Stream: compiler development

Topic: CI/CD

Notification Bot (Jun 23 2026 at 13:21):

This topic was moved here from #compiler development > ci by Anton.

Anton (Jun 23 2026 at 13:22):

I renamed this to prevent a naming collision with the zulip export.

Anton (Jun 27 2026 at 16:20):

It's nice to have all the github CI machines easily available in the weekend :)

Anton (Jun 29 2026 at 11:16):

The zig-benchmarks CI step server is temporarily offline for updates.

Anton (Jun 29 2026 at 13:12):

Back up

Anton (Jul 04 2026 at 18:04):

We have a minici error on the main branch, I'm on it

expected type 'canonical_names.ModuleIdentityId', found 'canonical_names.ModuleNameId

Anton (Jul 04 2026 at 18:19):

PR#9947

Anton (Jul 07 2026 at 16:39):

I'm working on a fix for flaky issue:

  | === CLI Test Runner ===
  | 632 tests, 3 workers, 240s timeout, backends: interpreter, dev
  |
  | Glue generator: /Users/runner/work/roc/roc/zig-out/bin/roc, glue-opt=default, full-targets=false
  |
  |   timed out    roc test round-trips JSON parse and encode  (239121.7ms, phase=run)
  |         child command timed out; preserved work dir: /Users/runner/work/roc/roc/.zig-cache/roc-test/work/5-4ac202caf347fc1e-5
  |         signal 9
  |         stderr: stuck: child command timed out after 238999ms:

Will Hawkins (Jul 08 2026 at 12:25):

Anton said:

I'm working on a fix for flaky issue:

  | === CLI Test Runner ===
  | 632 tests, 3 workers, 240s timeout, backends: interpreter, dev
  |
  | Glue generator: /Users/runner/work/roc/roc/zig-out/bin/roc, glue-opt=default, full-targets=false
  |
  |   timed out    roc test round-trips JSON parse and encode  (239121.7ms, phase=run)
  |         child command timed out; preserved work dir: /Users/runner/work/roc/roc/.zig-cache/roc-test/work/5-4ac202caf347fc1e-5
  |         signal 9
  |         stderr: stuck: child command timed out after 238999ms:

Thank you (again!) for looking into this!

Anton (Jul 08 2026 at 14:05):

I think it's fixed now, let me know if you still see it (with commit 876f7c2 already on your branch)

Anton (Jul 13 2026 at 15:28):

I'd like to give zlint (like clippy but for zig) a try. Mainly for new contributors it's nice to have a quick check of zig best practices. What do you think?

Richard Feldman (Jul 13 2026 at 15:49):

sure! I assume in addition to the current tidy.zig?

Anton (Jul 13 2026 at 16:19):

Yes, in addition

Will Hawkins (Jul 17 2026 at 22:21):

Sorry to bother, but I had a question about the test_runner artifact. In d50491d77b84dec1d21d7da71d8e7bae70462b91, it looks like all meaningful references were removed to the test_runner artifact. In particular, after that commit, test_runner_exe is never referenced by a b.addRunArtifact call. I am just curious whether I am missing something silly (I am trying to work my way through the build/test system) or whether a failure to remove the code that adds the test_runner (and the main in src/cli/test/test_runner.zig) was a simple oversight.

If the latter, I'd be glad to build a PR for removing that dead code, if that is okay! Please forgive me if this question is silly/stupid -- as I said, I am attempting to grok the build/test system!

Will Hawkins (Jul 17 2026 at 22:22):

Will Hawkins said:

Sorry to bother, but I had a question about the test_runner artifact. In d50491d77b84dec1d21d7da71d8e7bae70462b91, it looks like all meaningful references were removed to the test_runner artifact. In particular, after that commit, test_runner_exe is never referenced by a b.addRunArtifact call. I am just curious whether I am missing something silly (I am trying to work my way through the build/test system) or whether a failure to remove the code that adds the test_runner (and the main in src/cli/test/test_runner.zig) was a simple oversight.

If the latter, I'd be glad to build a PR for removing that dead code, if that is okay! Please forgive me if this question is silly/stupid -- as I said, I am attempting to grok the build/test system!

Also, I apologize if this channel was not the appropriate one for this question.

Luke Boswell (Jul 17 2026 at 23:02):

Maybe in the process of merging different PR's that wasn't removed (or somehow was added back).

Luke Boswell (Jul 17 2026 at 23:03):

I'm guessing an oversight... the parallel cli runner is an upgrade if I recall correctly

Will Hawkins (Jul 17 2026 at 23:05):

Luke Boswell said:

Maybe in the process of merging different PR's that wasn't removed (or somehow was added back).

Sounds good! Thank you for the confirmation! I will submit a PR with that code removed, if that sounds okay?

Will Hawkins (Jul 18 2026 at 04:20):

Will Hawkins said:

Luke Boswell said:

Maybe in the process of merging different PR's that wasn't removed (or somehow was added back).

Sounds good! Thank you for the confirmation! I will submit a PR with that code removed, if that sounds okay?

Thank you for triggering the CI on the PR -- it looks like there is a failure in CI on Windows. I will investigate ASAP!

Luke Boswell (Jul 18 2026 at 04:53):

That failure looks like it's happening on all the PRs

Luke Boswell (Jul 18 2026 at 04:53):

Windows is timing out for some reason.

Luke Boswell (Jul 18 2026 at 04:54):

If you are able to find the root cause that'd be great

Last updated: Jul 23 2026 at 13:15 UTC