gh-141732: Fix ExceptionGroup repr changing when original exception sequence is mutated

[dr-carlos]

• BaseExceptionGroup.__repr__ now overrides BaseException.__repr__, using the group's exception tuple instead of the mutable argument passed in.
• ComplexExtendsException (a private macro in Objects/exceptions.c) now has a parameter to modify the exception's __repr__ and not just its __str__.

• Issue: ExceptionGroup repr is misleading if original exception sequence is mutated #141732

---

📚 Documentation preview 📚: https://cpython-previews--141736.org.readthedocs.build/

[iritkatriel]

Thanks.

We don't make user-visible changes to builtin types without a PEP, so this PR cannot be merged as is.

We could document current behaviour (i.e., don't mutate the list) if it's not clear from the docs and then consider if we want to change it.

If we do want to change it, maybe it should be to make a shallow copy of the list rather than make it a tuple (so the repr/str won't change and break existing code that relies on current behaviour).

[dr-carlos]

Thanks.

We don't make user-visible changes to builtin types without a PEP, so this PR cannot be merged as is.

We could document current behaviour (i.e., don't mutate the list) if it's not clear from the docs and then consider if we want to change it.

If we do want to change it, maybe it should be to make a shallow copy of the list rather than make it a tuple (so the repr/str won't change and break existing code that relies on current behaviour).

Thanks for the quick response! That all makes sense.

Just to clarify - is c87b66b not a very similar change in this release without a PEP? But maybe I'm missing something :)

[iritkatriel]

If you implement BaseExceptionGroup_repr so that the output doesn't change, then this change makes sense and can be merged and backported.

[dr-carlos]

Is there anything blocking this PR?

I'll create an issue/PR updating docs to use tuples here (as suggested by @iritkatriel) once merged.

[dr-carlos]

Thanks! Not sure why the Android test is failing; but it seems unrelated to this PR.

[picnixz]

We'll need to wait for #142365 to be merged so that we can bypass the flaky tests on iOS and Android

[iritkatriel]

There are some test failures preventing this being merged.

[miss-islington-app]

Thanks @dr-carlos for the PR, and @iritkatriel for merging it 🌮🎉.. I'm working now to backport this PR to: 3.13, 3.14.
🐍🍒⛏🤖

[miss-islington-app]

Sorry, @dr-carlos and @iritkatriel, I could not cleanly backport this to 3.13 due to a conflict.
Please backport using cherry_picker on command line.

cherry_picker ff2577f56eb2170ef0afafa90f78c693df7ca562 3.13 

[bedevere-app]

GH-142388 is a backport of this pull request to the 3.14 branch.

[bedevere-app]

GH-142391 is a backport of this pull request to the 3.13 branch.

[bedevere-bot]

⚠️⚠️⚠️ Buildbot failure ⚠️⚠️⚠️

Hi! The buildbot s390x Fedora Stable Refleaks 3.13 (tier-3) has failed when building commit d768900.

What do you need to do:

1. Don't panic.
2. Check the buildbot page in the devguide if you don't know what the buildbots are or how they work.
3. Go to the page of the buildbot that failed (https://buildbot.python.org/#/builders/1662/builds/727) and take a look at the build logs.
4. Check if the failure is related to this commit (d768900) or if it is a false positive.
5. If the failure is related to this commit, please, reflect that on the issue and make a new Pull Request with a fix.

You can take a look at the buildbot page here:

https://buildbot.python.org/#/builders/1662/builds/727

Failed tests:

• test_import

Test leaking resources:

• test_import: references
• test_import: memory blocks

Summary of the results of the build (if available):

==

Click to see traceback logs

TracebackTests.test_exec_failure) ... ok TracebackTests.test_broken_parent_from) ... ok TracebackTests.test_exec_failure_nested) ... ok TracebackTests.test_nonexistent_module_nested) ... ok TracebackTests.test_broken_parent) ... ok TracebackTests.test_import_bug) ... ok TracebackTests.test_unencodable_filename) ... ok TracebackTests.test_nonexistent_module) ... ok TracebackTests.test_syntax_error) ... ok TracebackTests.test_broken_submodule) ... ok TracebackTests.test_broken_from) ... ok 

[dr-carlos]

Not sure why the buildbot is failing - had a look at the test output and the only thing I can find is

test_import leaked [1, 3, 0] references, sum=4 (this is fine) test_import leaked [1, 1, 6] memory blocks, sum=8 1 test failed again: test_import 

And I don't know enough about the import tests to know whether this is a problem (and whether this PR has anything to do with it).

[iritkatriel]

It's a problem - this is the reference leak test (something was not decref'ed). Entirely possible that it's not related to your PR. I kicked off the test again to see if it's intermittent.

You can run this test locally with -R3:3.

[dr-carlos]

It's a problem - this is the reference leak test (something was not decref'ed). Entirely possible that it's not related to your PR. I kicked off the test again to see if it's intermittent.

You can run this test locally with -R3:3.

Thanks for the heads up! I've tested locally with -R 3:3 and it passed, and the Buildbot seems to be running without errors again.