@std/webgpu/texture_with_data.ts has flaky test
https://github.com/denoland/std/actions/runs/13420418226/job/37491446085?pr=6408
ERRORS
createTextureWithData() => ./webgpu/texture_with_data_test.ts:7:6
error: TypeError: Failed to execute 'call' on 'GPUDevice': 'format' of 'GPUTextureDescriptor' (Argument 0) can not be converted to a enum
const texture = device.createTexture(descriptor);
^
at createTextureWithData (file:///D:/a/std/std/webgpu/texture_with_data.ts:130:26)
at fn (file:///D:/a/std/std/webgpu/texture_with_data_test.ts:16:5)
FAILURES
createTextureWithData() => ./webgpu/texture_with_data_test.ts:7:6
FAILED | 4142 passed (2719 steps) | 1 failed | 25 ignored (2m17s)
error: Test failed
Error: Process completed with exit code 1.
tried locally, however cannot reproduce. also, this being a flake makes no sense...
tried locally, however cannot reproduce. also, this being a flake makes no sense...
No idea why this happens, but I've seen this error multiple times on different @std PRs. Maybe it is a bug in deno's GPUDevice or GPUTextureDescriptor implementation?
The error message seems created in this line? https://github.com/denoland/deno_core/blob/7cbc5ae35c279972d46e5552b7a5978860df611e/core/webidl.rs#L96
Is it possible that CI machine's state affects this?