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
|
2018-05-22 12:34:41 -05:00
|
|
|
CHAR
|
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(Debug)]
|
|
|
|
pub enum Descriptor {
|
|
|
|
I8,
|
|
|
|
U8,
|
|
|
|
I16,
|
|
|
|
U16,
|
|
|
|
I32,
|
|
|
|
U32,
|
|
|
|
I64,
|
|
|
|
U64,
|
|
|
|
F32,
|
|
|
|
F64,
|
|
|
|
Boolean,
|
|
|
|
Function(Box<Function>),
|
2018-04-17 11:29:03 -07:00
|
|
|
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),
|
2018-05-22 12:34:41 -05:00
|
|
|
Char,
|
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(Debug)]
|
|
|
|
pub struct Function {
|
|
|
|
pub arguments: Vec<Descriptor>,
|
|
|
|
pub ret: Option<Descriptor>,
|
|
|
|
}
|
|
|
|
|
2018-04-17 11:29:03 -07:00
|
|
|
#[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,
|
2018-05-05 14:10:25 -07:00
|
|
|
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))),
|
2018-04-17 11:29:03 -07:00
|
|
|
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)
|
|
|
|
}
|
2018-05-22 12:34:41 -05:00
|
|
|
CHAR => Descriptor::Char,
|
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
|
|
|
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 {
|
2018-06-27 22:42:34 -07:00
|
|
|
Descriptor::I8
|
|
|
|
| Descriptor::U8
|
|
|
|
| Descriptor::I16
|
|
|
|
| Descriptor::U16
|
|
|
|
| Descriptor::I32
|
|
|
|
| Descriptor::U32
|
|
|
|
| Descriptor::F32
|
|
|
|
| Descriptor::F64
|
|
|
|
| Descriptor::Enum => 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 false,
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-05-05 14:10:25 -07:00
|
|
|
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,
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-04-17 11:29:03 -07:00
|
|
|
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,
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-04-17 11:29:03 -07:00
|
|
|
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 {
|
2018-04-17 11:29:03 -07:00
|
|
|
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,
|
2018-06-27 22:42:34 -07:00
|
|
|
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),
|
2018-05-05 14:10:25 -07:00
|
|
|
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),
|
2018-05-05 14:10:25 -07:00
|
|
|
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),
|
2018-06-27 22:42:34 -07:00
|
|
|
_ => 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 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,
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-04-13 14:13:14 -07:00
|
|
|
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 {
|
2018-04-13 14:13:14 -07:00
|
|
|
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 {
|
2018-06-27 22:42:34 -07:00
|
|
|
Descriptor::Ref(_) | Descriptor::RefMut(_) => 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
|
|
|
_ => false,
|
|
|
|
}
|
|
|
|
}
|
2018-05-01 10:06:35 -07:00
|
|
|
|
|
|
|
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
|
|
|
|
}
|
|
|
|
|
2018-04-17 11:29:03 -07:00
|
|
|
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",
|
2018-05-05 14:10:25 -07:00
|
|
|
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,
|
2018-05-05 14:10:25 -07:00
|
|
|
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,
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|