gh-121723: Relax constraints on queue objects for logging.handlers.QueueHandler.

[picnixz]

Not sure if I need to explain what the Queue public API is. I think people are smart enough to understand that we expect a Queue-like object but tell me if you want me to put more details here.

• Issue: multiprocessing.Pool in spawn mode breaks logging.handlers.QueueHandler configuration #121723

---

📚 Documentation preview 📚: https://cpython-previews--122154.org.readthedocs.build/

[vsajip]

Looks good, but I have two thoughts:

1. Should the actual test for a queue protocol be factored out into an is_queue function? I know it'll only be used once, but it might make the overall flow slightly more readable.
2. Is it worth adding a test for the specific failure condition of #121723 (multiprocessing.set_start_method('spawn'))?

[picnixz]

1. Yes but I wasn't sure you wanted this in the module. AFAICT, the logging library likes to have long functions :)
2. This one requires a debug build (otherwise the assert is not executed) but I'll see if I can test it in debug only (I think the behaviour will become buggy in release mode).

[bedevere-bot]

🤖 New build scheduled with the buildbot fleet by @vsajip for commit 7729fb6 🤖

If you want to schedule another build, you need to add the 🔨 test-with-buildbots label again.

[picnixz]

@vsajip Before merging, I'll need to fix some typos so don't enable the automerge (I'm waiting for the build bots to finish).

[picnixz]

I'm waiting for the build bots to finish

Good thing I didn't wait in front of my screen, thinking it would finish FAST! I'll update my branch tomorrow.

[bedevere-bot]

🤖 New build scheduled with the buildbot fleet by @vsajip for commit d6b6ef3 🤖

If you want to schedule another build, you need to add the 🔨 test-with-buildbots label again.

[picnixz]

• Ok, so https://buildbot.python.org/#/builders/1295/builds/224 is a legit failure (it's just that queue is empty, but I suspect it's either an issue with the FT build or the timeout is not just enough).
• The other on AMD64 Arch Linux Usan 3.x/4532 is something I found on other PRs as well.

[vsajip]

or the timeout is not just enough

Do you want to try increasing the timeout?

[picnixz]

Do you want to try increasing the timeout?

For the sake of the test, I'll try but I'm not sure whether it's really a timeout issue or not. I'll put it at 60s (if it fails, then it's probably something else but it's probably something related to multiprocessing then).

[picnixz]

Can you run a single build bot actually? namely AMD64 Windows Server 2022 NoGIL PR/224? it appears that there are also other issues with multiprocessing for it but I'm not sure if they are related...

[vsajip]

I've asked that bot to rerun. I assume it'll pick up your latest changes - https://buildbot.python.org/#/builders/1295/builds/226

[picnixz]

It appears that the only failure is not related to this PR:

FAIL: test_search_major_2 (test.test_launcher.TestLauncher.test_search_major_2)
FAIL: test_search_major_3 (test.test_launcher.TestLauncher.test_search_major_3)
Re-running test_launcher in verbose mode (matching: test_search_major_2, test_search_major_3)
FAIL: test_search_major_2 (test.test_launcher.TestLauncher.test_search_major_2)
FAIL: test_search_major_3 (test.test_launcher.TestLauncher.test_search_major_3)

[miss-islington-app]

Thanks @picnixz for the PR, and @vsajip for merging it 🌮🎉.. I'm working now to backport this PR to: 3.12, 3.13.
🐍🍒⛏🤖

[bedevere-app]

GH-122603 is a backport of this pull request to the 3.13 branch.

[bedevere-app]

GH-122604 is a backport of this pull request to the 3.12 branch.