str(", ");
2019-01-16 10:46:26 +00:00
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
imports.push_str(item);
if let Some(other) = rename {
imports.push_str(" as ");
imports.push_str(other)
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
imports.push_str(" } from '");
imports.push_str(module);
imports.push_str("';\n");
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
}
Ok(imports)
}
2020-03-04 06:20:58 +01:00
fn ts_for_init_fn(
&self,
has_memory: bool,
has_module_or_path_optional: bool,
) -> Result<String, Error> {
let output = crate::wasm2es6js::interface(&self.module)?;
2019-04-01 19:45:53 -03:00
let (memory_doc, memory_param) = if has_memory {
(
"* @param {WebAssembly.Memory} maybe_memory\n",
", maybe_memory: WebAssembly.Memory",
)
} else {
("", "")
};
let arg_optional = if has_module_or_path_optional { "?" } else { "" };
Ok(format!(
2019-04-01 19:45:53 -03:00
"\n\
export type InitInput = RequestInfo | URL | Response | BufferSource | WebAssembly.Module;\n\
\n\
export interface InitOutput {{\n\
{output}}}\n\
\n\
2019-04-01 19:45:53 -03:00
/**\n\
* If `module_or_path` is {{RequestInfo}} or {{URL}}, makes a request and\n\
2019-04-01 19:45:53 -03:00
* for everything else, calls `WebAssembly.instantiate` directly.\n\
*\n\
* @param {{InitInput | Promise<InitInput>}} module_or_path\n\
2019-04-01 19:45:53 -03:00
{}\
*\n\
* @returns {{Promise<InitOutput>}}\n\
2019-04-01 19:45:53 -03:00
*/\n\
2019-05-09 11:01:45 -07:00
export default function init \
(module_or_path{}: InitInput | Promise<InitInput>{}): Promise<InitOutput>;
2019-04-01 19:45:53 -03:00
",
memory_doc, arg_optional, memory_param,
output = output,
))
2019-04-01 19:45:53 -03:00
}
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
fn gen_init(
&mut self,
needs_manual_start: bool,
mut imports: Option<&mut String>,
) -> Result<(String, String), Error> {
let module_name = "wbg";
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let mut init_memory_arg = "";
let mut init_memory1 = String::new();
let mut init_memory2 = String::new();
let mut has_memory = false;
if let Some(mem) = self.module.memories.iter().next() {
if let Some(id) = mem.import {
self.module.imports.get_mut(id).module = module_name.to_string();
let mut memory = String::from("new WebAssembly.Memory({");
memory.push_str(&format!("initial:{}", mem.initial));
if let Some(max) = mem.maximum {
memory.push_str(&format!(",maximum:{}", max));
}
if mem.shared {
memory.push_str(",shared:true");
}
memory.push_str("})");
self.imports_post.push_str("let memory;\n");
init_memory1 = format!("memory = imports.{}.memory = maybe_memory;", module_name);
init_memory2 = format!("memory = imports.{}.memory = {};", module_name, memory);
init_memory_arg = ", maybe_memory";
has_memory = true;
}
}
let default_module_path = match self.config.mode {
OutputMode::Web => {
"\
if (typeof input === 'undefined') {
input = import.meta.url.replace(/\\.js$/, '_bg.wasm');
}"
}
OutputMode::NoModules { .. } => {
"\
if (typeof input === 'undefined') {
let src;
2020-03-10 15:13:50 +01:00
if (typeof document === 'undefined') {
src = location.href;
} else {
src = document.currentScript.src;
}
input = src.replace(/\\.js$/, '_bg.wasm');
}"
}
_ => "",
};
let ts = self.ts_for_init_fn(has_memory, !default_module_path.is_empty())?;
// Initialize the `imports` object for all import definitions that we're
// directed to wire up.
let mut imports_init = String::new();
if self.wasm_import_definitions.len() > 0 {
imports_init.push_str("imports.");
imports_init.push_str(module_name);
imports_init.push_str(" = {};\n");
}
Add tests for the interface types output of wasm-bindgen (#1898) * Add tests for the interface types output of wasm-bindgen This commit expands the test suite with assertions about the output of the interface types pass in wasm-bindgen. The goal here is to actually assert that we produce the right output and have a suite of reference files to show how the interface types output is changing over time. The `reference` test suite added in the previous PR has been updated to work for interface types as well, generating `*.wit` file assertions which are printed via the `wit-printer` crate on crates.io. Along the way a number of bugs were fixed with the interface types output, such as: * Non-determinism in output caused by iteration of a `HashMap` * Avoiding JS generation entirely in interface types mode, ensuring that we don't export extraneous intrinsics that aren't otherwise needed. * Fixing location of the stack pointer for modules where it's GC'd out. It's now rooted in the aux section of wasm-bindgen so it's available to later passes, like the multi-value pass. * Interface types emission now works in debug mode, meaning the `--release` flag is no longer required. This previously did not work because the `__wbindgen_throw` intrinsic was required in debug mode. This comes about because of the `malloc_failure` and `internal_error` functions in the anyref pass. The purpose of these functions is to signal fatal runtime errors, if any, in a way that's usable to the user. For wasm interface types though we can replace calls to these functions with `unreachable` to avoid needing to import the intrinsic. This has the accidental side effect of making `wasm_bindgen::throw_str` "just work" with wasm interface types by aborting the program, but that's not actually entirely intended. It's hoped that a split of a `wasm-bindgen-core` crate would solve this issue for the future. * Run the wasm interface types validator in tests * Add more gc roots for adapter gc * Improve stack pointer detection The stack pointer is never initialized to zero, but some other mutable globals are (TLS, thread ID, etc), so let's filter those out.
2019-12-04 15:19:48 -06:00
for (id, js) in crate::sorted_iter(&self.wasm_import_definitions) {
let import = self.module.imports.get_mut(*id);
import.module = module_name.to_string();
imports_init.push_str("imports.");
imports_init.push_str(module_name);
imports_init.push_str(".");
imports_init.push_str(&import.name);
imports_init.push_str(" = ");
imports_init.push_str(js.trim());
imports_init.push_str(";\n");
}
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
let extra_modules = self
.module
.imports
.iter()
.filter(|i| !self.wasm_import_definitions.contains_key(&i.id()))
.filter(|i| {
// Importing memory is handled specially in this area, so don't
// consider this a candidate for importing from extra modules.
match i.kind {
walrus::ImportKind::Memory(_) => false,
_ => true,
}
})
.map(|i| &i.module)
.collect::<BTreeSet<_>>();
for (i, extra) in extra_modules.iter().enumerate() {
let imports = match &mut imports {
Some(list) => list,
None => bail!(
"cannot import from modules (`{}`) with `--no-modules`",
extra
),
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
};
imports.push_str(&format!("import * as __wbg_star{} from '{}';\n", i, extra));
imports_init.push_str(&format!("imports['{}'] = __wbg_star{};\n", extra, i));
}
2019-04-01 19:45:53 -03:00
let js = format!(
"\
async function load(module, imports{init_memory_arg}) {{
if (typeof Response === 'function' && module instanceof Response) {{
{init_memory2}
if (typeof WebAssembly.instantiateStreaming === 'function') {{
try {{
return await WebAssembly.instantiateStreaming(module, imports);
}} catch (e) {{
if (module.headers.get('Content-Type') != 'application/wasm') {{
console.warn(\"`WebAssembly.instantiateStreaming` failed \
because your server does not serve wasm with \
`application/wasm` MIME type. Falling back to \
`WebAssembly.instantiate` which is slower. Original \
error:\\n\", e);
}} else {{
throw e;
}}
}}
}}
const bytes = await module.arrayBuffer();
return await WebAssembly.instantiate(bytes, imports);
}} else {{
{init_memory1}
const instance = await WebAssembly.instantiate(module, imports);
if (instance instanceof WebAssembly.Instance) {{
return {{ instance, module }};
}} else {{
return instance;
}}
}}
}}
async function init(input{init_memory_arg}) {{
{default_module_path}
const imports = {{}};
{imports_init}
if (typeof input === 'string' || (typeof Request === 'function' && input instanceof Request) || (typeof URL === 'function' && input instanceof URL)) {{
input = fetch(input);
}}
const {{ instance, module }} = await load(await input, imports{init_memory_arg});
wasm = instance.exports;
init.__wbindgen_wasm_module = module;
{start}
return wasm;
}}
",
2019-04-01 19:45:53 -03:00
init_memory_arg = init_memory_arg,
default_module_path = default_module_path,
init_memory1 = init_memory1,
init_memory2 = init_memory2,
start = if needs_manual_start {
"wasm.__wbindgen_start();"
} else {
""
},
imports_init = imports_init,
2019-04-01 19:45:53 -03:00
);
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
Ok((js, ts))
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
fn write_classes(&mut self) -> Result<(), Error> {
for (class, exports) in self.exported_classes.take().unwrap() {
self.write_class(&class, &exports)?;
}
Ok(())
}
fn write_class(&mut self, name: &str, class: &ExportedClass) -> Result<(), Error> {
let mut dst = format!("class {} {{\n", name);
let mut ts_dst = format!("export {}", dst);
if self.config.debug && !class.has_constructor {
dst.push_str(
2018-06-27 22:42:34 -07:00
"
constructor() {
throw new Error('cannot invoke `new` directly');
}
",
);
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
if class.wrap_needed {
dst.push_str(&format!(
"
static __wrap(ptr) {{
const obj = Object.create({}.prototype);
obj.ptr = ptr;
{}
return obj;
}}
",
name,
if self.config.weak_refs {
format!("{}FinalizationGroup.register(obj, obj.ptr, obj.ptr);", name)
} else {
String::new()
},
));
}
if self.config.weak_refs {
self.global(&format!(
"
const {}FinalizationGroup = new FinalizationGroup((items) => {{
for (const ptr of items) {{
wasm.{}(ptr);
}}
}});
",
name,
wasm_bindgen_shared::free_function(&name),
));
}
// If the class is inspectable, generate `toJSON` and `toString`
// to expose all readable properties of the class. Otherwise,
// the class shows only the "ptr" property when logged or serialized
if class.is_inspectable {
// Creates a `toJSON` method which returns an object of all readable properties
// This object looks like { a: this.a, b: this.b }
dst.push_str(&format!(
"
toJSON() {{
return {{{}}};
}}
toString() {{
return JSON.stringify(this);
}}
",
class
.readable_properties
.iter()
.fold(String::from("\n"), |fields, field_name| {
format!("{}{name}: this.{name},\n", fields, name = field_name)
})
));
if self.config.mode.nodejs() {
// `util.inspect` must be imported in Node.js to define [inspect.custom]
let module_name = self.import_name(&JsImport {
name: JsImportName::Module {
module: "util".to_string(),
name: "inspect".to_string(),
},
fields: Vec::new(),
})?;
// Node.js supports a custom inspect function to control the
// output of `console.log` and friends. The constructor is set
// to display the class name as a typical JavaScript class would
dst.push_str(&format!(
"
[{}.custom]() {{
return Object.assign(Object.create({{constructor: this.constructor}}), this.toJSON());
}}
",
module_name
));
}
}
2018-06-27 22:42:34 -07:00
dst.push_str(&format!(
"
free() {{
const ptr = this.ptr;
this.ptr = 0;
{}
wasm.{}(ptr);
}}
",
if self.config.weak_refs {
format!("{}FinalizationGroup.unregister(ptr);", name)
} else {
String::new()
},
wasm_bindgen_shared::free_function(&name),
2018-06-27 22:42:34 -07:00
));
2019-06-06 16:11:51 -03:00
ts_dst.push_str(" free(): void;\n");
dst.push_str(&class.contents);
ts_dst.push_str(&class.typescript);
2019-05-30 10:56:59 -07:00
let mut fields = class.typescript_fields.keys().collect::<Vec<_>>();
fields.sort(); // make sure we have deterministic output
for name in fields {
let (ty, docs, has_setter, is_optional) = &class.typescript_fields[name];
ts_dst.push_str(docs);
2019-06-06 16:11:51 -03:00
ts_dst.push_str(" ");
if !has_setter {
2019-05-30 10:56:59 -07:00
ts_dst.push_str("readonly ");
}
ts_dst.push_str(name);
if *is_optional {
ts_dst.push_str("?: ");
} else {
ts_dst.push_str(": ");
}
ts_dst.push_str(&ty);
2019-05-30 10:56:59 -07:00
ts_dst.push_str(";\n");
}
dst.push_str("}\n");
ts_dst.push_str("}\n");
self.export(&name, &dst, Some(&class.comments))?;
self.typescript.push_str(&class.comments);
self.typescript.push_str(&ts_dst);
Ok(())
}
fn expose_drop_ref(&mut self) {
if !self.should_write_global("drop_ref") {
2018-04-07 13:06:36 +05:45
return;
}
self.expose_global_heap();
self.expose_global_heap_next();
// Note that here we check if `idx` shouldn't actually be dropped. This
// is due to the fact that `JsValue::null()` and friends can be passed
// by value to JS where we'll automatically call this method. Those
// constants, however, cannot be dropped. See #1054 for removing this
// branch.
//
// Otherwise the free operation here is pretty simple, just appending to
// the linked list of heap slots that are free.
2018-06-27 22:42:34 -07:00
self.global(&format!(
"
function dropObject(idx) {{
Add support for optional slice types (#507) * Shard the `convert.rs` module into sub-modules Hopefully this'll make the organization a little nicer over time! * Start adding support for optional types This commit starts adding support for optional types to wasm-bindgen as arguments/return values to functions. The strategy here is to add two new traits, `OptionIntoWasmAbi` and `OptionFromWasmAbi`. These two traits are used as a blanket impl to implement `IntoWasmAbi` and `FromWasmAbi` for `Option<T>`. Some consequences of this design: * It should be possible to ensure `Option<SomeForeignType>` implements to/from wasm traits. This is because the option-based traits can be implemented for foreign types. * A specialized implementation is possible for all types, so there's no need for `Option<T>` to introduce unnecessary overhead. * Two new traits is a bit unforutnate but I can't currently think of an alternative design that works for the above two constraints, although it doesn't mean one doesn't exist! * The error messages for "can't use this type here" is actually halfway decent because it says these new traits need to be implemented, which provides a good place to document and talk about what's going on here! * Nested references like `Option<&T>` can't implement `FromWasmAbi`. This means that you can't define a function in Rust which takes `Option<&str>`. It may be possible to do this one day but it'll likely require more trait trickery than I'm capable of right now. * Add support for optional slices This commit adds support for optional slice types, things like strings and arrays. The null representation of these has a pointer value of 0, which should never happen in normal Rust. Otherwise the various plumbing is done throughout the tooling to enable these types in all locations. * Fix `takeObject` on global sentinels These don't have a reference count as they're always expected to work, so avoid actually dropping a reference on them. * Remove some no longer needed bindings * Add support for optional anyref types This commit adds support for optional imported class types. Each type imported with `#[wasm_bindgen]` automatically implements the relevant traits and now supports `Option<Foo>` in various argument/return positions. * Fix building without the `std` feature * Actually fix the build... * Add support for optional types to WebIDL Closes #502
2018-07-19 14:44:23 -05:00
if (idx < {}) return;
heap[idx] = heap_next;
heap_next = idx;
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}}
",
INITIAL_HEAP_OFFSET + INITIAL_HEAP_VALUES.len(),
2018-06-27 22:42:34 -07:00
));
}
fn expose_global_heap(&mut self) {
if !self.should_write_global("heap") {
2018-04-07 13:06:36 +05:45
return;
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
assert!(!self.config.anyref);
self.global(&format!(
"const heap = new Array({}).fill(undefined);",
INITIAL_HEAP_OFFSET
));
self.global(&format!("heap.push({});", INITIAL_HEAP_VALUES.join(", ")));
}
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
fn expose_global_heap_next(&mut self) {
if !self.should_write_global("heap_next") {
2018-04-07 13:06:36 +05:45
return;
}
self.expose_global_heap();
self.global("let heap_next = heap.length;");
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
fn expose_get_object(&mut self) {
if !self.should_write_global("get_object") {
2018-04-07 13:06:36 +05:45
return;
}
self.expose_global_heap();
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
// Accessing a heap object is just a simple index operation due to how
// the stack/heap are laid out.
self.global("function getObject(idx) { return heap[idx]; }");
}
fn expose_not_defined(&mut self) {
if !self.should_write_global("not_defined") {
return;
}
self.global("function notDefined(what) { return () => { throw new Error(`${what} is not defined`); }; }");
}
fn expose_assert_num(&mut self) {
if !self.should_write_global("assert_num") {
2018-04-07 13:06:36 +05:45
return;
}
2018-06-27 22:42:34 -07:00
self.global(&format!(
"
function _assertNum(n) {{
2018-06-15 12:55:37 -05:00
if (typeof(n) !== 'number') throw new Error('expected a number argument');
}}
"
2018-06-27 22:42:34 -07:00
));
}
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
fn expose_assert_bool(&mut self) {
if !self.should_write_global("assert_bool") {
2018-04-07 13:06:36 +05:45
return;
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
2018-06-27 22:42:34 -07:00
self.global(&format!(
"
function _assertBoolean(n) {{
2018-06-15 12:55:37 -05:00
if (typeof(n) !== 'boolean') {{
throw new Error('expected a boolean argument');
2018-06-15 12:55:37 -05:00
}}
}}
"
2018-06-27 22:42:34 -07:00
));
}
fn expose_wasm_vector_len(&mut self) {
if !self.should_write_global("wasm_vector_len") {
return;
}
self.global("let WASM_VECTOR_LEN = 0;");
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_pass_string_to_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
self.expose_wasm_vector_len();
let debug = if self.config.debug {
"
2018-06-15 12:55:37 -05:00
if (typeof(arg) !== 'string') throw new Error('expected a string argument');
"
} else {
""
};
// If we are targeting Node.js, it doesn't have `encodeInto` yet
// but it does have `Buffer::write` which has similar semantics but
// doesn't require creating intermediate view using `subarray`
// and also has `Buffer::byteLength` to calculate size upfront.
if self.config.mode.nodejs() {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let get_buf = self.expose_node_buffer_memory(memory);
let ret = MemView {
name: "passStringToWasm",
num: get_buf.num,
};
if !self.should_write_global(ret.to_string()) {
return Ok(ret);
}
self.global(&format!(
"
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
function {}(arg, malloc) {{
{}
const len = Buffer.byteLength(arg);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
const ptr = malloc(len);
{}().write(arg, ptr, len);
WASM_VECTOR_LEN = len;
return ptr;
}}
",
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
ret, debug, get_buf,
));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
return Ok(ret);
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let mem = self.expose_uint8_memory(memory);
let ret = MemView {
name: "passStringToWasm",
num: mem.num,
};
if !self.should_write_global(ret.to_string()) {
return Ok(ret);
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.expose_text_encoder()?;
// The first implementation we have for this is to use
// `TextEncoder#encode` which has been around for quite some time.
let encode = "function (arg, view) {
const buf = cachedTextEncoder.encode(arg);
view.set(buf);
return {
read: arg.length,
written: buf.length
};
}";
// Another possibility is to use `TextEncoder#encodeInto` which is much
// newer and isn't implemented everywhere yet. It's more efficient,
2019-08-24 19:49:26 +02:00
// however, becaues it allows us to elide an intermediate allocation.
let encode_into = "function (arg, view) {
return cachedTextEncoder.encodeInto(arg, view);
}";
// Looks like `encodeInto` doesn't currently work when the memory passed
// in is backed by a `SharedArrayBuffer`, so force usage of `encode` if
// a `SharedArrayBuffer` is in use.
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let shared = self.module.memories.get(memory).shared;
match self.config.encode_into {
EncodeInto::Always if !shared => {
2019-09-10 11:20:19 -07:00
self.global(&format!(
"
const encodeString = {};
2019-09-10 11:20:19 -07:00
",
encode_into
));
}
EncodeInto::Test if !shared => {
2019-09-10 11:20:19 -07:00
self.global(&format!(
"
const encodeString = (typeof cachedTextEncoder.encodeInto === 'function'
? {}
: {});
2019-09-10 11:20:19 -07:00
",
encode_into, encode
));
}
_ => {
2019-09-10 11:20:19 -07:00
self.global(&format!(
"
const encodeString = {};
2019-09-10 11:20:19 -07:00
",
encode
));
}
}
// A fast path that directly writes char codes into WASM memory as long
// as it finds only ASCII characters.
//
// This is much faster for common ASCII strings because it can avoid
// calling out into C++ TextEncoder code.
//
// This might be not very intuitive, but such calls are usually more
// expensive in mainstream engines than staying in the JS, and
// charCodeAt on ASCII strings is usually optimised to raw bytes.
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let encode_as_ascii = format!(
"\
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
if (realloc === undefined) {{
const buf = cachedTextEncoder.encode(arg);
const ptr = malloc(buf.length);
{mem}().subarray(ptr, ptr + buf.length).set(buf);
WASM_VECTOR_LEN = buf.length;
return ptr;
}}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let len = arg.length;
let ptr = malloc(len);
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
const mem = {mem}();
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let offset = 0;
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
for (; offset < len; offset++) {{
const code = arg.charCodeAt(offset);
if (code > 0x7F) break;
mem[ptr + offset] = code;
}}
",
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
mem = mem,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
);
// TODO:
// When converting a JS string to UTF-8, the maximum size is `arg.length * 3`,
// so we just allocate that. This wastes memory, so we should investigate
// looping over the string to calculate the precise size, or perhaps using
// `shrink_to_fit` on the Rust side.
self.global(&format!(
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
"function {name}(arg, malloc, realloc) {{
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
{debug}
{ascii}
if (offset !== len) {{
if (offset !== 0) {{
arg = arg.slice(offset);
}}
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
ptr = realloc(ptr, len, len = offset + arg.length * 3);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
const view = {mem}().subarray(ptr + offset, ptr + len);
const ret = encodeString(arg, view);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
{debug_end}
offset += ret.written;
}}
WASM_VECTOR_LEN = offset;
return ptr;
}}",
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
name = ret,
debug = debug,
ascii = encode_as_ascii,
mem = mem,
debug_end = if self.config.debug {
"if (ret.read !== arg.length) throw new Error('failed to pass whole string');"
2019-05-13 08:12:32 -07:00
} else {
""
},
));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
Ok(ret)
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_pass_array8_to_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
let view = self.expose_uint8_memory(memory);
self.pass_array_to_wasm("passArray8ToWasm", view, 1)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_pass_array16_to_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
let view = self.expose_uint16_memory(memory);
self.pass_array_to_wasm("passArray16ToWasm", view, 2)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_pass_array32_to_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
let view = self.expose_uint32_memory(memory);
self.pass_array_to_wasm("passArray32ToWasm", view, 4)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_pass_array64_to_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
let view = self.expose_uint64_memory(memory);
self.pass_array_to_wasm("passArray64ToWasm", view, 8)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_pass_array_f32_to_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
let view = self.expose_f32_memory(memory);
self.pass_array_to_wasm("passArrayF32ToWasm", view, 4)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_pass_array_f64_to_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
let view = self.expose_f64_memory(memory);
self.pass_array_to_wasm("passArrayF64ToWasm", view, 8)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_pass_array_jsvalue_to_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
let mem = self.expose_uint32_memory(memory);
let ret = MemView {
name: "passArrayJsValueToWasm",
num: mem.num,
};
if !self.should_write_global(ret.to_string()) {
return Ok(ret);
}
self.expose_wasm_vector_len();
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
match (self.aux.anyref_table, self.aux.anyref_alloc) {
(Some(table), Some(alloc)) => {
// TODO: using `addToAnyrefTable` goes back and forth between wasm
// and JS a lot, we should have a bulk operation for this.
let add = self.expose_add_to_anyref_table(table, alloc)?;
self.global(&format!(
"
function {}(array, malloc) {{
const ptr = malloc(array.length * 4);
const mem = {}();
for (let i = 0; i < array.length; i++) {{
mem[ptr / 4 + i] = {}(array[i]);
}}
WASM_VECTOR_LEN = array.length;
return ptr;
}}
",
ret, mem, add,
));
}
_ => {
self.expose_add_heap_object();
self.global(&format!(
"
function {}(array, malloc) {{
const ptr = malloc(array.length * 4);
const mem = {}();
for (let i = 0; i < array.length; i++) {{
mem[ptr / 4 + i] = addHeapObject(array[i]);
}}
WASM_VECTOR_LEN = array.length;
return ptr;
}}
",
ret, mem,
));
}
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
Ok(ret)
}
2018-06-27 22:42:34 -07:00
fn pass_array_to_wasm(
&mut self,
name: &'static str,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
view: MemView,
2018-06-27 22:42:34 -07:00
size: usize,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
) -> Result<MemView, Error> {
let ret = MemView {
name,
num: view.num,
};
if !self.should_write_global(ret.to_string()) {
return Ok(ret);
}
self.expose_wasm_vector_len();
2018-06-27 22:42:34 -07:00
self.global(&format!(
"
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
function {}(arg, malloc) {{
const ptr = malloc(arg.length * {size});
{}().set(arg, ptr / {size});
WASM_VECTOR_LEN = arg.length;
return ptr;
}}
",
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
ret,
view,
2018-06-27 22:42:34 -07:00
size = size
));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
Ok(ret)
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn expose_text_encoder(&mut self) -> Result<(), Error> {
if !self.should_write_global("text_encoder") {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
return Ok(());
}
self.expose_text_processor("TextEncoder", "('utf-8')")
2018-02-05 16:39:11 -08:00
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn expose_text_decoder(&mut self) -> Result<(), Error> {
if !self.should_write_global("text_decoder") {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
return Ok(());
2018-02-05 16:39:11 -08:00
}
2019-10-23 00:26:28 +02:00
// `ignoreBOM` is needed so that the BOM will be preserved when sending a string from Rust to JS
// `fatal` is needed to catch any weird encoding bugs when sending a string from Rust to JS
self.expose_text_processor("TextDecoder", "('utf-8', { ignoreBOM: true, fatal: true })")?;
2019-10-23 00:26:28 +02:00
// This is needed to workaround a bug in Safari
// See: https://github.com/rustwasm/wasm-bindgen/issues/1825
self.global("cachedTextDecoder.decode();");
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
Ok(())
}
fn expose_text_processor(&mut self, s: &str, args: &str) -> Result<(), Error> {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
if self.config.mode.nodejs() {
let name = self.import_name(&JsImport {
name: JsImportName::Module {
module: "util".to_string(),
name: s.to_string(),
},
fields: Vec::new(),
})?;
self.global(&format!("let cached{} = new {}{};", s, name, args));
} else if !self.config.mode.always_run_in_browser() {
2018-11-27 12:07:59 -08:00
self.global(&format!(
"
const l{0} = typeof {0} === 'undefined' ? \
(0, module.require)('util').{0} : {0};\
2018-11-27 12:07:59 -08:00
",
s
));
self.global(&format!("let cached{0} = new l{0}{1};", s, args));
} else {
self.global(&format!("let cached{0} = new {0}{1};", s, args));
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
Ok(())
}
2018-02-05 16:39:11 -08:00
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_string_from_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.expose_text_decoder()?;
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let mem = self.expose_uint8_memory(memory);
let ret = MemView {
name: "getStringFromWasm",
num: mem.num,
};
if !self.should_write_global(ret.to_string()) {
return Ok(ret);
}
// Typically we try to give a raw view of memory out to `TextDecoder` to
// avoid copying too much data. If, however, a `SharedArrayBuffer` is
// being used it looks like that is rejected by `TextDecoder` or
// otherwise doesn't work with it. When we detect a shared situation we
// use `slice` which creates a new array instead of `subarray` which
// creates just a view. That way in shared mode we copy more data but in
// non-shared mode there's no need to copy the data except for the
// string itself.
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let is_shared = self.module.memories.get(memory).shared;
let method = if is_shared { "slice" } else { "subarray" };
self.global(&format!(
"
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
function {}(ptr, len) {{
return cachedTextDecoder.decode({}().{}(ptr, ptr + len));
}}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
",
ret, mem, method
));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
Ok(ret)
2018-02-05 16:39:11 -08:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_cached_string_from_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
2019-07-16 23:03:18 +02:00
self.expose_get_object();
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let get = self.expose_get_string_from_wasm(memory)?;
let ret = MemView {
name: "getCachedStringFromWasm",
num: get.num,
};
if !self.should_write_global(ret.to_string()) {
return Ok(ret);
}
2019-07-16 23:03:18 +02:00
2019-07-18 00:11:40 +02:00
// This has support for both `&str` and `Option<&str>`.
//
// If `ptr` is not `0` then we know that it's a `&str` or `Some(&str)`, so we just decode it.
//
// If `ptr` is `0` then the `len` is a pointer to the cached `JsValue`, so we return that.
//
// If `ptr` and `len` are both `0` then that means it's `None`, in that case we rely upon
// the fact that `getObject(0)` is guaranteed to be `undefined`.
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
self.global(&format!(
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
"
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
function {}(ptr, len) {{
if (ptr === 0) {{
2019-07-18 00:11:40 +02:00
return getObject(len);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}} else {{
return {}(ptr, len);
}}
}}
",
ret, get,
));
Ok(ret)
2019-07-16 23:03:18 +02:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_js_value_from_wasm(&mut self, memory: MemoryId) -> Result<MemView, Error> {
let mem = self.expose_uint32_memory(memory);
let ret = MemView {
name: "getArrayJsValueFromWasm",
num: mem.num,
};
if !self.should_write_global(ret.to_string()) {
return Ok(ret);
2018-02-28 10:56:56 +01:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
match (self.aux.anyref_table, self.aux.anyref_drop_slice) {
(Some(table), Some(drop)) => {
let table = self.export_name_of(table);
let drop = self.export_name_of(drop);
self.global(&format!(
"
function {}(ptr, len) {{
const mem = {}();
const slice = mem.subarray(ptr / 4, ptr / 4 + len);
const result = [];
for (let i = 0; i < slice.length; i++) {{
result.push(wasm.{}.get(slice[i]));
}}
wasm.{}(ptr, len);
return result;
}}
",
ret, mem, table, drop,
));
}
_ => {
self.expose_take_object();
self.global(&format!(
"
function {}(ptr, len) {{
const mem = {}();
const slice = mem.subarray(ptr / 4, ptr / 4 + len);
const result = [];
for (let i = 0; i < slice.length; i++) {{
result.push(takeObject(slice[i]));
}}
return result;
}}
",
ret, mem,
));
}
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
Ok(ret)
2018-02-28 10:56:56 +01:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_i8_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_int8_memory(memory);
self.arrayget("getArrayI8FromWasm", view, 1)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_u8_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_uint8_memory(memory);
self.arrayget("getArrayU8FromWasm", view, 1)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_clamped_array_u8_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_clamped_uint8_memory(memory);
self.arrayget("getClampedArrayU8FromWasm", view, 1)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_i16_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_int16_memory(memory);
self.arrayget("getArrayI16FromWasm", view, 2)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_u16_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_uint16_memory(memory);
self.arrayget("getArrayU16FromWasm", view, 2)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_i32_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_int32_memory(memory);
self.arrayget("getArrayI32FromWasm", view, 4)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_u32_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_uint32_memory(memory);
self.arrayget("getArrayU32FromWasm", view, 4)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_i64_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_int64_memory(memory);
self.arrayget("getArrayI64FromWasm", view, 8)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_u64_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_uint64_memory(memory);
self.arrayget("getArrayU64FromWasm", view, 8)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_f32_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_f32_memory(memory);
self.arrayget("getArrayF32FromWasm", view, 4)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_array_f64_from_wasm(&mut self, memory: MemoryId) -> MemView {
let view = self.expose_f64_memory(memory);
self.arrayget("getArrayF64FromWasm", view, 8)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn arrayget(&mut self, name: &'static str, view: MemView, size: usize) -> MemView {
let ret = MemView {
name,
num: view.num,
};
if !self.should_write_global(name) {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
return ret;
}
2018-06-27 22:42:34 -07:00
self.global(&format!(
"
function {name}(ptr, len) {{
return {mem}().subarray(ptr / {size}, ptr / {size} + len);
}}
",
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
name = ret,
mem = view,
size = size,
));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
return ret;
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_node_buffer_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getNodeBufferMemory", "Buffer.from", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_int8_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getInt8Memory", "new Int8Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_uint8_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getUint8Memory", "new Uint8Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_clamped_uint8_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getUint8ClampedMemory", "new Uint8ClampedArray", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_int16_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getInt16Memory", "new Int16Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_uint16_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getUint16Memory", "new Uint16Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_int32_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getInt32Memory", "new Int32Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_uint32_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getUint32Memory", "new Uint32Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_int64_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getInt64Memory", "new BigInt64Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_uint64_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getUint64Memory", "new BigUint64Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_f32_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getFloat32Memory", "new Float32Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_f64_memory(&mut self, memory: MemoryId) -> MemView {
self.memview("getFloat64Memory", "new Float64Array", memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn memview_function(&mut self, t: VectorKind, memory: MemoryId) -> MemView {
match t {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
VectorKind::String => self.expose_uint8_memory(memory),
VectorKind::I8 => self.expose_int8_memory(memory),
VectorKind::U8 => self.expose_uint8_memory(memory),
VectorKind::ClampedU8 => self.expose_clamped_uint8_memory(memory),
VectorKind::I16 => self.expose_int16_memory(memory),
VectorKind::U16 => self.expose_uint16_memory(memory),
VectorKind::I32 => self.expose_int32_memory(memory),
VectorKind::U32 => self.expose_uint32_memory(memory),
VectorKind::I64 => self.expose_int64_memory(memory),
VectorKind::U64 => self.expose_uint64_memory(memory),
VectorKind::F32 => self.expose_f32_memory(memory),
VectorKind::F64 => self.expose_f64_memory(memory),
VectorKind::Anyref => self.expose_uint32_memory(memory),
}
}
fn memview(&mut self, name: &'static str, js: &str, memory: walrus::MemoryId) -> MemView {
let view = self.memview_memory(name, memory);
if !self.should_write_global(name.to_string()) {
return view;
}
let mem = self.export_name_of(memory);
2018-06-27 22:42:34 -07:00
self.global(&format!(
"
let cache{name} = null;
function {name}() {{
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
if (cache{name} === null || cache{name}.buffer !== wasm.{mem}.buffer) {{
cache{name} = {js}(wasm.{mem}.buffer);
2018-06-15 12:55:37 -05:00
}}
return cache{name};
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}}
",
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
name = view,
js = js,
mem = mem,
));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
return view;
}
fn memview_memory(&mut self, name: &'static str, memory: walrus::MemoryId) -> MemView {
let next = self.memory_indices.len();
let num = *self.memory_indices.entry(memory).or_insert(next);
MemView { name, num }
}
fn memview_table(&mut self, name: &'static str, table: walrus::TableId) -> MemView {
let next = self.table_indices.len();
let num = *self.table_indices.entry(table).or_insert(next);
MemView { name, num }
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
fn expose_assert_class(&mut self) {
if !self.should_write_global("assert_class") {
2018-04-07 13:06:36 +05:45
return;
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
self.global(
2018-06-27 22:42:34 -07:00
"
function _assertClass(instance, klass) {
if (!(instance instanceof klass)) {
throw new Error(`expected instance of ${klass.name}`);
}
return instance.ptr;
}
",
);
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
fn expose_global_stack_pointer(&mut self) {
if !self.should_write_global("stack_pointer") {
return;
}
self.global(&format!("let stack_pointer = {};", INITIAL_HEAP_OFFSET));
}
fn expose_borrowed_objects(&mut self) {
if !self.should_write_global("borrowed_objects") {
2018-04-07 13:06:36 +05:45
return;
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
self.expose_global_heap();
self.expose_global_stack_pointer();
// Our `stack_pointer` points to where we should start writing stack
// objects, and the `stack_pointer` is incremented in a `finally` block
// after executing this. Once we've reserved stack space we write the
// value. Eventually underflow will throw an exception, but JS sort of
// just handles it today...
self.global(
2018-06-27 22:42:34 -07:00
"
function addBorrowedObject(obj) {
if (stack_pointer == 1) throw new Error('out of js stack');
heap[--stack_pointer] = obj;
return stack_pointer;
}
2018-11-30 13:04:05 -08:00
",
);
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
fn expose_take_object(&mut self) {
if !self.should_write_global("take_object") {
2018-04-07 13:06:36 +05:45
return;
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
self.expose_get_object();
self.expose_drop_ref();
self.global(
2018-06-27 22:42:34 -07:00
"
function takeObject(idx) {
const ret = getObject(idx);
dropObject(idx);
return ret;
}
",
);
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
fn expose_add_heap_object(&mut self) {
if !self.should_write_global("add_heap_object") {
2018-04-07 13:06:36 +05:45
return;
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
self.expose_global_heap();
self.expose_global_heap_next();
let set_heap_next = if self.config.debug {
2018-06-27 22:42:34 -07:00
String::from(
"
if (typeof(heap_next) !== 'number') throw new Error('corrupt heap');
",
2018-06-27 22:42:34 -07:00
)
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
} else {
String::new()
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
};
// Allocating a slot on the heap first goes through the linked list
// (starting at `heap_next`). Once that linked list is exhausted we'll
// be pointing beyond the end of the array, at which point we'll reserve
// one more slot and use that.
2018-06-27 22:42:34 -07:00
self.global(&format!(
"
function addHeapObject(obj) {{
if (heap_next === heap.length) heap.push(heap.length + 1);
const idx = heap_next;
heap_next = heap[idx];
{}
heap[idx] = obj;
return idx;
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}}
",
set_heap_next
2018-06-27 22:42:34 -07:00
));
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
}
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
fn expose_handle_error(&mut self) -> Result<(), Error> {
2019-02-04 02:08:08 +01:00
if !self.should_write_global("handle_error") {
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
return Ok(());
2019-02-04 02:08:08 +01:00
}
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
let store = self
.aux
.exn_store
.ok_or_else(|| anyhow!("failed to find `__wbindgen_exn_store` intrinsic"))?;
let store = self.export_name_of(store);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
match (self.aux.anyref_table, self.aux.anyref_alloc) {
(Some(table), Some(alloc)) => {
let add = self.expose_add_to_anyref_table(table, alloc)?;
self.global(&format!(
"
function handleError(f) {{
return function () {{
try {{
return f.apply(this, arguments);
}} catch (e) {{
const idx = {}(e);
wasm.{}(idx);
}}
}};
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}}
",
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
add, store,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
));
}
_ => {
self.expose_add_heap_object();
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
self.global(&format!(
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
"
function handleError(f) {{
return function () {{
try {{
return f.apply(this, arguments);
}} catch (e) {{
wasm.{}(addHeapObject(e));
}}
}};
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
}}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
",
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
store,
));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
Ok(())
2019-02-04 02:08:08 +01:00
}
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
fn expose_log_error(&mut self) {
if !self.should_write_global("log_error") {
return;
}
self.global(
"\
function logError(f) {
return function () {
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
try {
return f.apply(this, arguments);
} catch (e) {
let error = (function () {
try {
return e instanceof Error \
? `${e.message}\\n\\nStack:\\n${e.stack}` \
: e.toString();
} catch(_) {
return \"<failed to stringify thrown value>\";
}
}());
console.error(\"wasm-bindgen: imported JS function that \
was not marked as `catch` threw an error:\", \
error);
throw e;
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
};
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
",
);
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn pass_to_wasm_function(&mut self, t: VectorKind, memory: MemoryId) -> Result<MemView, Error> {
match t {
VectorKind::String => self.expose_pass_string_to_wasm(memory),
VectorKind::I8 | VectorKind::U8 | VectorKind::ClampedU8 => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
self.expose_pass_array8_to_wasm(memory)
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
VectorKind::U16 | VectorKind::I16 => self.expose_pass_array16_to_wasm(memory),
VectorKind::I32 | VectorKind::U32 => self.expose_pass_array32_to_wasm(memory),
VectorKind::I64 | VectorKind::U64 => self.expose_pass_array64_to_wasm(memory),
VectorKind::F32 => self.expose_pass_array_f32_to_wasm(memory),
VectorKind::F64 => self.expose_pass_array_f64_to_wasm(memory),
VectorKind::Anyref => self.expose_pass_array_jsvalue_to_wasm(memory),
}
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_get_vector_from_wasm(
&mut self,
ty: VectorKind,
memory: MemoryId,
) -> Result<MemView, Error> {
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
Ok(match ty {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
VectorKind::String => self.expose_get_string_from_wasm(memory)?,
VectorKind::I8 => self.expose_get_array_i8_from_wasm(memory),
VectorKind::U8 => self.expose_get_array_u8_from_wasm(memory),
VectorKind::ClampedU8 => self.expose_get_clamped_array_u8_from_wasm(memory),
VectorKind::I16 => self.expose_get_array_i16_from_wasm(memory),
VectorKind::U16 => self.expose_get_array_u16_from_wasm(memory),
VectorKind::I32 => self.expose_get_array_i32_from_wasm(memory),
VectorKind::U32 => self.expose_get_array_u32_from_wasm(memory),
VectorKind::I64 => self.expose_get_array_i64_from_wasm(memory),
VectorKind::U64 => self.expose_get_array_u64_from_wasm(memory),
VectorKind::F32 => self.expose_get_array_f32_from_wasm(memory),
VectorKind::F64 => self.expose_get_array_f64_from_wasm(memory),
VectorKind::Anyref => self.expose_get_array_js_value_from_wasm(memory)?,
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
})
}
Start removal of vector special-casing This commit starts wasm-bindgen down a path of removing the special casing it currently has around vectors, slices, and strings. This has long been a thorn in wasm-bindgen's side as it doesn't handle other kinds of vectors and otherwise is very inflexible with future additions. Additionally it leads to a lot of duplicated-ish code throughout various portions of codegen. The fundamental reason for this was that two arguments were required to be passed back to wasm, and I couldn't figure out a way to shove both those arguments into a function argument. The new strategy here is that there is one global stack well known to both JS and Rust which arguments *may* also be transferred between. By default all ABI arguments pass as literal function arguments, but if two or more arguments need to be passed then the extra ones are all passed through this global stack. The stack is effectively temporary scratch space when crossing the JS/Rust boundary (both ways). No long term storage is intended here. The `simple` test is passing as a result of this commit, using strings internally. The `Vector` type in the AST has been removed (yay!) and the bulk of the implementation of slices and vectors now resides in the `wasm-bindgen` crate itself, defining how to pass all these arguments around. The JS generator, however, still needs to know about all the sorts of vectors so it can generate appropriate code for JS. Future commits will continue cleanup and get the rest of the tests working.
2018-03-31 07:57:47 -07:00
fn expose_get_inherited_descriptor(&mut self) {
if !self.should_write_global("get_inherited_descriptor") {
2018-06-27 22:42:34 -07:00
return;
}
// It looks like while rare some browsers will move descriptors up the
// property chain which runs the risk of breaking wasm-bindgen-generated
// code because we're looking for precise descriptor functions rather
// than relying on the prototype chain like most "normal JS" projects
// do.
