Fix some non-Linux test failures #6473

TheRealFalcon · 2025-09-14T01:13:44Z

Proposed Commit Message

Fix some non-Linux test failures

At least partially fixes GH-5336

Additional Context

With these changes (and changes to the default TMPDIR), tests pass for me on a Mac. This PR contains multiple commits and each commit should be independently reviewable. I'll leave it up to whoever merges to decide if they want to squash or not.

Test Steps

Merge type

Squash merge using "Proposed Commit Message"
Rebase and merge unique commits. Requires commit messages per-commit each referencing the pull request number (#<PR_NUM>)

TheRealFalcon · 2025-09-14T01:17:01Z

Bwahaha...CLA bites me too. I'll get that taken care of

TheRealFalcon · 2025-09-18T20:13:52Z

Update: we're looking into doing the company-wide CLA, so it may be a while

holmanb · 2025-09-18T20:25:28Z

Update: we're looking into doing the company-wide CLA, so it may be a while

Thanks for the update. Keep us posted!

TheRealFalcon · 2025-10-17T19:36:40Z

Update: still waiting on CLA. Mostly just commenting to stop the stale checker from flagging this again 😄

TheRealFalcon · 2025-10-29T18:49:25Z

CLA has been signed on our end, so now just waiting for whatever magic needs to happen on the Canonical side for the CLA checker to be updated.

blackboxsw · 2025-10-29T22:07:04Z

CLA has been signed on our end, so now just waiting for whatever magic needs to happen on the Canonical side for the CLA checker to be updated.

Woo hoo! Welcome back!

blackboxsw · 2025-10-29T22:09:05Z

@TheRealFalcon in prepartion for landing, a couple of hacktoberfest commits landed dropping tests/helpers.py::TestCase which resulted in a merge conflict. I'll make sure I touch base with Canonical CLA folks to ensure the CLA signing is permitted now for your company so we can clear the CI workflow.

TheRealFalcon · 2025-10-30T15:13:25Z

Noted, I'll do the rebase once the CLA checker says I'm good. I've just been manually curling it to see.

holmanb

The code changes look good to me. There is a merge conflict, and the CLA job hasn't passed yet, but once those are resolved I'd say that this is ready for a rebase merge.

Thanks @TheRealFalcon!

holmanb · 2025-11-14T00:50:06Z

tests/unittests/conftest.py

 from cloudinit import user_data as ud
-from cloudinit import util
+from cloudinit import (
+    util,


nit: I assume this change wasn't intended? if this is what black wants now, maybe we could just move this up with the import list above this line?

Hrm, isort insists on it. It's because the as ud line has to be called out separately but is ordered before util. Seems like a stupid isort quirk, but I'd rather just leave it rather than try to work around it.

holmanb · 2025-11-14T00:56:27Z

tests/unittests/test_subp.py

        cmd = [
            SH,
            "-c",
-            'echo -n "$@"',


interesting - I didn't realize that this isn't defined in posix

Maybe this fix should be separated in a PR of its own to avoid being buried in this one.

I wouldn't bother with that. Since this is test code this shouldn't affect the end user in any way or be worth treating differently than this OS compatibility PR.

TheRealFalcon · 2025-11-20T22:39:11Z

Thanks!

the CLA job hasn't passed yet

Still waiting on this unfortunately. I believe we submitted it a few weeks ago. Is there something stuck on the Canonical side?

blackboxsw · 2025-11-20T23:07:37Z

Thanks!

the CLA job hasn't passed yet

Still waiting on this unfortunately. I believe we submitted it a few weeks ago. Is there something stuck on the Canonical side?

I've copied internal folks here saying we're blocked with "external" contributors having signed the CLA. My visibility here is that legal has been asked about this directly and they CLA-folks are awaiting a response. I'll bump this topic again so we can unblock.

TheRealFalcon · 2025-12-03T21:00:42Z

sigh looks like there's still more to sort on our end. Commenting again to keep the auto-closer quiet

The "has_gnu_date" function checks "date --help" output, but some versions of "date" don't include a "--help" flag.

`time.strftime()` has platform-specific behavior when dealing with a `gmtime()` struct.

It was removed in canonical#6356, but is still needed in some tests that interact with `/tmp`.

TheRealFalcon · 2026-01-05T16:27:49Z

Happy new year! CLA check finally passes!

I rebased and applied the one comment. I'll clean up the other test failure/linter stuff so we can hopefully finally get this through.

Many socket attributes are OS-specific, so ensure we have attributes that work for the tests that need them. Also, mock socket calls where needed.

'echo -n' isn't universal

Note that this test wasn't always Linux-specific, but when the pid stuff moved under Distro, the test was changed to test the base class (i.e., Linux) functionality only.

Implementation differences cause different results.

TheRealFalcon · 2026-01-05T18:47:07Z

This should be ready for review again

holmanb

@TheRealFalcon thanks for this! One minor nitpick and then I'm happy to merge it.

holmanb · 2026-01-05T19:07:59Z

tests/unittests/conftest.py

 from cloudinit import user_data as ud
-from cloudinit import util
+from cloudinit import (
+    util,


nit: an import list one item long seems messy. We should be able to include this in the multi-line list starting on line 10

Not sure if you saw my response on the earlier comment, but

isort insists on it. It's because the as ud line has to be called out separately but is ordered before util. Seems like a stupid isort quirk, but I'd rather just leave it rather than try to work around it.

If I move this one to the other list, the lint fails. Would you rather I leave it or try to silence isort?

Oops, thanks for restating that. It is fine as-is. Thanks!

Oops, thanks for restating that.

