Commit Graph

12 Commits

Author SHA1 Message Date
Alex Crichton
cf08aee341 Tidy up a test 2018-06-30 11:19:27 -07:00
Alex Crichton
e06255fba5 Don't generate JS bindings for unused imports
If a JS import's shim isn't actually imported that means that somewhere along
the way it was optimized out or it was never used in the first place! In that
case we can skip generation of the JS bindings for it as it's not needed.
2018-06-29 15:56:12 -07:00
Alex Crichton
9a3ff77ea9
Support returning custom types in imports (#350)
Closes #320
2018-06-28 20:08:02 -05:00
Alex Crichton
4138583dff
Support wildcard arguments in foreign functions (#351)
No real reason to not support them!

Closes #346
2018-06-28 20:06:35 -05:00
R. Andrew Ohana
9127a0419f rustfmt all the things 2018-06-27 22:42:34 -07:00
Alex Crichton
ca8809b4e9 Don't try to codegen static into an impl
Closes #194
2018-05-11 08:28:09 -07:00
Alex Crichton
06664b34ce Fix parsing some Rust keywords in attributes
Closes #193
2018-05-09 08:01:57 -07:00
Alex Crichton
139b7a1aae Don't use the global stack for string lengths
This commit updates the `Abi` associated type for all slice types to a
`WasmSlice` type, an aggregate of two `u32` integers. This translates to an ABI
where when passed as a function argument it expands to two integer arguments,
and when passed as a return value it passes a return pointer as the first
argument to get filled in.

This is hopefully more forwards-compatible with the host bindings proposal which
uses this strategy for passing string arguments at least. It's a little sketchy
what we're doing as there's not really a stable ABI yet, but hopefully this'll
all be relatively stable for awhile!
2018-05-02 21:03:50 -07:00
Alex Crichton
412bebca72 Add support for version specifications
This commit adds a `#[wasm_bindgen(version = "...")]` attribute support. This
information is eventually written into a `__wasm_pack_unstable` section.
Currently this is a strawman for the proposal in ashleygwilliams/wasm-pack#101
2018-04-25 22:23:02 -07:00
Alex Crichton
d9a71b43db Assert empty JS heap/stack in tests
Turns out there was a bug when passing a vector of `JsValue` instances back to
JS all objects were leaked rather than correctly removed from the global slab.
2018-04-25 22:15:28 -07:00
Alex Crichton
792a8e132e Fix unused variables in generated code
Also deny all warnings in tests to prevent this creeping back in.

Closes #141
2018-04-19 13:16:59 -07:00
Alex Crichton
656d69816d Move all tests to the same suite
Nowadays the compile times are mitigated with incremental compilation and
otherwise it's much more ergonomic to run only one test if they're all in the
same suite.
2018-04-09 15:32:06 -07:00