//
// As a result we have a small helper here which will walk the prototype
// chain looking for a descriptor. For some more information on this see
// #109
2018-06-27 22:42:34 -07:00
self.global(
"
function GetOwnOrInheritedPropertyDescriptor(obj, id) {
while (obj) {
let desc = Object.getOwnPropertyDescriptor(obj, id);
if (desc) return desc;
obj = Object.getPrototypeOf(obj);
}
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
return {};
}
",
2018-06-27 22:42:34 -07:00
);
}
fn expose_u32_cvt_shim(&mut self) -> &'static str {
let name = "u32CvtShim";
if !self.should_write_global(name) {
2018-06-27 22:42:34 -07:00
return name;
}
self.global(&format!("const {} = new Uint32Array(2);", name));
name
}
fn expose_int64_cvt_shim(&mut self) -> &'static str {
let name = "int64CvtShim";
if !self.should_write_global(name) {
2018-06-27 22:42:34 -07:00
return name;
}
let n = self.expose_u32_cvt_shim();
2018-06-27 22:42:34 -07:00
self.global(&format!(
"const {} = new BigInt64Array({}.buffer);",
name, n
));
name
}
fn expose_uint64_cvt_shim(&mut self) -> &'static str {
let name = "uint64CvtShim";
if !self.should_write_global(name) {
2018-06-27 22:42:34 -07:00
return name;
}
let n = self.expose_u32_cvt_shim();
2018-06-27 22:42:34 -07:00
self.global(&format!(
"const {} = new BigUint64Array({}.buffer);",
name, n
));
name
}
Add support for optional slice types (#507) * Shard the `convert.rs` module into sub-modules Hopefully this'll make the organization a little nicer over time! * Start adding support for optional types This commit starts adding support for optional types to wasm-bindgen as arguments/return values to functions. The strategy here is to add two new traits, `OptionIntoWasmAbi` and `OptionFromWasmAbi`. These two traits are used as a blanket impl to implement `IntoWasmAbi` and `FromWasmAbi` for `Option<T>`. Some consequences of this design: * It should be possible to ensure `Option<SomeForeignType>` implements to/from wasm traits. This is because the option-based traits can be implemented for foreign types. * A specialized implementation is possible for all types, so there's no need for `Option<T>` to introduce unnecessary overhead. * Two new traits is a bit unforutnate but I can't currently think of an alternative design that works for the above two constraints, although it doesn't mean one doesn't exist! * The error messages for "can't use this type here" is actually halfway decent because it says these new traits need to be implemented, which provides a good place to document and talk about what's going on here! * Nested references like `Option<&T>` can't implement `FromWasmAbi`. This means that you can't define a function in Rust which takes `Option<&str>`. It may be possible to do this one day but it'll likely require more trait trickery than I'm capable of right now. * Add support for optional slices This commit adds support for optional slice types, things like strings and arrays. The null representation of these has a pointer value of 0, which should never happen in normal Rust. Otherwise the various plumbing is done throughout the tooling to enable these types in all locations. * Fix `takeObject` on global sentinels These don't have a reference count as they're always expected to work, so avoid actually dropping a reference on them. * Remove some no longer needed bindings * Add support for optional anyref types This commit adds support for optional imported class types. Each type imported with `#[wasm_bindgen]` automatically implements the relevant traits and now supports `Option<Foo>` in various argument/return positions. * Fix building without the `std` feature * Actually fix the build... * Add support for optional types to WebIDL Closes #502
2018-07-19 14:44:23 -05:00
fn expose_is_like_none(&mut self) {
if !self.should_write_global("is_like_none") {
return;
Add support for optional slice types (#507) * Shard the `convert.rs` module into sub-modules Hopefully this'll make the organization a little nicer over time! * Start adding support for optional types This commit starts adding support for optional types to wasm-bindgen as arguments/return values to functions. The strategy here is to add two new traits, `OptionIntoWasmAbi` and `OptionFromWasmAbi`. These two traits are used as a blanket impl to implement `IntoWasmAbi` and `FromWasmAbi` for `Option<T>`. Some consequences of this design: * It should be possible to ensure `Option<SomeForeignType>` implements to/from wasm traits. This is because the option-based traits can be implemented for foreign types. * A specialized implementation is possible for all types, so there's no need for `Option<T>` to introduce unnecessary overhead. * Two new traits is a bit unforutnate but I can't currently think of an alternative design that works for the above two constraints, although it doesn't mean one doesn't exist! * The error messages for "can't use this type here" is actually halfway decent because it says these new traits need to be implemented, which provides a good place to document and talk about what's going on here! * Nested references like `Option<&T>` can't implement `FromWasmAbi`. This means that you can't define a function in Rust which takes `Option<&str>`. It may be possible to do this one day but it'll likely require more trait trickery than I'm capable of right now. * Add support for optional slices This commit adds support for optional slice types, things like strings and arrays. The null representation of these has a pointer value of 0, which should never happen in normal Rust. Otherwise the various plumbing is done throughout the tooling to enable these types in all locations. * Fix `takeObject` on global sentinels These don't have a reference count as they're always expected to work, so avoid actually dropping a reference on them. * Remove some no longer needed bindings * Add support for optional anyref types This commit adds support for optional imported class types. Each type imported with `#[wasm_bindgen]` automatically implements the relevant traits and now supports `Option<Foo>` in various argument/return positions. * Fix building without the `std` feature * Actually fix the build... * Add support for optional types to WebIDL Closes #502
2018-07-19 14:44:23 -05:00
}
self.global(
"
Add support for optional slice types (#507) * Shard the `convert.rs` module into sub-modules Hopefully this'll make the organization a little nicer over time! * Start adding support for optional types This commit starts adding support for optional types to wasm-bindgen as arguments/return values to functions. The strategy here is to add two new traits, `OptionIntoWasmAbi` and `OptionFromWasmAbi`. These two traits are used as a blanket impl to implement `IntoWasmAbi` and `FromWasmAbi` for `Option<T>`. Some consequences of this design: * It should be possible to ensure `Option<SomeForeignType>` implements to/from wasm traits. This is because the option-based traits can be implemented for foreign types. * A specialized implementation is possible for all types, so there's no need for `Option<T>` to introduce unnecessary overhead. * Two new traits is a bit unforutnate but I can't currently think of an alternative design that works for the above two constraints, although it doesn't mean one doesn't exist! * The error messages for "can't use this type here" is actually halfway decent because it says these new traits need to be implemented, which provides a good place to document and talk about what's going on here! * Nested references like `Option<&T>` can't implement `FromWasmAbi`. This means that you can't define a function in Rust which takes `Option<&str>`. It may be possible to do this one day but it'll likely require more trait trickery than I'm capable of right now. * Add support for optional slices This commit adds support for optional slice types, things like strings and arrays. The null representation of these has a pointer value of 0, which should never happen in normal Rust. Otherwise the various plumbing is done throughout the tooling to enable these types in all locations. * Fix `takeObject` on global sentinels These don't have a reference count as they're always expected to work, so avoid actually dropping a reference on them. * Remove some no longer needed bindings * Add support for optional anyref types This commit adds support for optional imported class types. Each type imported with `#[wasm_bindgen]` automatically implements the relevant traits and now supports `Option<Foo>` in various argument/return positions. * Fix building without the `std` feature * Actually fix the build... * Add support for optional types to WebIDL Closes #502
2018-07-19 14:44:23 -05:00
function isLikeNone(x) {
return x === undefined || x === null;
}
",
);
Add support for optional slice types (#507) * Shard the `convert.rs` module into sub-modules Hopefully this'll make the organization a little nicer over time! * Start adding support for optional types This commit starts adding support for optional types to wasm-bindgen as arguments/return values to functions. The strategy here is to add two new traits, `OptionIntoWasmAbi` and `OptionFromWasmAbi`. These two traits are used as a blanket impl to implement `IntoWasmAbi` and `FromWasmAbi` for `Option<T>`. Some consequences of this design: * It should be possible to ensure `Option<SomeForeignType>` implements to/from wasm traits. This is because the option-based traits can be implemented for foreign types. * A specialized implementation is possible for all types, so there's no need for `Option<T>` to introduce unnecessary overhead. * Two new traits is a bit unforutnate but I can't currently think of an alternative design that works for the above two constraints, although it doesn't mean one doesn't exist! * The error messages for "can't use this type here" is actually halfway decent because it says these new traits need to be implemented, which provides a good place to document and talk about what's going on here! * Nested references like `Option<&T>` can't implement `FromWasmAbi`. This means that you can't define a function in Rust which takes `Option<&str>`. It may be possible to do this one day but it'll likely require more trait trickery than I'm capable of right now. * Add support for optional slices This commit adds support for optional slice types, things like strings and arrays. The null representation of these has a pointer value of 0, which should never happen in normal Rust. Otherwise the various plumbing is done throughout the tooling to enable these types in all locations. * Fix `takeObject` on global sentinels These don't have a reference count as they're always expected to work, so avoid actually dropping a reference on them. * Remove some no longer needed bindings * Add support for optional anyref types This commit adds support for optional imported class types. Each type imported with `#[wasm_bindgen]` automatically implements the relevant traits and now supports `Option<Foo>` in various argument/return positions. * Fix building without the `std` feature * Actually fix the build... * Add support for optional types to WebIDL Closes #502
2018-07-19 14:44:23 -05:00
}
fn expose_make_mut_closure(&mut self) -> Result<(), Error> {
if !self.should_write_global("make_mut_closure") {
return Ok(());
}
let table = self.export_function_table()?;
// For mutable closures they can't be invoked recursively.
// To handle that we swap out the `this.a` pointer with zero
// while we invoke it. If we finish and the closure wasn't
// destroyed, then we put back the pointer so a future
// invocation can succeed.
self.global(&format!(
"
function makeMutClosure(arg0, arg1, dtor, f) {{
const state = {{ a: arg0, b: arg1, cnt: 1 }};
const real = (...args) => {{
// First up with a closure we increment the internal reference
// count. This ensures that the Rust closure environment won't
// be deallocated while we're invoking it.
state.cnt++;
const a = state.a;
state.a = 0;
try {{
return f(a, state.b, ...args);
}} finally {{
if (--state.cnt === 0) wasm.{}.get(dtor)(a, state.b);
else state.a = a;
}}
}};
real.original = state;
return real;
}}
",
table
));
Ok(())
}
fn expose_make_closure(&mut self) -> Result<(), Error> {
if !self.should_write_global("make_closure") {
return Ok(());
}
let table = self.export_function_table()?;
// For shared closures they can be invoked recursively so we
// just immediately pass through `this.a`. If we end up
// executing the destructor, however, we clear out the
// `this.a` pointer to prevent it being used again the
// future.
self.global(&format!(
"
function makeClosure(arg0, arg1, dtor, f) {{
const state = {{ a: arg0, b: arg1, cnt: 1 }};
const real = (...args) => {{
// First up with a closure we increment the internal reference
// count. This ensures that the Rust closure environment won't
// be deallocated while we're invoking it.
state.cnt++;
try {{
return f(state.a, state.b, ...args);
}} finally {{
if (--state.cnt === 0) {{
wasm.{}.get(dtor)(state.a, state.b);
state.a = 0;
}}
}}
}};
real.original = state;
return real;
}}
",
table
));
Ok(())
}
2018-04-14 09:37:49 -07:00
fn global(&mut self, s: &str) {
let s = s.trim();
// Ensure a blank line between adjacent items, and ensure everything is
// terminated with a newline.
while !self.globals.ends_with("\n\n\n") && !self.globals.ends_with("*/\n") {
2018-04-14 09:37:49 -07:00
self.globals.push_str("\n");
}
self.globals.push_str(s);
self.globals.push_str("\n");
2018-04-14 09:37:49 -07:00
}
2019-04-30 10:26:03 -03:00
fn require_class_wrap(&mut self, name: &str) {
require_class(&mut self.exported_classes, name).wrap_needed = true;
}
fn add_module_import(&mut self, module: String, name: &str, actual: &str) {
let rename = if name == actual {
None
} else {
Some(actual.to_string())
};
self.js_imports
.entry(module)
.or_insert(Vec::new())
.push((name.to_string(), rename));
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn import_name(&mut self, import: &JsImport) -> Result<String, Error> {
if let Some(name) = self.imported_names.get(&import.name) {
let mut name = name.clone();
for field in import.fields.iter() {
name.push_str(".");
name.push_str(field);
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
return Ok(name.clone());
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
let mut name = match &import.name {
JsImportName::Module { module, name } => {
let unique_name = self.generate_identifier(name);
self.add_module_import(module.clone(), name, &unique_name);
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
unique_name
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
JsImportName::LocalModule { module, name } => {
let unique_name = self.generate_identifier(name);
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
let module = self.config.local_module_name(module);
self.add_module_import(module, name, &unique_name);
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
unique_name
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
JsImportName::InlineJs {
unique_crate_identifier,
snippet_idx_in_crate,
name,
} => {
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
let module = self
.config
.inline_js_module_name(unique_crate_identifier, *snippet_idx_in_crate);
let unique_name = self.generate_identifier(name);
self.add_module_import(module, name, &unique_name);
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
unique_name
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
JsImportName::VendorPrefixed { name, prefixes } => {
self.imports_post.push_str("const l");
self.imports_post.push_str(&name);
self.imports_post.push_str(" = ");
switch(&mut self.imports_post, name, "", prefixes);
self.imports_post.push_str(";\n");
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn switch(dst: &mut String, name: &str, prefix: &str, left: &[String]) {
if left.len() == 0 {
dst.push_str(prefix);
return dst.push_str(name);
}
dst.push_str("(typeof ");
dst.push_str(prefix);
dst.push_str(name);
dst.push_str(" !== 'undefined' ? ");
dst.push_str(prefix);
dst.push_str(name);
dst.push_str(" : ");
switch(dst, name, &left[0], &left[1..]);
dst.push_str(")");
}
format!("l{}", name)
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
JsImportName::Global { name } => {
let unique_name = self.generate_identifier(name);
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
if unique_name != *name {
bail!("cannot import `{}` from two locations", name);
}
unique_name
Migrate `wasm-bindgen` to using `walrus` This commit moves `wasm-bindgen` the CLI tool from internally using `parity-wasm` for wasm parsing/serialization to instead use `walrus`. The `walrus` crate is something we've been working on recently with an aim to replace the usage of `parity-wasm` in `wasm-bindgen` to make the current CLI tool more maintainable as well as more future-proof. The `walrus` crate provides a much nicer AST to work with as well as a structured `Module`, whereas `parity-wasm` provides a very raw interface to the wasm module which isn't really appropriate for our use case. The many transformations and tweaks that wasm-bindgen does have a huge amount of ad-hoc index management to carefully craft a final wasm binary, but this is all entirely taken care for us with the `walrus` crate. Additionally, `wasm-bindgen` will ingest and rewrite the wasm file, often changing the binary offsets of functions. Eventually with DWARF debug information we'll need to be sure to preserve the debug information throughout the transformations that `wasm-bindgen` does today. This is practically impossible to do with the `parity-wasm` architecture, but `walrus` was designed from the get-go to solve this problem transparently in the `walrus` crate itself. (it doesn't today, but this is planned work) It is the intention that this does not end up regressing any `wasm-bindgen` use cases, neither in functionality or in speed. As a large change and refactoring, however, it's likely that at least something will arise! We'll want to continue to remain vigilant to any issues that come up with this commit. Note that the `gc` crate has been deleted as part of this change, as the `gc` crate is no longer necessary since `walrus` does it automatically. Additionally the `gc` crate was one of the main problems with preserving debug information as it often deletes wasm items! Finally, this also starts moving crates to the 2018 edition where necessary since `walrus` requires the 2018 edition, and in general it's more pleasant to work within the 2018 edition!
