8000 ci: test on clang/arm64 by isaacbrodsky · Pull Request #975 · uber/h3 · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

ci: test on clang/arm64 #975

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

Merged
merged 6 commits into from
Mar 4, 2025
Merged

Conversation

isaacbrodsky
Copy link
Collaborator

No description provided.

@coveralls
Copy link
coveralls commented Feb 25, 2025

Coverage Status

coverage: 98.784%. remained the same
when pulling 502d568 on isaacbrodsky:linux-arm-clang
into 3a02395 on uber:master.

@ajfriend
Copy link
Contributor
ajfriend commented Feb 25, 2025

I can't immediately tell why things are failing. Maybe try fail-fast: false to get the complete list of what passes and what doesn't?

@dfellis
Copy link
Collaborator
dfellis commented Feb 25, 2025

I can't immediately tell why things are failing. Maybe try fail-fast: false to get the complete list of what passes and what doesn't?

https://github.com/uber/h3/actions/runs/13526744257/job/37799141287#step:8:570

Every single test fails immediately.

We should see if we can get it to dump the logs on failure, but I suspect it is accidentally trying to run x86-64 binaries in an arm64 environment.

@ajfriend
Copy link
Contributor

Every single test fails immediately.

Only one check fails, and the rest in that group are cancelled early Screenshot_20250225-113021~2.png

@dfellis
Copy link
Collaborator
dfellis commented Feb 25, 2025

Every single test fails immediately.

Only one check fails, and the rest in that group are cancelled early Screenshot_20250225-113021~2.png

Not what I meant. If you open the link, you will see that every test is executed for that run and they all fail. That implies that it is not generating binaries that work for the platform.

@isaacbrodsky
Copy link
Collaborator Author
isaacbrodsky commented Feb 25, 2025

I'll enable the print-output-on-failure option in CMake so we can find out. I wonder if some of these sanitizers are not implemented/available on arm64.

edit: dart-lang/sdk#55303 Seems to suggest we need a new version of Clang in order to use msan.

@dfellis
Copy link
Collaborator
dfellis commented Feb 25, 2025

Definitely better logging here. https://github.com/uber/h3/actions/runs/13529596356/job/37808380267#step:8:10522

Do these compile arg recommendations make sense for this test?

@ajfriend
Copy link
Contributor

@dfellis, ah, we were talking about different things. I mentioned fail-fast: false as a way to determine which checks are failing (as in what combinations of options/sanitizers aren't working).

That still might be helpful here, even if we only leave it on temporarily for this PR.

@ajfriend
Copy link
Contributor
ajfriend commented Mar 4, 2025

From what I can tell, -fsanitize=memory -fno-sanitize-recover=memory just isn't supported on ARM. Let's just drop it for ARM?

edit: OK, I'm just catching up: Maybe a newer version of clang would allow this, like you mentioned above, @isaacbrodsky

@ajfriend
Copy link
Contributor
ajfriend commented Mar 4, 2025

So maybe drop this test for now to get things passing? And we can create a follow up issue to update clang and add msan.

@isaacbrodsky isaacbrodsky merged commit f530729 into uber:master Mar 4, 2025
46 checks passed
@isaacbrodsky isaacbrodsky deleted the linux-arm-clang branch March 4, 2025 19:27
ajfriend pushed a commit to ajfriend/h3 that referenced this pull request Mar 6, 2025
* ci: test on clang/arm64

* CTEST_OUTPUT_ON_FAILURE

* test fail-fast

* fix

* exclude msan

* simplify
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
0