Message 371988 - Python tracker

➜

This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author	pashkin
Recipients	pashkin
Date	2020-06-21.08:53:47
SpamBayes Score	-1.0
Marked as misclassified	Yes
Message-id	<1592729628.06.0.446212297946.issue41062@roundup.psfhosted.org>
In-reply-to

Content
To me it seems like Advanced Debugger Support C-API doesn't make sense without HEAD_LOCK() and HEAD_UNLOCK() which are private right now. When researching how C-API works I've found this comment in the source code: https://github.com/python/cpython/blob/e838a9324c1719bb917ca81ede8d766b5cb551f4/Python/pystate.c#L1176 It says that the lists of interpreter-state and thread-state objects (that Adv. Debugger Support API operates on) could be mutated even when GIL is held so there is need to acquire head mutex when accessing them. But there is no way to acquire head mutex using public C-API. Am I right? If yes - it seems like HEAD_(UN)LOCK() should be made public.

To me it seems like Advanced Debugger Support C-API doesn't make sense without HEAD_LOCK() and HEAD_UNLOCK() which are private right now.

When researching how C-API works I've found this comment in the source code:
https://github.com/python/cpython/blob/e838a9324c1719bb917ca81ede8d766b5cb551f4/Python/pystate.c#L1176

It says that the lists of interpreter-state and thread-state objects (that Adv. Debugger Support API operates on) could be mutated even when GIL is held so there is need to acquire head mutex when accessing them. But there is no way to acquire head mutex using public C-API.

Am I right? If yes - it seems like HEAD_(UN)LOCK() should be made public.

History
Date	User	Action	Args
2020-06-21 08:53:48	pashkin	set	recipients: + pashkin
2020-06-21 08:53:48	pashkin	set	messageid: <1592729628.06.0.446212297946.issue41062@roundup.psfhosted.org>
2020-06-21 08:53:48	pashkin	link	issue41062 messages
2020-06-21 08:53:47	pashkin	create