328 lines
8.2 KiB
Rust
Raw Normal View History

Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
use std::char;
macro_rules! tys {
($($a:ident)*) => (tys! { @ ($($a)*) 0 });
(@ () $v:expr) => {};
(@ ($a:ident $($b:ident)*) $v:expr) => {
const $a: u32 = $v;
tys!(@ ($($b)*) $v+1);
}
}
// NB: this list must be kept in sync with `src/describe.rs`
tys! {
I8
U8
I16
U16
I32
U32
I64
U64
F32
F64
BOOLEAN
FUNCTION
CLOSURE
STRING
REF
REFMUT
SLICE
VECTOR
ANYREF
ENUM
RUST_STRUCT
}
#[derive(Debug)]
pub enum Descriptor {
I8,
U8,
I16,
U16,
I32,
U32,
I64,
U64,
F32,
F64,
Boolean,
Function(Box<Function>),
Closure(Box<Closure>),
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
Ref(Box<Descriptor>),
RefMut(Box<Descriptor>),
Slice(Box<Descriptor>),
Vector(Box<Descriptor>),
String,
Anyref,
Enum,
RustStruct(String),
}
#[derive(Debug)]
pub struct Function {
pub arguments: Vec<Descriptor>,
pub ret: Option<Descriptor>,
}
#[derive(Debug)]
pub struct Closure {
pub function: Function,
pub mutable: bool,
}
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
#[derive(Copy, Clone)]
pub enum VectorKind {
I8,
U8,
I16,
U16,
I32,
U32,
I64,
U64,
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
F32,
F64,
String,
Anyref,
}
impl Descriptor {
pub fn decode(mut data: &[u32]) -> Descriptor {
let descriptor = Descriptor::_decode(&mut data);
assert!(data.is_empty());
descriptor
}
fn _decode(data: &mut &[u32]) -> Descriptor {
match get(data) {
I8 => Descriptor::I8,
I16 => Descriptor::I16,
I32 => Descriptor::I32,
I64 => Descriptor::I64,
U8 => Descriptor::U8,
U16 => Descriptor::U16,
U32 => Descriptor::U32,
U64 => Descriptor::U64,
F32 => Descriptor::F32,
F64 => Descriptor::F64,
BOOLEAN => Descriptor::Boolean,
FUNCTION => Descriptor::Function(Box::new(Function::decode(data))),
CLOSURE => Descriptor::Closure(Box::new(Closure::decode(data))),
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
REF => Descriptor::Ref(Box::new(Descriptor::_decode(data))),
REFMUT => Descriptor::RefMut(Box::new(Descriptor::_decode(data))),
SLICE => Descriptor::Slice(Box::new(Descriptor::_decode(data))),
VECTOR => Descriptor::Vector(Box::new(Descriptor::_decode(data))),
STRING => Descriptor::String,
ANYREF => Descriptor::Anyref,
ENUM => Descriptor::Enum,
RUST_STRUCT => {
let name = (0..get(data))
.map(|_| char::from_u32(get(data)).unwrap())
.collect();
Descriptor::RustStruct(name)
}
other => panic!("unknown descriptor: {}", other),
}
}
pub fn unwrap_function(&self) -> &Function {
match *self {
Descriptor::Function(ref f) => f,
_ => panic!("not a function"),
}
}
pub fn is_number(&self) -> bool {
match *self {
Descriptor::I8 |
Descriptor::U8 |
Descriptor::I16 |
Descriptor::U16 |
Descriptor::I32 |
Descriptor::U32 |
Descriptor::F32 |
Descriptor::F64 |
Descriptor::Enum => true,
_ => return false,
}
}
pub fn get_64bit(&self) -> Option<bool> {
match *self {
Descriptor::I64 => Some(true),
Descriptor::U64 => Some(false),
_ => None,
}
}
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
pub fn is_ref_anyref(&self) -> bool {
match *self {
Descriptor::Ref(ref s) => s.is_anyref(),
_ => return false,
}
}
pub fn ref_closure(&self) -> Option<&Closure> {
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
match *self {
Descriptor::Ref(ref s) => s.closure(),
_ => None,
}
}
pub fn closure(&self) -> Option<&Closure> {
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
match *self {
Descriptor::Closure(ref s) => Some(s),
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
_ => None,
}
}
pub fn is_anyref(&self) -> bool {
match *self {
Descriptor::Anyref => true,
_ => false,
}
}
pub fn vector_kind(&self) -> Option<VectorKind> {
let inner = match *self {
Descriptor::String => return Some(VectorKind::String),
Descriptor::Vector(ref d) => &**d,
Descriptor::Ref(ref d) => {
match **d {
Descriptor::Slice(ref d) => &**d,
Descriptor::String => return Some(VectorKind::String),
_ => return None,
}
}
Descriptor::RefMut(ref d) => {
match **d {
Descriptor::Slice(ref d) => &**d,
_ => return None,
}
}
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
_ => return None,
};
match *inner {
Descriptor::I8 => Some(VectorKind::I8),
Descriptor::I16 => Some(VectorKind::I16),
Descriptor::I32 => Some(VectorKind::I32),
Descriptor::I64 => Some(VectorKind::I64),
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
Descriptor::U8 => Some(VectorKind::U8),
Descriptor::U16 => Some(VectorKind::U16),
Descriptor::U32 => Some(VectorKind::U32),
Descriptor::U64 => Some(VectorKind::U64),
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
Descriptor::F32 => Some(VectorKind::F32),
