Skip to content

Disable strict compiles for systems not in our CI#1291

Closed
scotts wants to merge 4 commits intopytorch:mainfrom
scotts:restrict_strict
Closed

Disable strict compiles for systems not in our CI#1291
scotts wants to merge 4 commits intopytorch:mainfrom
scotts:restrict_strict

Conversation

@scotts
Copy link
Copy Markdown
Contributor

@scotts scotts commented Mar 10, 2026

This PR has some unrelated changes, but I'm doing them all in one so we can unblock upstreaming to PyTorch.

We're still running into trouble bumping PyTorch's submodule hash (pytorch/pytorch#176963). We turned on strict compiles for everything, but we only have tests in our CI for Linux systems for CPU only, CUDA and ROCm. We don't have tests for Mac, Windows and XPU. This PR only enables strict compiles for what we test in our CI. We should add these to our CI to avoid these problems when we upstream.

This also fixes some easy strictness stuff in those files that I'm confident won't cause any problems (casts; maybe unused).

Finally, I'm disabling the linting workflow because it's not lining up with internal. We can deal with that later.

@meta-cla meta-cla bot added the cla signed label Mar 10, 2026
@scotts scotts marked this pull request as ready for review March 10, 2026 14:03
@meta-codesync
Copy link
Copy Markdown

meta-codesync bot commented Mar 10, 2026

@scotts has imported this pull request. If you are a Meta employee, you can view this in D95949728.

@meta-codesync meta-codesync bot closed this in 0035505 Mar 10, 2026
@meta-codesync
Copy link
Copy Markdown

meta-codesync bot commented Mar 10, 2026

@scotts merged this pull request in 0035505.

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

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants