David Neto

this seems like it should be filed upstream in the Glslang project https://github.com/KhronosGroup/glslang

The WebGPU maxComputeInvocationsPerWorkgroup limit corresponds to Vulkan's `maxComputeWorkGroupInvocations` limit. See Sascha Willems' GPUinfo.or for this data on Vulkan: https://vulkan.gpuinfo.org/displaydevicelimit.php?name=maxComputeWorkGroupInvocations&platform=all That counts distinct GPUs reported, and is not weighted by number...

Familiarity with C is not a goal. The comma operator adds unnecessary complexity. (Design DRY. See https://extensiblewebmanifesto.org/) As for multi-variable updates in loops, I refer you to my comment at...

Copying request from @MichealRead in #5152 --- Currently, WGSL supports only 32-bit integer types (i32 and u32). Many applications could greatly benefit from native support for additional integer types, such...

> Enhanced integer support can benefit many things but primarily AI and video work where 8 and 16 bit buffers are common. When the buffer is a large array of...

Works in dawn.node

> Looks like the tests here do not cover the case where the is a break if in the continuing block. Is the corresponding spec PR supposed to cover that...

I want to add more cases, as found from testing the Tint fixes. See the DeadLoopTests in https://dawn-review.googlesource.com/c/dawn/+/267474 that cover `loop`, `for-with-cond`, `for-without-cond`, and `while`. It's tricky enough that it...

I've reworked the statement uniformity checks to use an explicit `sensitive` data field. And I've added the variants from the Tint unit test, as described in the updated commit comment.

Passes on dawn.node when Dawn is patched with https://dawn-review.googlesource.com/c/dawn/+/267474