A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

Addressed review feedback:

• moved str() coercion to re.sub() call site in _expand_help()
• removed _format_action changes
• trimmed regression coverage down to focused interpolated help proxy test

Testing:

• ./python.exe -m test test_argparse

I have made the requested changes; please review again

Thanks for making the requested changes!

@savannahostrowski: please review the changes made to this pull request.

Inspected the Win32 failure. test_argparse passed there; the red job is unrelated: test_profiling.test_sampling_profiler.test_generator_not_under_consumer_arithmetic in Windows / Build and test (Win32, switch-case).

This matches gh-150429, with follow-up fix in #150433. I pushed an empty commit to retrigger CI on the current PR head.

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

Addressed latest review feedback:

• fixed string-like proxy handling for argument help without a % format specifier
• kept scope limited to argument help only
• narrowed NEWS wording to match actual fix scope

Testing:

• ./python.exe -m test -v test_argparse -m test_argument_help_without_interpolation_accepts_string_like_proxy
• ./python.exe -m test -v test_argparse -m test_argument_help_interpolation_accepts_string_like_proxy
• ./python.exe -m test test_argparse

I have made the requested changes; please review again

Thanks for making the requested changes!

@savannahostrowski: please review the changes made to this pull request.

Alright, something about this issue had been nagging at me this afternoon, so I went back through the commit history again. I didn't realize this before but there's some asymmetry here between proxies with and without format specifiers. While proxies with a format specifier didn't raise before #142960, proxies without did, even with color off. As such, I think we should keep this PR minimally scoped to just resolving the immediate regression with proxies with format specifiers, and leave the discussion re. consistency for another issue/PR.

With that in mind, your commit dc8b7d7 was actually correct. The whole fix should just be the single str() coercion on the re.sub subject in _expand_help.

However, for next steps, I'd like to slow down a bit here. Instead of another revision, it would be helpful to hear your thoughts on the fix first. I would like to make sure that the changes landed are ones that work but are also ones you understand. We have a policy for this sort of thing in our devguide; please give our AI policy a read.

@savannahostrowski - I understand this PR should stay narrowly scoped to the regression from #142960. In _expand_help(), the issue is that re.sub() now receives a proxy help object on the interpolation path, so coercing that re.sub subject with str() fixes the regression without changing unrelated behavior. I also understand that proxy help without a format specifier is separate pre-existing behavior, so this PR should not try to address that here. I’ve read the AI policy and wanted to confirm that understanding before making any further changes.

@savannahostrowski - anything you want me to address before we take it further? Please let me know.

Thanks for working on this but after several rounds of review, I’m going to close this and handle the regression directly. The required fix is very small, and the current branch still includes behavior outside the regression’s scope.