Skip to content

Modified Webgpu to work with the stable webgpu-header.#572

Merged
cwfitzgerald merged 33 commits into
gfx-rs:trunkfrom
Murmele:webgpu-header-1.0
Apr 8, 2026
Merged

Modified Webgpu to work with the stable webgpu-header.#572
cwfitzgerald merged 33 commits into
gfx-rs:trunkfrom
Murmele:webgpu-header-1.0

Conversation

@Murmele
Copy link
Copy Markdown
Contributor

@Murmele Murmele commented Apr 8, 2026

Fixed the ci issues from and closes #536

@cwfitzgerald
Copy link
Copy Markdown
Member

Thank you!

@cwfitzgerald cwfitzgerald merged commit f16f716 into gfx-rs:trunk Apr 8, 2026
33 checks passed
@Murmele
Copy link
Copy Markdown
Contributor Author

Murmele commented Apr 8, 2026

Thanks for merging!

@kolkov
Copy link
Copy Markdown
Contributor

kolkov commented Apr 20, 2026

Great work @Murmele — this covered everything we flagged in our earlier review of #536. Thanks for getting this across the finish line.

Two small observations (if I'm reading the code correctly):

  1. requiredFeatures in wgpuCreateInstancemap_instance_descriptor in conv.rs seems to ignore the base descriptor (_base parameter unused), so requiredFeatures/requiredFeatureCount are never validated. A user passing WGPUInstanceFeatureName_TimedWaitAny wouldn't get an error even though TimedWaitAny is unsupported.

  2. Stubs with unimplemented!()wgpuGetInstanceFeatures, wgpuHasInstanceFeature, wgpuBufferReadMappedRange, wgpuBufferWriteMappedRange will panic if called. Downstream bindings that probe capabilities at startup may hit these unexpectedly. Returning empty results or WGPUStatus_Error would be safer.

Neither is blocking — just flagging for awareness. Happy to submit a follow-up PR if useful.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants