This commit switches to executing `rustfmt` by default on
`web-sys`-generated bindings. This improves situations like "view
source" in Rustdoc as well as the IDE interactive debugging experience.
This was initially disabled by default because `rustfmt` took so long to
execute, but nowadays `web-sys` is by default much smaller so there's
much less need to avoid running `rustfmt` in fear of it taking too
long.
Closes#1457
Treats any object of shape `{ next: function }` as an iterator via new `is_type_of` method. This is consistent with JavaScript iteration protocol which we already respect.
Also fixes a minor issue that `is_function` was unnecessarily called twice (once explicitly and once as part of `dyn_into` which now does the same check).
This commit adds `#[derive(PartialEq, Eq)]` to many types throughout
`js-sys`. These types are basically all based on `Object`, which means
that `Object.is` can be used for `PartialEq` and the `Eq` requirements
are upheld.
The macro has also been updated to internally store the deref target
instead of unconditionally storing `JsValue`, allowing `#[derive]` to
work a bit better in these situations.
* Ensure `PartialEq` is implemented from these types to native Rust types
* Implement `From` between these type and native Rust types
* Deprecated `Number::new` and `Boolean::new` to discourage use of the
object forms, recommending the `from` constructors instead.
Closes#1446
The last write accidentally wasn't accounted for in the returned length
of the string and we unfortunately don't have any test coverage of
`encodeInto` since it requires Firefox nightly right now (and doesn't
work in Node yet).
Closes#1436
This commit adds an intrinsics to the `wasm_bindgen` crate which
accesses the `WebAssembly.Table` which is the function table of the
module. Eventually the thinking is that a module would import its own
function table via native wasm functionality (via `anyref` and such),
but until that's implemented let's add a binding for it ourselves!
Closes#1427
This commit aims to address #1348 via a number of strategies:
* Documentation is updated to warn about UTF-16 vs UTF-8 problems
between JS and Rust. Notably documenting that `as_string` and handling
of arguments is lossy when there are lone surrogates.
* A `JsString::is_valid_utf16` method was added to test whether
`as_string` is lossless or not.
The intention is that most default behavior of `wasm-bindgen` will
remain, but where necessary bindings will use `JsString` instead of
`str`/`String` and will manually check for `is_valid_utf16` as
necessary. It's also hypothesized that this is relatively rare and not
too performance critical, so an optimized intrinsic for `is_valid_utf16`
is not yet provided.
Closes#1348
Because of some incorrect use of `js.push_str(..)`, we could sometimes emit code
before the ES modules imports, which is syntactically invalid:
const __exports = {};
import { Thing } from '...'; // Syntax error!
This has been fixed by making sure that the correct `imports` or `imports_post`
string is built up. We now also assert that the `js` string is empty at the
location where we add imports if we're using ES modules.
This commit fixes the `init` function when passed a
`WebAssembly.Module`. Upon closer reading of the [spec] we see there's
two possible return values from `WebAssembly.instantiate`. If passed a
`Module`, it will return only the `Instance`. If passed a buffer source,
though, it'll return an object with the module/instance.
The fix here is to check the result value is an `Instance`, and if so
assume the input must have been a module so it's paired up in the
output.
Closes#1418
[spec]: http://webassembly.github.io/spec/js-api/index.html#webassembly-namespace
This is work towards #1399, although it's just for one-argument closures
where the first argument is a reference. No other closures with
references in argument position are supported yet
Instead of doubling the size on each iteration, use precise upper limit (3 * JS length) if the string turned out not to be ASCII-only. This results in maximum of 1 reallocation instead of O(log N).
Some dummy examples of what this would change:
- 1000 of ASCII chars: no change, allocates 1000 bytes and bails out.
- 1000 ASCII chars + 1 '😃': before allocated 1000 bytes and reallocated to 2000; now allocates 1000 bytes and reallocates to 1006.
- 1000 of '😃' chars: before allocated 1000 bytes, reallocated to 2000, finally reallocated again to 4000; now allocates 1000 bytes and reallocates to 4000 right away.
Related issue: #1313