That's silly, but it doesn't need to change. I'm fine with it as-is.

holmanb · 2026-01-05T19:12:01Z

tests/unittests/runs/test_merge_run.py

 def cfg(tmp_path, mocker):
-    mocker.patch("cloudinit.util.os.chown")
+    # root group doesn't exist everywhere
+    mocker.patch("cloudinit.util.chownbyname")


+1, we shouldn't have been depending on that anyways

holmanb · 2026-01-05T19:18:13Z

tests/unittests/test_subp.py

        cmd = [
            SH,
            "-c",
-            'echo -n "$@"',


I wouldn't bother with that. Since this is test code this shouldn't affect the end user in any way or be worth treating differently than this OS compatibility PR.

holmanb · 2026-01-05T19:41:39Z

tests/unittests/conftest.py

    TMPDIR = os.getenv("TMPDIR", "/tmp")
-    Path(tmpdir, TMPDIR[1:]).mkdir(exist_ok=True)
+    Path(tmpdir, TMPDIR[1:]).mkdir(parents=True, exist_ok=True)
+    Path(tmpdir, "tmp").mkdir(exist_ok=True)


I see that setting TMPDIR causes tests to fail before this change. Tests also fail when TMPDIR includes a trailing slash, but fixing that isn't a very high priority in my opinion.

TheRealFalcon force-pushed the mac-tests branch from 101edc6 to cdbd582 Compare September 14, 2025 01:26

github-actions bot added the stale-pr Pull request is stale; will be auto-closed soon label Oct 3, 2025

holmanb removed the stale-pr Pull request is stale; will be auto-closed soon label Oct 3, 2025

canonical deleted a comment from github-actions bot Oct 28, 2025

github-actions bot added the stale-pr Pull request is stale; will be auto-closed soon label Nov 14, 2025

holmanb removed the stale-pr Pull request is stale; will be auto-closed soon label Nov 14, 2025

holmanb approved these changes Nov 14, 2025

View reviewed changes

canonical deleted a comment from github-actions bot Dec 9, 2025

github-actions bot added the stale-pr Pull request is stale; will be auto-closed soon label Dec 23, 2025

canonical deleted a comment from github-actions bot Dec 24, 2025

blackboxsw removed the stale-pr Pull request is stale; will be auto-closed soon label Dec 24, 2025

holmanb self-assigned this Jan 5, 2026

TheRealFalcon added 7 commits January 5, 2026 10:15

fix: don't traceback when using BSD date

9d9481b

The "has_gnu_date" function checks "date --help" output, but some versions of "date" don't include a "--help" flag.

test: ensure TMPDIR parents exist

5d9fba7

fix: ensure timestamp timezone encoded correctly in status.py

a1fcd9c

`time.strftime()` has platform-specific behavior when dealing with a `gmtime()` struct.

test: ensure we don't actually rename ntp.conf in test

629fff4

test: bring back fake tmpdir creation

4489bef

It was removed in canonical#6356, but is still needed in some tests that interact with `/tmp`.

test: mock chown to account for distro grp differences

4dd4ed9

test: hardcode passlib usage in Azure test

1166818

TheRealFalcon force-pushed the mac-tests branch from cdbd582 to 73d99a8 Compare January 5, 2026 16:24

TheRealFalcon added 7 commits January 5, 2026 10:38

test: add a socket fixtures and mocks

5c37d1e

Many socket attributes are OS-specific, so ensure we have attributes that work for the tests that need them. Also, mock socket calls where needed.

test: ensure symlinks are resolved in certain tests

9be643b

test: Replace 'echo -n' with 'printf'

cc61476

'echo -n' isn't universal

fix: ensure mount type is used if passed

d892b50

test: use realpath for tmp directories

80d369d

test: skip Linux-specific test on non-Linux

767510d

Note that this test wasn't always Linux-specific, but when the pid stuff moved under Distro, the test was changed to test the base class (i.e., Linux) functionality only.

test: Skip Azure openssl tests on non-Linux

2ea6d8d

Implementation differences cause different results.

TheRealFalcon force-pushed the mac-tests branch from 73d99a8 to 2ea6d8d Compare January 5, 2026 18:41

holmanb reviewed Jan 5, 2026

View reviewed changes

holmanb approved these changes Jan 6, 2026

View reviewed changes

holmanb merged commit b5080b9 into canonical:main Jan 6, 2026
20 checks passed

holmanb mentioned this pull request Jan 6, 2026

ci: exercise TMPDIR now that it works #6652

Merged

Fix some non-Linux test failures #6473

Fix some non-Linux test failures #6473

Uh oh!

Conversation

TheRealFalcon commented Sep 14, 2025

Proposed Commit Message

Additional Context

Test Steps

Merge type

Uh oh!

TheRealFalcon commented Sep 14, 2025

Uh oh!

TheRealFalcon commented Sep 18, 2025

Uh oh!

holmanb commented Sep 18, 2025

Uh oh!

TheRealFalcon commented Oct 17, 2025

Uh oh!

TheRealFalcon commented Oct 29, 2025

Uh oh!

blackboxsw commented Oct 29, 2025

Uh oh!

blackboxsw commented Oct 29, 2025

Uh oh!

TheRealFalcon commented Oct 30, 2025

Uh oh!

holmanb left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

TheRealFalcon commented Nov 20, 2025

Uh oh!

blackboxsw commented Nov 20, 2025

Uh oh!

TheRealFalcon commented Dec 3, 2025

Uh oh!

TheRealFalcon commented Jan 5, 2026

Uh oh!

TheRealFalcon commented Jan 5, 2026

Uh oh!

holmanb left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

holmanb Jan 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

holmanb Jan 6, 2026 •

edited

Loading