Descriptor::F64 => Some(VectorKind::F64),
Descriptor::Anyref => Some(VectorKind::Anyref),
_ => None
}
}
pub fn rust_struct(&self) -> Option<&str> {
let inner = match *self {
Descriptor::Ref(ref d) => &**d,
Descriptor::RefMut(ref d) => &**d,
ref d => d,
};
match *inner {
Descriptor::RustStruct(ref s) => Some(s),
_ => None,
}
}
pub fn stack_closure(&self) -> Option<(&Function, bool)> {
let (inner, mutable) = match *self {
Descriptor::Ref(ref d) => (&**d, false),
Descriptor::RefMut(ref d) => (&**d, true),
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
_ => return None,
};
match *inner {
Descriptor::Function(ref f) => Some((f, mutable)),
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
_ => None,
}
}
pub fn is_by_ref(&self) -> bool {
match *self {
Descriptor::Ref(_) |
Descriptor::RefMut(_) => true,
_ => false,
}
}
pub fn is_mut_ref(&self) -> bool {
match *self {
Descriptor::RefMut(_) => true,
_ => false,
}
}
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
}
fn get(a: &mut &[u32]) -> u32 {
let ret = a[0];
*a = &a[1..];
ret
}
impl Closure {
fn decode(data: &mut &[u32]) -> Closure {
let mutable = get(data) == REFMUT;
assert_eq!(get(data), FUNCTION);
Closure {
mutable,
function: Function::decode(data),
}
}
}
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
impl Function {
fn decode(data: &mut &[u32]) -> Function {
let arguments = (0..get(data))
.map(|_| Descriptor::_decode(data))
.collect::<Vec<_>>();
let ret = if get(data) == 0 {
None
} else {
Some(Descriptor::_decode(data))
};
Function { arguments, ret }
}
}
impl VectorKind {
pub fn js_ty(&self) -> &str {
match *self {
VectorKind::String => "string",
VectorKind::I8 => "Int8Array",
VectorKind::U8 => "Uint8Array",
VectorKind::I16 => "Int16Array",
VectorKind::U16 => "Uint16Array",
VectorKind::I32 => "Int32Array",
VectorKind::U32 => "Uint32Array",
VectorKind::I64 => "BigInt64Array",
VectorKind::U64 => "BigUint64Array",
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
VectorKind::F32 => "Float32Array",
VectorKind::F64 => "Float64Array",
VectorKind::Anyref => "any[]",
}
}
pub fn size(&self) -> usize {
match *self {
VectorKind::String => 1,
VectorKind::I8 => 1,
VectorKind::U8 => 1,
VectorKind::I16 => 2,
VectorKind::U16 => 2,
VectorKind::I32 => 4,
VectorKind::U32 => 4,
VectorKind::I64 => 8,
VectorKind::U64 => 8,
Overhaul how type information gets to the CLI This commit is a complete overhaul of how the `#[wasm_bindgen]` macro communicates type information to the CLI tool, and it's done in a somewhat... unconventional fashion. Today we've got a problem where the generated JS needs to understand the types of each function exported or imported. This understanding is what enables it to generate the appropriate JS wrappers and such. We want to, however, be quite flexible and extensible in types that are supported across the boundary, which means that internally we rely on the trait system to resolve what's what. Communicating the type information historically was done by creating a four byte "descriptor" and using associated type projections to communicate that to the CLI tool. Unfortunately four bytes isn't a lot of space to cram information like arguments to a generic function, tuple types, etc. In general this just wasn't flexible enough and the way custom references were treated was also already a bit of a hack. This commit takes a radical step of creating a **descriptor function** for each function imported/exported. The really crazy part is that the `wasm-bindgen` CLI tool now embeds a wasm interpreter and executes these functions when the CLI tool is invoked. By allowing arbitrary functions to get executed it's now *much* easier to inform `wasm-bindgen` about complicated structures of types. Rest assured though that all these descriptor functions are automatically unexported and gc'd away, so this should not have any impact on binary sizes A new internal trait, `WasmDescribe`, is added to represent a description of all types, sort of like a serialization of the structure of a type that `wasm-bindgen` can understand. This works by calling a special exported function with a `u32` value a bunch of times. This means that when we run a descriptor we effectively get a `Vec<u32>` in the `wasm-bindgen` CLI tool. This list of integers can then be parsed into a rich `enum` for the JS generation to work with. This commit currently only retains feature parity with the previous implementation. I hope to soon solve issues like #123, #104, and #111 with this support.
2018-04-13 07:33:46 -07:00
VectorKind::F32 => 4,
VectorKind::F64 => 8,
VectorKind::Anyref => 4,
}
}
}