gh-94732: Fix KeyboardInterrupt race in asyncio run_forever()

[hetmankp]

This addresses issue #94732, ensuring that the event loop's _thread_id
attribute and the asyncgen hooks set by sys.set_asyncgen_hooks() are
always restored no matter where a KeyboardInterrupt exception is raised.

• Issue: Race with KeyboardInterrupt in asyncio loop run_forever() method #94732

[ghost]

The following commit authors need to sign the Contributor License Agreement:

• hetmankp@users.noreply.github.com

Click the button to sign:

[bedevere-bot]

Most changes to Python require a NEWS entry.

Please add it using the blurb_it web app or the blurb command-line tool.

[gvanrossum]

I am okay with skipping news, it please do sign the CLA.

[hetmankp]

I am okay with skipping news, it please do sign the CLA.

I have attempted to sign the CLA (i.e. enable the app) but I have the "Keep my email addresses private" option selected on Github, and the CLA app doesn't seem to accept my Github noreply email address. (My Github account is old enough, the noreply address shouldn't need the ID number).

Is this form of signing the agreement not supported at all or did I do something wrong?

[gvanrossum]

Maybe @ambv can help with the CLA issue?

[ambv]

the CLA app doesn't seem to accept my Github noreply email address

@hetmankp, can you say what exactly seems to be happening? We do have handling for the old-style noreply addresses, too, maybe there's a bug in there.

[hetmankp]

@hetmankp, can you say what exactly seems to be happening? We do have handling for the old-style noreply addresses, too, maybe there's a bug in there.

Sorry @ambv , I should have been more specific. When I click the "SIGN IN WITH GITHUB TO AGREE" button on the The PSF Individual Contributor Agreement page, I get taken to a white page with one line of text which reads:

Thank you for authorizing our application, but the CLA must be signed by the users who contributed to the PR. Authors emails are: hetmankp@users.noreply.github.com.

[bedevere-bot]

Most changes to Python require a NEWS entry.

Please add it using the blurb_it web app or the blurb command-line tool.

[bedevere-bot]

Most changes to Python require a NEWS entry.

Please add it using the blurb_it web app or the blurb command-line tool.

[gvanrossum]

How's the CLA signing going? Still no resolution?

[gvanrossum]

@hetmankp I cannot merge this without a signed CLA, alas. @ambv What can we do?

[hetmankp]

@hetmankp I cannot merge this without a signed CLA, alas. @ambv What can we do?

@gvanrossum@ambv , more than happy to work with whoever is available to provide further debugging information from my end.

Obviously this commit is simple enough we can probably find an alternative way forward if it comes to that, but if at all possible I'd rather we address the licencing system bug since it could affect other committers as well (plus it is the second time it has caused an issue for me attempting to submit a merge request into CPython).

[gvanrossum]

I don't know what's going on in the CLA bot, but presumably it would be avoidable if you linked a proper email address to your GitHub account and created a new PR using that email? (IIRC you can set the email address when committing with a git flag, or you can change your git config to set it.)

[hetmankp]

I don't know what's going on in the CLA bot, but presumably it would be avoidable if you linked a proper email address to your GitHub account and created a new PR using that email? (IIRC you can set the email address when committing with a git flag, or you can change your git config to set it.)

@gvanrossum I'd prefer not to do that though I think the issue might have been with my GitHub account anyway. It's an old account so as far as I'm aware it should have been using the simple anonymous email format without the ID number. However, I've just opened up a new pull request with the aforementioned ID number included and the CLA signing bot seems happy with that one.

The new pull request with the CLA signed is #97765 , so I think this one can be closed and that one can be used instead. Let me know if you'd like this done some other way though. Thanks and sorry for the complications.