The handle dependency validation code was using the handle's index
directly while trying to erase the handle type. This would cause the
validator to crash while processing the first handle of the arena since
it would be trying to construct a `NonZeroU32` with a zero.
This commit fixes the issue by adding 1 to the index which not only
fixes this panic but also makes so that the created Handle is equal to
the passed handle (minus the type that was erased)
It also fixes the error message not including the subject's kind
Before proceeding with any other validation, check that all Handles are valid for their arenas, and refer only to older handles than themselves. This allows subsequent stages to simply use indexing without panics, assuming validation has passed.
* Add support for WGSL's `atomicCompareExchangeWeak` with the `__atomic_compare_exchange_result` struct, and add SPIR-V codegen for it.
Partially addresses https://github.com/gpuweb/gpuweb/pull/2113, #1755.
* Add tests for `atomicCompareExchangeWeak`, and support both u32 and i32 atomics with it.
* More thorough typechecking of the struct returned by `atomicCompareExchangeWeak`.
* Make some (currently hacky) changes to enable multiview in webgl
* Fix ViewIndex built in for this extension
* Run cargo fmt, fix tests
* Allow specifying if we're targetting webgl in the glsl version
* Document multiview2 extension
* fn embedded -> const fn embedded
* Fix tests
* Fix benches
* Add snapshot tests
* Revamp so that the glsl options have some multiview options. Also add tests
* Make clippy happier
* Go back to having is_webgl be part of Version
* Use wgsl as input for tests
* Rename Version::new_embedded to Version::new_gles, fix glsl validation
* Run cargo fmt
* Fix brand new clippy warnings
Introduce a new `TypeFlags::CONSTRUCTIBLE` flag, corresponding to
WGSL's "constructible types". Set this on the appropriate types.
Check for this flag on function return types.
Add `valid::Capabilities` flags for the `ClipDistance` and
`CullDistance` builtins, which are not supported by all back ends.
Have the CLI perform valation with only those capabilities that the
requested back ends support.
Fixes#1961.
A `ValuePointer` type should always be handled exactly like the
equivalent `Pointer` to `Scalar` type. To this end:
- Adjust the handling of `TypeFlags::ARGUMENT`.
- Neither `Pointer` nor `ValuePointer` types are ever `DATA`.
When `VaryingContext::validate` discovers that it needs to examine
each member of a struct, pass the type as an argument to `validate`
and `validate_impl`, rather than mutating `self.ty` on each iteration
of the loop. This matches the way we're handling the binding, and
avoids stashing per-iteration state in a surprising place.
This should have no effect on execution.
* use strip_prefix
* make fn const (resolving TODO)
* make use of nested OR patterns in match arms
* warn on clippy::missing_const_for_fn
* constify functions
* ignore clippy::missing_const_for_fn for into_inner functions
* Allow vecN<i32> and vecN<u32> in `dot()`, first changes
* Added a test case
* Fix the test
* Changes to baking of expressions, incl args of integer dot product
* Implemented requested changes for glsl backend
* Added support for integer dot product on MSL backend
* Removed outdated code for hlsl and wgls writers
* Implement in spv backend
* Commit modified outputs from running the tests
* cargo fmt
* Applied requested changes for both MSL and GLSL back
* Changes to spv back
* Committed all test output changes
* Cargo fmt
* Added a comment w.r.t. VK_KHR_shader_integer_dot_product
* Implemented requested svp change
* Minor change to test case
This is because I wanted to highlight the fact that the correct
id is used in the last sum of the integer dot product expression
* Changed function signature
since it could not fail, changed it to simply return `void`
* Rewrite front/back doc summaries
- Use line comments instead of block comments
- Standardize language for each front/backend
- Add reference link for each format
- Minor punctuation changes
* Add documentation for keywords module
* Clarify contents of keywords module in summary
* Refer to modules by their type name
* Add basic summary for valid module
* Adjust EarlyDepthTest and ConservativeDepth docs
* Remove "in" from list
* Adjust wording
* Standardize format of docstrings
* Adjust module links to be consistent with other links
* Add summary for reserved keywords list
* Remove extraneous doc spaces with `cargo fmt`
* Correct spelling of whether and rewrite some lines
* Fill out GLSL backend docs
* Remove unnecessary link targets
* Fill out DOT backend docs
* Change module line comments to block comments
* Remove unnecessary spaces
* Fix mistake during rebasing