2019-01-31 09:54:23 -08:00
}
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
};
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.imported_names
.insert(import.name.clone(), name.clone());
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
// After we've got an actual name handle field projections
for field in import.fields.iter() {
name.push_str(".");
name.push_str(field);
}
Ok(name)
}
/// If a start function is present, it removes it from the `start` section
/// of the wasm module and then moves it to an exported function, named
/// `__wbindgen_start`.
fn unstart_start_function(&mut self) -> bool {
Migrate `wasm-bindgen` to using `walrus` This commit moves `wasm-bindgen` the CLI tool from internally using `parity-wasm` for wasm parsing/serialization to instead use `walrus`. The `walrus` crate is something we've been working on recently with an aim to replace the usage of `parity-wasm` in `wasm-bindgen` to make the current CLI tool more maintainable as well as more future-proof. The `walrus` crate provides a much nicer AST to work with as well as a structured `Module`, whereas `parity-wasm` provides a very raw interface to the wasm module which isn't really appropriate for our use case. The many transformations and tweaks that wasm-bindgen does have a huge amount of ad-hoc index management to carefully craft a final wasm binary, but this is all entirely taken care for us with the `walrus` crate. Additionally, `wasm-bindgen` will ingest and rewrite the wasm file, often changing the binary offsets of functions. Eventually with DWARF debug information we'll need to be sure to preserve the debug information throughout the transformations that `wasm-bindgen` does today. This is practically impossible to do with the `parity-wasm` architecture, but `walrus` was designed from the get-go to solve this problem transparently in the `walrus` crate itself. (it doesn't today, but this is planned work) It is the intention that this does not end up regressing any `wasm-bindgen` use cases, neither in functionality or in speed. As a large change and refactoring, however, it's likely that at least something will arise! We'll want to continue to remain vigilant to any issues that come up with this commit. Note that the `gc` crate has been deleted as part of this change, as the `gc` crate is no longer necessary since `walrus` does it automatically. Additionally the `gc` crate was one of the main problems with preserving debug information as it often deletes wasm items! Finally, this also starts moving crates to the 2018 edition where necessary since `walrus` requires the 2018 edition, and in general it's more pleasant to work within the 2018 edition!
2019-01-31 09:54:23 -08:00
let start = match self.module.start.take() {
Some(id) => id,
None => return false,
};
self.module.exports.add("__wbindgen_start", start);
true
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn expose_add_to_anyref_table(
&mut self,
table: TableId,
alloc: FunctionId,
) -> Result<MemView, Error> {
let view = self.memview_table("addToAnyrefTable", table);
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
assert!(self.config.anyref);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
if !self.should_write_global(view.to_string()) {
return Ok(view);
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let alloc = self.export_name_of(alloc);
let table = self.export_name_of(table);
self.global(&format!(
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
"
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
function {}(obj) {{
const idx = wasm.{}();
wasm.{}.set(idx, obj);
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
return idx;
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}}
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
",
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
view, alloc, table,
));
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
Ok(view)
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
pub fn generate(&mut self) -> Result<(), Error> {
self.prestore_global_import_identifiers()?;
Add tests for the interface types output of wasm-bindgen (#1898) * Add tests for the interface types output of wasm-bindgen This commit expands the test suite with assertions about the output of the interface types pass in wasm-bindgen. The goal here is to actually assert that we produce the right output and have a suite of reference files to show how the interface types output is changing over time. The `reference` test suite added in the previous PR has been updated to work for interface types as well, generating `*.wit` file assertions which are printed via the `wit-printer` crate on crates.io. Along the way a number of bugs were fixed with the interface types output, such as: * Non-determinism in output caused by iteration of a `HashMap` * Avoiding JS generation entirely in interface types mode, ensuring that we don't export extraneous intrinsics that aren't otherwise needed. * Fixing location of the stack pointer for modules where it's GC'd out. It's now rooted in the aux section of wasm-bindgen so it's available to later passes, like the multi-value pass. * Interface types emission now works in debug mode, meaning the `--release` flag is no longer required. This previously did not work because the `__wbindgen_throw` intrinsic was required in debug mode. This comes about because of the `malloc_failure` and `internal_error` functions in the anyref pass. The purpose of these functions is to signal fatal runtime errors, if any, in a way that's usable to the user. For wasm interface types though we can replace calls to these functions with `unreachable` to avoid needing to import the intrinsic. This has the accidental side effect of making `wasm_bindgen::throw_str` "just work" with wasm interface types by aborting the program, but that's not actually entirely intended. It's hoped that a split of a `wasm-bindgen-core` crate would solve this issue for the future. * Run the wasm interface types validator in tests * Add more gc roots for adapter gc * Improve stack pointer detection The stack pointer is never initialized to zero, but some other mutable globals are (TLS, thread ID, etc), so let's filter those out.
2019-12-04 15:19:48 -06:00
for (id, adapter) in crate::sorted_iter(&self.wit.adapters) {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let instrs = match &adapter.kind {
AdapterKind::Import { .. } => continue,
AdapterKind::Local { instructions } => instructions,
};
self.generate_adapter(*id, adapter, instrs)?;
Add experimental support for the `anyref` type This commit adds experimental support to `wasm-bindgen` to emit and leverage the `anyref` native wasm type. This native type is still in a proposal status (the reference-types proposal). The intention of `anyref` is to be able to directly hold JS values in wasm and pass the to imported functions, namely to empower eventual host bindings (now renamed WebIDL bindings) integration where we can skip JS shims altogether for many imports. This commit doesn't actually affect wasm-bindgen's behavior at all as-is, but rather this support requires an opt-in env var to be configured. Once the support is stable in browsers it's intended that this will add a CLI switch for turning on this support, eventually defaulting it to `true` in the far future. The basic strategy here is to take the `stack` and `slab` globals in the generated JS glue and move them into wasm using a table. This new table in wasm is managed at the fringes via injected shims. At `wasm-bindgen`-time the CLI will rewrite exports and imports with shims that actually use `anyref` if needed, performing loads/stores inside the wasm module instead of externally in the wasm module. This should provide a boost over what we have today, but it's not a fantastic strategy long term. We have a more grand vision for `anyref` being a first-class type in the language, but that's on a much longer horizon and this is currently thought to be the best we can do in terms of integration in the near future. The stack/heap JS tables are combined into one wasm table. The stack starts at the end of the table and grows down with a stack pointer (also injected). The heap starts at the end and grows up (state managed in linear memory). The anyref transformation here will hook up various intrinsics in wasm-bindgen to the runtime functionality if the anyref supoprt is enabled. The main tricky treatment here was applied to closures, where we need JS to use a different function pointer than the one Rust gives it to use a JS function pointer empowered with anyref. This works by switching up a bit how descriptors work, embedding the shims to call inside descriptors rather than communicated at runtime. This means that we're accessing constant values in the generated JS and we can just update the constant value accessed.
2018-10-18 08:43:36 -07:00
}
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
let mut pairs = self.aux.export_map.iter().collect::<Vec<_>>();
pairs.sort_by_key(|(k, _)| *k);
check_duplicated_getter_and_setter_names(&pairs)?;
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
for e in self.aux.enums.iter() {
self.generate_enum(e)?;
2018-02-22 12:01:38 +01:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
for s in self.aux.structs.iter() {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.generate_struct(s)?;
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
self.typescript.push_str(&self.aux.extra_typescript);
Rewrite wasm-bindgen with ES6 modules in mind This commit is a mostly-rewrite of the `wasm-bindgen` tool. After some recent discussions it's clear that the previous model wasn't quite going to cut it, and this iteration is one which primarily embraces ES6 modules and the idea that this is a polyfill for host bindings. The overall interface and functionality hasn't changed much but the underlying technology has now changed significantly. Previously `wasm-bindgen` would emit a JS file that acted as an ES6 module but had a bit of a wonky interface. It exposed an async function for instantiation of the wasm module, but that's the bundler's job, not ours! Instead this iteration views each input and output as a discrete ES6 module. The input wasm file is interpreted as "this *should* be an ES6 module with rich types" and the output is "well here's some ES6 modules that fulfill that contract". Notably the tool now replaces the original wasm ES6 module with a JS ES6 module that has the "rich interface". Additionally a second ES6 module is emitted (the actual wasm file) which imports and exports to the original ES6 module. This strategy is hoped to be much more amenable to bundlers and controlling how the wasm itself is instantiated. The emitted files files purely assume ES6 modules and should be able to work as-is once ES6 module integration for wasm is completed. Note that there aren't a ton of tools to pretend a wasm module is an ES6 module at the moment but those should be coming soon! In the meantime a local `wasm2es6js` hack was added to help make *something* work today. The README has also been updated with instructions for interacting with this model.
2018-01-29 21:20:38 -08:00
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
for path in self.aux.package_jsons.iter() {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.process_package_json(path)?;
}
Ok(())
}
/// Registers import names for all `Global` imports first before we actually
/// process any adapters.
///
/// `Global` names must be imported as their exact name, so if the same name
/// from a global is also imported from a module we have to be sure to
/// import the global first to ensure we don't shadow the actual global
/// value. Otherwise we have no way of accessing the global value!
///
/// This function will iterate through the import map up-front and generate
/// a cache entry for each import name which is a `Global`.
fn prestore_global_import_identifiers(&mut self) -> Result<(), Error> {
for import in self.aux.import_map.values() {
let js = match import {
2020-04-20 22:10:33 +08:00
AuxImport::Value(AuxValue::Bare(js))
| AuxImport::Value(AuxValue::ClassGetter(js, ..))
| AuxImport::Value(AuxValue::Getter(js, ..))
| AuxImport::Value(AuxValue::ClassSetter(js, ..))
| AuxImport::Value(AuxValue::Setter(js, ..))
| AuxImport::ValueWithThis(js, ..)
| AuxImport::Instanceof(js)
| AuxImport::Static(js)
| AuxImport::StructuralClassGetter(js, ..)
| AuxImport::StructuralClassSetter(js, ..)
| AuxImport::IndexingGetterOfClass(js)
| AuxImport::IndexingSetterOfClass(js)
| AuxImport::IndexingDeleterOfClass(js) => js,
_ => continue,
};
if let JsImportName::Global { .. } = js.name {
self.import_name(js)?;
}
}
Ok(())
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn generate_adapter(
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
&mut self,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
id: AdapterId,
adapter: &Adapter,
instrs: &[InstructionData],
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
) -> Result<(), Error> {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
enum Kind<'a> {
Export(&'a AuxExport),
Import(walrus::ImportId),
Adapter,
}
let kind = match self.aux.export_map.get(&id) {
Some(export) => Kind::Export(export),
None => {
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
let core = self.wit.implements.iter().find(|pair| pair.2 == id);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
match core {
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
Some((core, _, _)) => Kind::Import(*core),
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
None => Kind::Adapter,
}
}
};
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let catch = self.aux.imports_with_catch.contains(&id);
if let Kind::Import(core) = kind {
if !catch && self.attempt_direct_import(core, instrs)? {
return Ok(());
}
}
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
// Construct a JS shim builder, and configure it based on the kind of
// export that we're generating.
let mut builder = binding::Builder::new(self);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
builder.log_error(match kind {
Kind::Export(_) | Kind::Adapter => false,
Kind::Import(_) => builder.cx.config.debug,
});
builder.catch(catch);
let mut arg_names = &None;
match kind {
Kind::Export(export) => {
arg_names = &export.arg_names;
match &export.kind {
AuxExportKind::Function(_) => {}
AuxExportKind::StaticFunction { .. } => {}
AuxExportKind::Constructor(class) => builder.constructor(class),
AuxExportKind::Getter { .. } | AuxExportKind::Setter { .. } => {
builder.method(false)
}
AuxExportKind::Method { consumed, .. } => builder.method(*consumed),
}
}
Kind::Import(_) => {}
Kind::Adapter => {}
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
// Process the `binding` and generate a bunch of JS/TypeScript/etc.
let binding::JsFunction {
ts_sig,
ts_arg_tys,
ts_ret_ty,
js_doc,
code,
might_be_optional_field,
catch,
log_error,
} = builder
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
.process(&adapter, instrs, arg_names)
.with_context(|| match kind {
Kind::Export(e) => format!("failed to generate bindings for `{}`", e.debug_name),
Kind::Import(i) => {
let i = builder.cx.module.imports.get(i);
format!(
"failed to generate bindings for import of `{}::{}`",
i.module, i.name
)
}
Kind::Adapter => format!("failed to generates bindings for adapter"),
})?;
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
2019-07-10 09:44:04 -07:00
// Once we've got all the JS then put it in the right location depending
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
// on what's being exported.
