-
Notifications
You must be signed in to change notification settings - Fork 853
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Run WebGPU CTS on CI #1421
Comments
Our friends at Deno provided this example https://github.com/lucacasonato/wgpu_cts_runner |
This isn't actually a requirement for the milestone, but still a very important issue to get sorted out. |
#1859 brings the Deno bindings in-repo. Let's consider this option (1). (2) Deno_webgpu could live in a separate crate. If there is a major PR here that we want to be safe about, we'd have to ask the author to also have a branch that updates the bindings. This is cumbersome, it's easy to fail following this. (3) Deno_webgpu could generate the API traces for all of the tests and store them somewhere, to be picked up by CI. This would still be problematic to test if the format of an API trace changes. (4) Keep relying on manual CTS runs. This is fragile. The costs of integrating Deno:
I think the costs are justified: Deno team has been pro-active in adopting WebGPU and helping us with CTS. And the benefit of a consistent CI coverage is worth it, especially now that PR would be able to unlock more tests. So let's do it! |
Is your feature request related to a problem? Please describe.
We need to have the strongest CI coverage on this repo if we want it to live independent.
Describe the solution you'd like
My first guess is integrating with Deno. Obviously, this will only work for non-breaking changes.
It could be OK, given the expectation that most changes in the future are non-breaking.
Describe alternatives you've considered
If that doesn't work out, we can also try Servo and Gecko.
Additional context
This is a requirement for moving move things into wgpu repo. It was born out of discussion in gfx-rs/gfx#3768, followed by an internal meeting at Mozilla.
The text was updated successfully, but these errors were encountered: