GeezerVison

Note to myself or whoever who merges this. Please add @wangxiang-hz to co-authors when merging as this is the fixed version of their PR, and the work should also be attributed to them.

I will review and hopefully merge this on Friday when I have more time. Thanks.

#103551 has a simple test case for the issue with the original PR, might be useful to add as a unit test for this PR.

#103551 has a simple test case for the issue with the original PR, might be useful to add as a unit test for this PR.

@property raises and __getattr__ not called --> added in #103358
@property calls __getattr__, and __getattr__ raises --> added in #103336

The test of #103551 should be covered by #103358. But a little duplication is okay. I will add it later.

🤖 New build scheduled with the buildbot fleet by @Fidget-Spinner for commit c3aebf6 🤖

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

Does anyone have any idea why this fails?
https://buildbot.python.org/all/#/builders/738/builds/3112
The failure is a timing test. This PR seems to have changed the time usage of one particular test on ARM.

Got a timeout here: https://buildbot.python.org/all/#/builders/403/builds/1066

all builds: https://buildbot.python.org/all/#/changes/22158

Got a timeout here: https://buildbot.python.org/all/#/builders/403/builds/1066

This specific test seems flaky. I would ignore it.

@Fidget-Spinner How is it going on? We are running out of time. New features will soon be frozen in May. 7th.
https://peps.python.org/pep-0693/

Or will this be considered as a bug fix?

🤖 New build scheduled with the buildbot fleet by @Fidget-Spinner for commit 75d9a70 🤖

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

@Fidget-Spinner How is it going on? We are running out of time. New features will soon be frozen in May. 7th. https://peps.python.org/pep-0693/

Or will this be considered as a bug fix?

I'm triggering a buildbot rerun to see if that specific buildbot is actually impacted.

https://buildbot.python.org/all/#builders/178/builds/1189

test_monitoring failure is tracked here
#103845
test_import failure is tracked here
#103879 (comment)

Seems no more failures

Please add news. You can copy the previous PR's news item. I will attribute the commit to both of you.

Please, don't worry about 229 random files reported as changed. It's a GitHub after-merge artifact, the final commit 59c27fa squash-merged into main is fine.

Sorry guys, I just noticed this pull request. I was buy with my work these weeks and just finished a major update at the end of April..
I think @sunmy2019 is right and his commits is better. call_attribute should only be called when attribute is not found, then we call it, and clear the PyExc_AttributeError being raised.
I made a mistake, PyObject_GenericGetAttr may not raise attribute error but other scripts may still raise other exceptions. When this happen, call_attribute should not be called.
Thank you all for finding this and sorry for the bugs triggered.