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
match kind {
Kind::Export(export) => {
assert_eq!(catch, false);
assert_eq!(log_error, false);
let ts_sig = match export.generate_typescript {
true => Some(ts_sig.as_str()),
false => None,
};
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let docs = format_doc_comments(&export.comments, Some(js_doc));
match &export.kind {
AuxExportKind::Function(name) => {
if let Some(ts_sig) = ts_sig {
self.typescript.push_str(&docs);
self.typescript.push_str("export function ");
self.typescript.push_str(&name);
self.typescript.push_str(ts_sig);
self.typescript.push_str(";\n");
}
self.export(&name, &format!("function{}", code), Some(&docs))?;
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
self.globals.push_str("\n");
}
AuxExportKind::Constructor(class) => {
let exported = require_class(&mut self.exported_classes, class);
if exported.has_constructor {
bail!("found duplicate constructor for class `{}`", class);
}
exported.has_constructor = true;
exported.push(&docs, "constructor", "", &code, ts_sig);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}
AuxExportKind::Getter { class, field } => {
let ret_ty = match export.generate_typescript {
true => match &ts_ret_ty {
Some(s) => Some(s.as_str()),
_ => None,
},
false => None,
};
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let exported = require_class(&mut self.exported_classes, class);
exported.push_getter(&docs, field, &code, ret_ty);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}
AuxExportKind::Setter { class, field } => {
let arg_ty = match export.generate_typescript {
true => Some(ts_arg_tys[0].as_str()),
false => None,
};
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let exported = require_class(&mut self.exported_classes, class);
exported.push_setter(&docs, field, &code, arg_ty, might_be_optional_field);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}
AuxExportKind::StaticFunction { class, name } => {
let exported = require_class(&mut self.exported_classes, class);
exported.push(&docs, name, "static ", &code, ts_sig);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}
AuxExportKind::Method { class, name, .. } => {
let exported = require_class(&mut self.exported_classes, class);
exported.push(&docs, name, "", &code, ts_sig);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
}
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
Kind::Import(core) => {
let code = if catch {
format!("handleError(function{})", code)
} else if log_error {
format!("logError(function{})", code)
} else {
format!("function{}", code)
};
self.wasm_import_definitions.insert(core, code);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}
Kind::Adapter => {
assert_eq!(catch, false);
assert_eq!(log_error, false);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
self.globals.push_str("function ");
self.globals.push_str(&self.adapter_name(id));
self.globals.push_str(&code);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
self.globals.push_str("\n\n");
}
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
return Ok(());
}
/// Returns whether we should disable the logic, in debug mode, to catch an
/// error, log it, and rethrow it. This is only intended for user-defined
/// imports, not all imports of everything.
fn import_never_log_error(&self, import: &AuxImport) -> bool {
match import {
// Some intrinsics are intended to exactly throw errors, and in
// general we shouldn't have exceptions in our intrinsics to debug,
// so skip these.
AuxImport::Intrinsic(_) => true,
// Otherwise assume everything else gets a debug log of errors
// thrown in debug mode.
_ => false,
}
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
/// Attempts to directly hook up the `id` import in the wasm module with
/// the `instrs` specified.
///
/// If this succeeds it returns `Ok(true)`, otherwise if it cannot be
/// directly imported then `Ok(false)` is returned.
fn attempt_direct_import(
&mut self,
id: ImportId,
instrs: &[InstructionData],
) -> Result<bool, Error> {
// First up extract the ID of the single called adapter, if any.
let mut call = None;
for instr in instrs {
match instr.instr {
Instruction::CallAdapter(id) => {
if call.is_some() {
return Ok(false);
} else {
call = Some(id);
}
}
Instruction::CallExport(_)
| Instruction::CallTableElement(_)
| Instruction::Standard(wit_walrus::Instruction::CallCore(_))
| Instruction::Standard(wit_walrus::Instruction::CallAdapter(_)) => {
return Ok(false)
}
_ => {}
}
}
let adapter = match call {
Some(id) => id,
None => return Ok(false),
};
match &self.wit.adapters[&adapter].kind {
AdapterKind::Import { kind, .. } => match kind {
AdapterJsImportKind::Normal => {}
// method/constructors need glue because we either need to
// invoke them as `new` or we need to invoke them with
// method-call syntax to get the `this` parameter right.
AdapterJsImportKind::Method | AdapterJsImportKind::Constructor => return Ok(false),
},
// This is an adapter-to-adapter call, so it needs a shim.
AdapterKind::Local { .. } => return Ok(false),
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
// Next up check to make sure that this import is to a bare JS value
// itself, no extra fluff intended.
let js = match &self.aux.import_map[&adapter] {
AuxImport::Value(AuxValue::Bare(js)) => js,
_ => return Ok(false),
};
// Make sure this isn't variadic in any way which means we need some
// sort of adapter glue.
if self.aux.imports_with_variadic.contains(&adapter) {
return Ok(false);
}
// Ensure that every single instruction can be represented without JS
// glue being generated, aka it's covered by the JS ECMAScript bindings
// for wasm.
if !self.representable_without_js_glue(instrs) {
return Ok(false);
}
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
// If there's no field projection happening here and this is a direct
// import from an ES-looking module, then we can actually just hook this
// up directly in the wasm file itself. Note that this is covered in the
// various output formats as well:
//
// * `bundler` - they think wasm is an ES module anyway
// * `web` - we're sure to emit more `import` directives during
// `gen_init` and we update the import object accordingly.
// * `nodejs` - the polyfill we have for requiring a wasm file as a node
// module will naturally emit `require` directives for the module
// listed on each wasm import.
// * `no-modules` - imports aren't allowed here anyway from other
// modules and an error is generated.
if js.fields.len() == 0 {
match &js.name {
JsImportName::Module { module, name } => {
let import = self.module.imports.get_mut(id);
import.module = module.clone();
import.name = name.clone();
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
return Ok(true);
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
}
JsImportName::LocalModule { module, name } => {
let module = self.config.local_module_name(module);
let import = self.module.imports.get_mut(id);
import.module = module;
import.name = name.clone();
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
return Ok(true);
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
}
JsImportName::InlineJs {
unique_crate_identifier,
snippet_idx_in_crate,
name,
} => {
let module = self
.config
.inline_js_module_name(unique_crate_identifier, *snippet_idx_in_crate);
let import = self.module.imports.get_mut(id);
import.module = module;
import.name = name.clone();
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
return Ok(true);
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
}
// Fall through below to requiring a JS shim to create an item
// that we can import. These are plucked from the global
// environment so there's no way right now to describe these
// imports in an ES module-like fashion.
JsImportName::Global { .. } | JsImportName::VendorPrefixed { .. } => {}
}
}
self.expose_not_defined();
let name = self.import_name(js)?;
let js = format!(
"typeof {name} == 'function' ? {name} : notDefined('{name}')",
name = name,
);
self.wasm_import_definitions.insert(id, js);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
Ok(true)
}
fn representable_without_js_glue(&self, instrs: &[InstructionData]) -> bool {
use Instruction::*;
let standard_enabled = self.config.wasm_interface_types;
let mut last_arg = None;
let mut saw_call = false;
for instr in instrs {
match instr.instr {
// Is an adapter section getting emitted? If so, then every
// standard operation is natively supported!
Standard(_) if standard_enabled => {}
// Fetching arguments is just that, a fetch, so no need for
// glue. Note though that the arguments must be fetched in order
// for this to actually work,
Standard(wit_walrus::Instruction::ArgGet(i)) => {
if saw_call {
return false;
}
match (i, last_arg) {
(0, None) => last_arg = Some(0),
(n, Some(i)) if n == i + 1 => last_arg = Some(n),
_ => return false,
}
}
// Similarly calling a function is the same as in JS, no glue
// needed.
CallAdapter(_) => saw_call = true,
// Conversions to wasm integers are always supported since
// they're coerced into i32/f32/f64 appropriately.
Standard(wit_walrus::Instruction::IntToWasm { .. }) => {}
// Converts from wasm to JS, however, only supports most
// integers. Converting into a u32 isn't supported because we
// need to generate glue to change the sign.
Standard(wit_walrus::Instruction::WasmToInt {
output: wit_walrus::ValType::U32,
..
}) => return false,
Standard(wit_walrus::Instruction::WasmToInt { .. }) => {}
// JS spec automatically coerces boolean values to i32 of 0 or 1
// depending on true/false
I32FromBool => {}
_ => return false,
}
}
return true;
Support emitting direct imports in wasm files Support was previously (re-)added in #1654 for importing direct JS values into a WebAssembly module by completely skipping JS shim generation. This commit takes that PR one step further by *also* embedding a direct import in the wasm file, where supported. The wasm file currently largely just imports from the JS shim file that we generate, but this allows it to directly improt from ES modules where supported and where possible. Note that like #1654 this only happens when the function signature doesn't actually require any conversions to happen in JS (such as handling closures). For imports from ES modules, local snippets, or inline JS they'll all have their import directives directly embedded into the final WebAssembly binary without any shims necessary to hook it all up. For imports from the global namespace or possibly vendor-prefixed items these still unconditionally require an import shim to be generated because there's no way to describe that import in an ES-friendly way (yet). There's a few consequences of this commit which are also worth noting: * The logic in `wasm-bindgen` where it gracefully handles (to some degree) not-defined items now only is guaranteed to be applied to the global namespace. If you import from a module, it'll be an instantiation time error rather than today's runtime error when the import is called. * Handling imports in the wasm module not registered with `#[wasm_bindgen]` has become more strict. Previously these imports were basically ignored, leaving them up for interpretation depending on the output format. The changes for each output target are: * `bundler` - not much has changed here. Previously these ignored imports would have been treated as ES module imports, and after this commit there might just be some more of these imports for bundlers to resolve. * `web` - previously the ignored imports would likely cause instantiation failures because the import object never actually included a binding for other imports. After this commit though the JS glue which instantiates the module now interprets all unrecognized wasm module imports as ES module imports, emitting an `import` directive. This matches what we want for the direct import functionality, and is also largely what we want for modules in general. * `nodejs` - previously ignored imports were handled in the translation shim for Node to generate `require` statements, so they were actually "correctly handled" sort of with module imports. The handling of this hasn't changed, and reflects what we want for direct imports of values where loading a wasm module in Node ends up translating the module field of each import to a `require`. * `no-modules` - this is very similar to the `web` target where previously this didn't really work one way or the other because we'd never fill in more fields of the import object when instantiating the module. After this PR though this is a hard-error to have unrecognized imports from `#[wasm_bindgen]` with the `no-modules` output type, because we don't know how to handle the imports. Note that this touches on #1584 and will likely break the current use case being mentioned there. I think though that this tightening up of how we handle imports is what we'll want in the long run where everything is interpreted as modules, and we'll need to figure out best how wasi fits into this. This commit is unlikely to have any real major immediate effects. The goal here is to continue to inch us towards a world where there's less and less JS glue necessary and `wasm-bindgen` is just a polyfill for web standards that otherwise all already exist. Also note that there's no explicitly added tests for this since this is largely just a refactoring of an internal implementation detail of `wasm-bindgen`, but the main `wasm` test suite has many instances of this path being taken, for example having imports like: (import "tests/wasm/duplicates_a.js" "foo" (func $__wbg_foo_969c253238f136f0 (type 1))) (import "tests/wasm/duplicates_b.js" "foo" (func $__wbg_foo_027958cb2e320a94 (type 0))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "trivial" (func $__wbg_trivial_75e27c84882af23b (type 1))) (import "./snippets/wasm-bindgen-3dff2bc911f0a20c/inline0.js" "incoming_bool" (func $__wbg_incomingbool_0f2d9f55f73a256f (type 0)))
2019-07-31 11:35:36 -07:00
}
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
/// Generates a JS snippet appropriate for invoking `import`.
///
/// This is generating code for `binding` where `bindings` has more type
/// infomation. The `args` array is the list of JS expressions representing
/// the arguments to pass to JS. Finally `variadic` indicates whether the
/// last argument is a list to be splatted in a variadic way, and `prelude`
/// is a location to push some more initialization JS if necessary.
///
/// The returned value here is a JS expression which evaluates to the
/// purpose of `AuxImport`, which depends on the kind of import.
fn invoke_import(
&mut self,
import: &AuxImport,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
kind: AdapterJsImportKind,
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
args: &[String],
variadic: bool,
prelude: &mut String,
) -> Result<String, Error> {
let variadic_args = |js_arguments: &[String]| {
Ok(if !variadic {
format!("{}", js_arguments.join(", "))
} else {
let (last_arg, args) = match js_arguments.split_last() {
Some(pair) => pair,
None => bail!("a function with no arguments cannot be variadic"),
};
if args.len() > 0 {
format!("{}, ...{}", args.join(", "), last_arg)
} else {
format!("...{}", last_arg)
}
})
};
match import {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
AuxImport::Value(val) => match kind {
AdapterJsImportKind::Constructor => {
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
let js = match val {
AuxValue::Bare(js) => self.import_name(js)?,
_ => bail!("invalid import set for constructor"),
};
Ok(format!("new {}({})", js, variadic_args(&args)?))
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
AdapterJsImportKind::Method => {
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
let descriptor = |anchor: &str, extra: &str, field: &str, which: &str| {
format!(
"GetOwnOrInheritedPropertyDescriptor({}{}, '{}').{}",
anchor, extra, field, which
)
};
let js = match val {
AuxValue::Bare(js) => self.import_name(js)?,
AuxValue::Getter(class, field) => {
self.expose_get_inherited_descriptor();
let class = self.import_name(class)?;
descriptor(&class, ".prototype", field, "get")
}
AuxValue::ClassGetter(class, field) => {
self.expose_get_inherited_descriptor();
let class = self.import_name(class)?;
descriptor(&class, "", field, "get")
}
AuxValue::Setter(class, field) => {
self.expose_get_inherited_descriptor();
let class = self.import_name(class)?;
descriptor(&class, ".prototype", field, "set")
}
AuxValue::ClassSetter(class, field) => {
self.expose_get_inherited_descriptor();
let class = self.import_name(class)?;
descriptor(&class, "", field, "set")
}
};
Ok(format!("{}.call({})", js, variadic_args(&args)?))
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
AdapterJsImportKind::Normal => {
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
let js = match val {
AuxValue::Bare(js) => self.import_name(js)?,
_ => bail!("invalid import set for free function"),
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
};
Ok(format!("{}({})", js, variadic_args(&args)?))
}
},
AuxImport::ValueWithThis(class, name) => {
let class = self.import_name(class)?;
Ok(format!("{}.{}({})", class, name, variadic_args(&args)?))
}
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
AuxImport::Instanceof(js) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 1);
let js = self.import_name(js)?;
Ok(format!("{} instanceof {}", args[0], js))
}
AuxImport::Static(js) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 0);
self.import_name(js)
}
AuxImport::Closure {
dtor,
mutable,
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
adapter,
nargs: _,
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
} => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 3);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let call = self.adapter_name(*adapter);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
if *mutable {
self.expose_make_mut_closure()?;
Ok(format!(
"makeMutClosure({arg0}, {arg1}, {dtor}, {call})",
arg0 = &args[0],
arg1 = &args[1],
dtor = dtor,
call = call,
))
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
} else {
self.expose_make_closure()?;
Ok(format!(
"makeClosure({arg0}, {arg1}, {dtor}, {call})",
arg0 = &args[0],
arg1 = &args[1],
dtor = dtor,
call = call,
))
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
}
AuxImport::StructuralMethod(name) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
let (receiver, args) = match args.split_first() {
Some(pair) => pair,
None => bail!("structural method calls must have at least one argument"),
};
Ok(format!("{}.{}({})", receiver, name, variadic_args(args)?))
}
AuxImport::StructuralGetter(field) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 1);
Ok(format!("{}.{}", args[0], field))
}
AuxImport::StructuralClassGetter(class, field) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 0);
let class = self.import_name(class)?;
Ok(format!("{}.{}", class, field))
}
AuxImport::StructuralSetter(field) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 2);
Ok(format!("{}.{} = {}", args[0], field, args[1]))
}
AuxImport::StructuralClassSetter(class, field) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 1);
let class = self.import_name(class)?;
Ok(format!("{}.{} = {}", class, field, args[0]))
}
AuxImport::IndexingGetterOfClass(class) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 1);
let class = self.import_name(class)?;
Ok(format!("{}[{}]", class, args[0]))
}
AuxImport::IndexingGetterOfObject => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 2);
Ok(format!("{}[{}]", args[0], args[1]))
}
AuxImport::IndexingSetterOfClass(class) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 2);
let class = self.import_name(class)?;
Ok(format!("{}[{}] = {}", class, args[0], args[1]))
}
AuxImport::IndexingSetterOfObject => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 3);
Ok(format!("{}[{}] = {}", args[0], args[1], args[2]))
}
AuxImport::IndexingDeleterOfClass(class) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 1);
let class = self.import_name(class)?;
Ok(format!("delete {}[{}]", class, args[0]))
}
AuxImport::IndexingDeleterOfObject => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 2);
Ok(format!("delete {}[{}]", args[0], args[1]))
}
AuxImport::WrapInExportedClass(class) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
assert_eq!(args.len(), 1);
self.require_class_wrap(class);
Ok(format!("{}.__wrap({})", class, args[0]))
}
AuxImport::Intrinsic(intrinsic) => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert!(kind == AdapterJsImportKind::Normal);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
assert!(!variadic);
self.invoke_intrinsic(intrinsic, args, prelude)
}
}
}
/// Same as `invoke_import` above, except more specialized and only used for
/// generating the JS expression needed to implement a particular intrinsic.
fn invoke_intrinsic(
&mut self,
intrinsic: &Intrinsic,
args: &[String],
prelude: &mut String,
) -> Result<String, Error> {
let expr = match intrinsic {
Intrinsic::JsvalEq => {
assert_eq!(args.len(), 2);
format!("{} === {}", args[0], args[1])
}
Intrinsic::IsFunction => {
assert_eq!(args.len(), 1);
format!("typeof({}) === 'function'", args[0])
}
Intrinsic::IsUndefined => {
assert_eq!(args.len(), 1);
format!("{} === undefined", args[0])
}
Intrinsic::IsNull => {
assert_eq!(args.len(), 1);
format!("{} === null", args[0])
}
Intrinsic::IsObject => {
assert_eq!(args.len(), 1);
prelude.push_str(&format!("const val = {};\n", args[0]));
format!("typeof(val) === 'object' && val !== null")
}
Intrinsic::IsSymbol => {
assert_eq!(args.len(), 1);
format!("typeof({}) === 'symbol'", args[0])
}
Intrinsic::IsString => {
assert_eq!(args.len(), 1);
format!("typeof({}) === 'string'", args[0])
}
2019-07-02 18:44:06 +02:00
Intrinsic::IsFalsy => {
assert_eq!(args.len(), 1);
format!("!{}", args[0])
}
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
Intrinsic::ObjectCloneRef => {
assert_eq!(args.len(), 1);
args[0].clone()
}
Intrinsic::ObjectDropRef => {
assert_eq!(args.len(), 1);
args[0].clone()
}
Intrinsic::CallbackDrop => {
assert_eq!(args.len(), 1);
prelude.push_str(&format!("const obj = {}.original;\n", args[0]));
prelude.push_str("if (obj.cnt-- == 1) {\n");
prelude.push_str("obj.a = 0;\n");
prelude.push_str("return true;\n");
prelude.push_str("}\n");
"false".to_string()
}
Intrinsic::CallbackForget => {
assert_eq!(args.len(), 1);
args[0].clone()
}
Intrinsic::NumberNew => {
assert_eq!(args.len(), 1);
args[0].clone()
}
Intrinsic::StringNew => {
assert_eq!(args.len(), 1);
args[0].clone()
}
Intrinsic::SymbolNamedNew => {
assert_eq!(args.len(), 1);
format!("Symbol({})", args[0])
}
Intrinsic::SymbolAnonymousNew => {
assert_eq!(args.len(), 0);
"Symbol()".to_string()
}
Intrinsic::NumberGet => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert_eq!(args.len(), 1);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
prelude.push_str(&format!("const obj = {};\n", args[0]));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
format!("typeof(obj) === 'number' ? obj : undefined")
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
Intrinsic::StringGet => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
assert_eq!(args.len(), 1);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
prelude.push_str(&format!("const obj = {};\n", args[0]));
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
format!("typeof(obj) === 'string' ? obj : undefined")
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
Intrinsic::BooleanGet => {
assert_eq!(args.len(), 1);
prelude.push_str(&format!("const v = {};\n", args[0]));
format!("typeof(v) === 'boolean' ? (v ? 1 : 0) : 2")
}
Intrinsic::Throw => {
assert_eq!(args.len(), 1);
format!("throw new Error({})", args[0])
}
Intrinsic::Rethrow => {
assert_eq!(args.len(), 1);
format!("throw {}", args[0])
}
Intrinsic::Module => {
assert_eq!(args.len(), 0);
if !self.config.mode.no_modules() && !self.config.mode.web() {
bail!(
"`wasm_bindgen::module` is currently only supported with \
`--target no-modules` and `--target web`"
);
}
format!("init.__wbindgen_wasm_module")
}
Intrinsic::Memory => {
assert_eq!(args.len(), 0);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let mut memories = self.module.memories.iter();
let memory = memories
.next()
.ok_or_else(|| anyhow!("no memory found to return in memory intrinsic"))?
.id();
if memories.next().is_some() {
bail!(
"multiple memories found, unsure which to return \
from memory intrinsic"
);
}
drop(memories);
format!("wasm.{}", self.export_name_of(memory))
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
Intrinsic::FunctionTable => {
assert_eq!(args.len(), 0);
let name = self.export_function_table()?;
format!("wasm.{}", name)
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
Intrinsic::DebugString => {
assert_eq!(args.len(), 1);
self.expose_debug_string();
format!("debugString({})", args[0])
}
Intrinsic::JsonParse => {
assert_eq!(args.len(), 1);
format!("JSON.parse({})", args[0])
}
Intrinsic::JsonSerialize => {
assert_eq!(args.len(), 1);
// Turns out `JSON.stringify(undefined) === undefined`, so if
// we're passed `undefined` reinterpret it as `null` for JSON
// purposes.
prelude.push_str(&format!("const obj = {};\n", args[0]));
"JSON.stringify(obj === undefined ? null : obj)".to_string()
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
Intrinsic::AnyrefHeapLiveCount => {
assert_eq!(args.len(), 0);
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
self.expose_global_heap();
prelude.push_str(
"
let free_count = 0;
let next = heap_next;
while (next < heap.length) {
free_count += 1;
next = heap[next];
}
",
);
format!(
"heap.length - free_count - {} - {}",
INITIAL_HEAP_OFFSET,
INITIAL_HEAP_VALUES.len(),
)
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
Intrinsic::InitAnyrefTable => {
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
let table = self
.aux
.anyref_table
.ok_or_else(|| anyhow!("must enable anyref to use anyref intrinsic"))?;
let name = self.export_name_of(table);
// Grow the table to insert our initial values, and then also
// set the 0th slot to `undefined` since that's what we've
// historically used for our ABI which is that the index of 0
// returns `undefined` for types like `None` going out.
let mut base = format!(
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
"
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
const table = wasm.{};
const offset = table.grow({});
table.set(0, undefined);
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
",
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
name,
INITIAL_HEAP_VALUES.len(),
);
for (i, value) in INITIAL_HEAP_VALUES.iter().enumerate() {
base.push_str(&format!("table.set(offset + {}, {});\n", i, value));
}
base
Second large refactor for WebIDL bindings This commit is the second, and hopefully last massive, refactor for using WebIDL bindings internally in `wasm-bindgen`. This commit actually fully executes on the task at hand, moving `wasm-bindgen` to internally using WebIDL bindings throughout its code generation, anyref passes, etc. This actually fixes a number of issues that have existed in the anyref pass for some time now! The main changes here are to basically remove the usage of `Descriptor` from generating JS bindings. Instead two new types are introduced: `NonstandardIncoming` and `NonstandardOutgoing` which are bindings lists used for incoming/outgoing bindings. These mirror the standard terminology and literally have variants which are the standard values. All `Descriptor` types are now mapped into lists of incoming/outgoing bindings and used for process in wasm-bindgen. All JS generation has been refactored and updated to now process these lists of bindings instead of the previous `Descriptor`. In other words this commit takes `js2rust.rs` and `rust2js.rs` and first splits them in two. Interpretation of `Descriptor` and what to do for conversions is in the binding selection modules. The actual generation of JS from the binding selection is now performed by `incoming.rs` and `outgoing.rs`. To boot this also deduplicates all the code between the argument handling of `js2rust.rs` and return value handling of `rust2js.rs`. This means that to implement a new binding you only need to implement it one place and it's implemented for free in the other! This commit is not the end of the story though. I would like to add a mdoe to `wasm-bindgen` that literally emits a WebIDL bindings section. That's left for a third (and hopefully final) refactoring which is also intended to optimize generated JS for bindings. This commit currently loses the optimization where an imported is hooked up by value directly whenever a shim isn't needed. It's planned that the next refactoring to emit a webidl binding section that can be added back in. It shouldn't be too too hard hopefully since all the scaffolding is in place now. cc #1524
2019-06-10 07:09:51 -07:00
}
};
Ok(expr)
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn generate_enum(&mut self, enum_: &AuxEnum) -> Result<(), Error> {
let docs = format_doc_comments(&enum_.comments, None);
2018-02-22 12:01:38 +01:00
let mut variants = String::new();
if enum_.generate_typescript {
self.typescript.push_str(&docs);
self.typescript
.push_str(&format!("export enum {} {{", enum_.name));
}
for (name, value, comments) in enum_.variants.iter() {
let variant_docs = if comments.is_empty() {
String::new()
} else {
format_doc_comments(&comments, None)
};
if !variant_docs.is_empty() {
variants.push_str("\n");
variants.push_str(&variant_docs);
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
variants.push_str(&format!("{}:{},", name, value));
if enum_.generate_typescript {
self.typescript.push_str("\n");
if !variant_docs.is_empty() {
self.typescript.push_str(&variant_docs);
}
self.typescript.push_str(&format!(" {},", name));
}
}
if enum_.generate_typescript {
self.typescript.push_str("\n}\n");
2018-02-22 12:01:38 +01:00
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.export(
2018-06-27 22:42:34 -07:00
&enum_.name,
&format!("Object.freeze({{ {} }})", variants),
Some(&docs),
)?;
2018-02-23 17:30:18 +01:00
Ok(())
2018-02-22 12:01:38 +01:00
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn generate_struct(&mut self, struct_: &AuxStruct) -> Result<(), Error> {
let class = require_class(&mut self.exported_classes, &struct_.name);
class.comments = format_doc_comments(&struct_.comments, None);
class.is_inspectable = struct_.is_inspectable;
Ok(())
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn process_package_json(&mut self, path: &Path) -> Result<(), Error> {
if self.config.mode.no_modules() {
bail!(
"NPM dependencies have been specified in `{}` but \
this is incompatible with the `no-modules` target",
2019-05-30 10:56:59 -07:00
path.display(),
);
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
let contents =
fs::read_to_string(path).context(format!("failed to read `{}`", path.display()))?;
let json: serde_json::Value = serde_json::from_str(&contents)?;
let object = match json.as_object() {
Some(s) => s,
None => bail!(
"expected `package.json` to have an JSON object in `{}`",
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
path.display()
),
};
let mut iter = object.iter();
let mut value = None;
while let Some((key, v)) = iter.next() {
if key == "dependencies" {
value = Some(v);
break;
}
}
let value = if let Some(value) = value {
value
} else {
return Ok(());
};
let value = match value.as_object() {
Some(s) => s,
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
None => bail!(
"expected `dependencies` to be a JSON object in `{}`",
path.display()
),
};
for (name, value) in value.iter() {
let value = match value.as_str() {
Some(s) => s,
None => bail!(
"keys in `dependencies` are expected to be strings in `{}`",
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
path.display()
),
};
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
if let Some((prev, _prev_version)) = self.npm_dependencies.get(name) {
bail!(
"dependency on NPM package `{}` specified in two `package.json` files, \
which at the time is not allowed:\n * {}\n * {}",
name,
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
path.display(),
prev.display(),
)
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.npm_dependencies
.insert(name.to_string(), (path.to_path_buf(), value.to_string()));
}
Ok(())
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn expose_debug_string(&mut self) {
if !self.should_write_global("debug_string") {
return;
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.global(
"
function debugString(val) {
// primitive types
const type = typeof val;
if (type == 'number' || type == 'boolean' || val == null) {
return `${val}`;
}
if (type == 'string') {
return `\"${val}\"`;
}
if (type == 'symbol') {
const description = val.description;
if (description == null) {
return 'Symbol';
} else {
return `Symbol(${description})`;
}
}
if (type == 'function') {
const name = val.name;
if (typeof name == 'string' && name.length > 0) {
return `Function(${name})`;
} else {
return 'Function';
}
}
// objects
if (Array.isArray(val)) {
const length = val.length;
let debug = '[';
if (length > 0) {
debug += debugString(val[0]);
}
for(let i = 1; i < length; i++) {
debug += ', ' + debugString(val[i]);
}
debug += ']';
return debug;
}
// Test for built-in
const builtInMatches = /\\[object ([^\\]]+)\\]/.exec(toString.call(val));
let className;
if (builtInMatches.length > 1) {
className = builtInMatches[1];
} else {
// Failed to match the standard '[object ClassName]'
return toString.call(val);
}
if (className == 'Object') {
// we're a user defined class or Object
// JSON.stringify avoids problems with cycles, and is generally much
// easier than looping through ownProperties of `val`.
try {
return 'Object(' + JSON.stringify(val) + ')';
} catch (_) {
return 'Object';
}
}
// errors
if (val instanceof Error) {
return `${val.name}: ${val.message}\\n${val.stack}`;
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
}
// TODO we could test for more things here, like `Set`s and `Map`s.
return className;
}
",
);
}
fn export_function_table(&mut self) -> Result<String, Error> {
match self.aux.function_table {
Some(id) => Ok(self.export_name_of(id)),
None => bail!("no function table found in module"),
}
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
fn export_name_of(&mut self, id: impl Into<walrus::ExportItem>) -> String {
let id = id.into();
let export = self.module.exports.iter().find(|e| {
use walrus::ExportItem::*;
match (e.item, id) {
(Function(a), Function(b)) => a == b,
(Table(a), Table(b)) => a == b,
(Memory(a), Memory(b)) => a == b,
(Global(a), Global(b)) => a == b,
_ => false,
}
});
if let Some(export) = export {
return export.name.clone();
}
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
let default_name = format!("__wbindgen_export_{}", self.next_export_idx);
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
self.next_export_idx += 1;
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
let name = match id {
walrus::ExportItem::Function(f) => match &self.module.funcs.get(f).name {
Some(s) => to_js_identifier(s),
None => default_name,
},
_ => default_name,
};
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
self.module.exports.add(&name, id);
return name;
Add reference output tests for JS operations (#1894) * Add reference output tests for JS operations This commit starts adding a test suite which checks in, to the repository, test assertions for both the JS and wasm file outputs of a Rust crate compiled with `#[wasm_bindgen]`. These aren't intended to be exhaustive or large scale tests, but rather micro-tests to help observe the changes in `wasm-bindgen`'s output over time. The motivation for this commit is basically overhauling how all the GC passes work in `wasm-bindgen` today. The reorganization is also included in this commit as well. Previously `wasm-bindgen` would, in an ad-hoc fashion, run the GC passes of `walrus` in a bunch of places to ensure that less "garbage" was seen by future passes. This not only was a source of slowdown but it also was pretty brittle since `wasm-bindgen` kept breaking if extra iteams leaked through. The strategy taken in this commit is to have one precise location for a GC pass, and everything goes through there. This is achieved by: * All internal exports are removed immediately when generating the nonstandard wasm interface types section. Internal exports, intrinsics, and runtime support are all referenced by the various instructions and/or sections that use them. This means that we now have precise tracking of what an adapter uses. * This in turn enables us to implement the `add_gc_roots` function for `walrus` custom sections, which in turn allows walrus GC passes to do what `unexport_unused_intrinsics` did before. That function is now no longer necessary, but effectively works the same way. All intrinsics are unexported at the beginning and then they're selectively re-imported and re-exported through the JS glue generation pass as necessary and defined by the bindings. * Passes like the `anyref` pass are now much more precise about the intrinsics that they work with. The `anyref` pass also deletes any internal intrinsics found and also does some rewriting of the adapters aftewards now to hook up calls to the heap count import to the heap count intrinsic in the wasm module. * Fix handling of __wbindgen_realloc The final user of the `require_internal_export` function was `__wbindgen_realloc`. This usage has now been removed by updating how we handle usage of the `realloc` function. The wasm interface types standard doesn't have a `realloc` function slot, nor do I think it ever will. This means that as a polyfill for wasm interface types we'll always have to support the lack of `realloc`. For direct Rust to JS, however, we can still optionally handle `realloc`. This is all handled with a few internal changes. * Custom `StringToMemory` instructions now exist. These have an extra `realloc` slot to store an intrinsic, if found. * Our custom instructions are lowered to the standard instructions when generating an interface types section. * The `realloc` function, if present, is passed as an argument like the malloc function when passing strings to wasm. If it's not present we use a slower fallback, but if it's present we use the faster implementation. This should mean that there's little-to-no impact on existing users of `wasm-bindgen`, but this should continue to still work for wasm interface types polyfills and such. Additionally the GC passes now work in that they don't delete `__wbindgen_realloc` which we later try to reference. * Add an empty test for the anyref pass * Precisely track I32FromOptionAnyref's dependencies This depends on the anyref table and a function to allocate an index if the anyref pass is running, so be sure to track that in the instruction itself for GC rooting. * Trim extraneous exports from nop anyref module Or if you're otherwise not using anyref slices, don't force some intrinsics to exist. * Remove globals from reference tests Looks like these values adjust in slight but insignificant ways over time * Update the anyref xform tests
2019-12-04 12:01:39 -06:00
// Not really an exhaustive list, but works for our purposes.
fn to_js_identifier(name: &str) -> String {
name.chars()
.map(|c| {
if c.is_ascii() && (c.is_alphabetic() || c.is_numeric()) {
c
} else {
'_'
}
})
.collect()
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
}
fn adapter_name(&self, id: AdapterId) -> String {
format!("__wbg_adapter_{}", id.0)
}
fn generate_identifier(&mut self, name: &str) -> String {
let cnt = self
.defined_identifiers
.entry(name.to_string())
.or_insert(0);
*cnt += 1;
// We want to mangle `default` at once, so we can support default exports and don't generate
// invalid glue code like this: `import { default } from './module';`.
if *cnt == 1 && name != "default" {
name.to_string()
} else {
format!("{}{}", name, cnt)
}
}
}
fn check_duplicated_getter_and_setter_names(
exports: &[(&AdapterId, &AuxExport)],
) -> Result<(), Error> {
let verify_exports =
|first_class, first_field, second_class, second_field| -> Result<(), Error> {
let both_are_in_the_same_class = first_class == second_class;
let both_are_referencing_the_same_field = first_field == second_field;
if both_are_in_the_same_class && both_are_referencing_the_same_field {
bail!(format!(
"There can be only one getter/setter definition for `{}` in `{}`",
first_field, first_class
));
}
Ok(())
};
for (idx, (_, first_export)) in exports.iter().enumerate() {
for (_, second_export) in exports.iter().skip(idx + 1) {
match (&first_export.kind, &second_export.kind) {
(
AuxExportKind::Getter {
class: first_class,
field: first_field,
},
AuxExportKind::Getter {
class: second_class,
field: second_field,
},
) => verify_exports(first_class, first_field, second_class, second_field)?,
(
AuxExportKind::Setter {
class: first_class,
field: first_field,
},
AuxExportKind::Setter {
class: second_class,
field: second_field,
},
) => verify_exports(first_class, first_field, second_class, second_field)?,
_ => {}
}
}
}
Ok(())
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
fn format_doc_comments(comments: &str, js_doc_comments: Option<String>) -> String {
let body: String = comments.lines().map(|c| format!("*{}\n", c)).collect();
let doc = if let Some(docs) = js_doc_comments {
2018-07-13 22:36:51 -07:00
docs.lines().map(|l| format!("* {} \n", l)).collect()
} else {
String::new()
};
format!("/**\n{}{}*/\n", body, doc)
2018-06-27 22:42:34 -07:00
}
2019-04-30 10:26:03 -03:00
fn require_class<'a>(
exported_classes: &'a mut Option<BTreeMap<String, ExportedClass>>,
name: &str,
) -> &'a mut ExportedClass {
exported_classes
.as_mut()
.expect("classes already written")
.entry(name.to_string())
.or_insert_with(ExportedClass::default)
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
impl ExportedClass {
fn push(
&mut self,
docs: &str,
function_name: &str,
function_prefix: &str,
js: &str,
ts: Option<&str>,
) {
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
self.contents.push_str(docs);
self.contents.push_str(function_prefix);
self.contents.push_str(function_name);
self.contents.push_str(js);
self.contents.push_str("\n");
if let Some(ts) = ts {
self.typescript.push_str(docs);
self.typescript.push_str(" ");
self.typescript.push_str(function_prefix);
self.typescript.push_str(function_name);
self.typescript.push_str(ts);
self.typescript.push_str(";\n");
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
}
2019-05-30 10:56:59 -07:00
2019-06-06 16:11:51 -03:00
/// Used for adding a getter to a class, mainly to ensure that TypeScript
2019-05-30 10:56:59 -07:00
/// generation is handled specially.
fn push_getter(&mut self, docs: &str, field: &str, js: &str, ret_ty: Option<&str>) {
self.push_accessor(docs, field, js, "get ");
if let Some(ret_ty) = ret_ty {
self.push_accessor_ts(docs, field, ret_ty, false);
}
self.readable_properties.push(field.to_string());
2019-06-06 16:11:51 -03:00
}
/// Used for adding a setter to a class, mainly to ensure that TypeScript
/// generation is handled specially.
fn push_setter(
&mut self,
docs: &str,
field: &str,
js: &str,
ret_ty: Option<&str>,
might_be_optional_field: bool,
) {
self.push_accessor(docs, field, js, "set ");
if let Some(ret_ty) = ret_ty {
let is_optional = self.push_accessor_ts(docs, field, ret_ty, true);
*is_optional = might_be_optional_field;
}
2019-06-06 16:11:51 -03:00
}
fn push_accessor_ts(
&mut self,
docs: &str,
field: &str,
ret_ty: &str,
is_setter: bool,
) -> &mut bool {
let (ty, accessor_docs, has_setter, is_optional) = self
2019-05-30 10:56:59 -07:00
.typescript_fields
.entry(field.to_string())
.or_insert_with(Default::default);
2019-06-06 16:11:51 -03:00
*ty = ret_ty.to_string();
// Deterministic output: always use the getter's docs if available
if !docs.is_empty() && (accessor_docs.is_empty() || !is_setter) {
*accessor_docs = docs.to_owned();
}
*has_setter |= is_setter;
is_optional
2019-05-30 10:56:59 -07:00
}
fn push_accessor(&mut self, docs: &str, field: &str, js: &str, prefix: &str) {
self.contents.push_str(docs);
self.contents.push_str(prefix);
self.contents.push_str(field);
self.contents.push_str(js);
self.contents.push_str("\n");
}
First refactor for WebIDL bindings This commit starts the `wasm-bindgen` CLI tool down the road to being a true polyfill for WebIDL bindings. This refactor is probably the first of a few, but is hopefully the largest and most sprawling and everything will be a bit more targeted from here on out. The goal of this refactoring is to separate out the massive `crates/cli-support/src/js/mod.rs` into a number of separate pieces of functionality. It currently takes care of basically everything including: * Binding intrinsics * Handling anyref transformations * Generating all JS for imports/exports * All the logic for how to import and how to name imports * Execution and management of wasm-bindgen closures Many of these are separable concerns and most overlap with WebIDL bindings. The internal refactoring here is intended to make it more clear who's responsible for what as well as making some existing operations much more straightforward. At a high-level, the following changes are done: 1. A `src/webidl.rs` module is introduced. The purpose of this module is to take all of the raw wasm-bindgen custom sections from the module and transform them into a WebIDL bindings section. This module has a placeholder `WebidlCustomSection` which is nowhere near the actual custom section but if you squint is in theory very similar. It's hoped that this will eventually become the true WebIDL custom section, currently being developed in an external crate. Currently, however, the WebIDL bindings custom section only covers a subset of the functionality we export to wasm-bindgen users. To avoid leaving them high and dry this module also contains an auxiliary custom section named `WasmBindgenAux`. This custom section isn't intended to have a binary format, but is intended to represent a theoretical custom section necessary to couple with WebIDL bindings to achieve all our desired functionality in `wasm-bindgen`. It'll never be standardized, but it'll also never be serialized :) 2. The `src/webidl.rs` module now takes over quite a bit of functionality from `src/js/mod.rs`. Namely it handles synthesis of an `export_map` and an `import_map` mapping export/import IDs to exactly what's expected to be hooked up there. This does not include type information (as that's in the bindings section) but rather includes things like "this is the method of class A" or "this import is from module `foo`" and things like that. These could arguably be subsumed by future JS features as well, but that's for another time! 3. All handling of wasm-bindgen "descriptor functions" now happens in a dedicated `src/descriptors.rs` module. The output of this module is its own custom section (intended to be immediately consumed by the WebIDL module) which is in theory what we want to ourselves emit one day but rustc isn't capable of doing so right now. 4. Invocations and generations of imports are completely overhauled. Using the `import_map` generated in the WebIDL step all imports are now handled much more precisely in one location rather than haphazardly throughout the module. This means we have precise information about each import of the module and we only modify exactly what we're looking at. This also vastly simplifies intrinsic generation since it's all simply a codegen part of the `rust2js.rs` module now. 5. Handling of direct imports which don't have a JS shim generated is slightly different from before and is intended to be future-compatible with WebIDL bindings in its full glory, but we'll need to update it to handle cases for constructors and method calls eventually as well. 6. Intrinsic definitions now live in their own file (`src/intrinsic.rs`) and have a separated definition for their symbol name and signature. The actual implementation of each intrinsic lives in `rust2js.rs` There's a number of TODO items to finish before this merges. This includes reimplementing the anyref pass and actually implementing import maps for other targets. Those will come soon in follow-up commits, but the entire `tests/wasm/main.rs` suite is currently passing and this seems like a good checkpoint.
2019-05-23 09:15:26 -07:00
}
Rewrite wasm-bindgen with updated interface types proposal (#1882) This commit is a pretty large scale rewrite of the internals of wasm-bindgen. No user-facing changes are expected as a result of this PR, but due to the scale of changes here it's likely inevitable that at least something will break. I'm hoping to get more testing in though before landing! The purpose of this PR is to update wasm-bindgen to the current state of the interface types proposal. The wasm-bindgen tool was last updated when it was still called "WebIDL bindings" so it's been awhile! All support is now based on https://github.com/bytecodealliance/wasm-interface-types which defines parsers/binary format/writers/etc for wasm-interface types. This is a pretty massive PR and unfortunately can't really be split up any more afaik. I don't really expect realistic review of all the code here (or commits), but some high-level changes are: * Interface types now consists of a set of "adapter functions". The IR in wasm-bindgen is modeled the same way not. * Each adapter function has a list of instructions, and these instructions work at a higher level than wasm itself, for example with strings. * The wasm-bindgen tool has a suite of instructions which are specific to it and not present in the standard. (like before with webidl bindings) * The anyref/multi-value transformations are now greatly simplified. They're simply "optimization passes" over adapter functions, removing instructions that are otherwise present. This way we don't have to juggle so much all over the place, and instructions always have the same meaning.
2019-12-03 11:16:44 -06:00
struct MemView {
name: &'static str,
num: usize,
}
impl fmt::Display for MemView {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "{}{}", self.name, self.num)
}
}