Author nascheme
Recipients Mark.Shannon, christian.heimes, inada.naoki, jdemeyer, lukasz.langa, miss-islington, nascheme, pablogsal, petr.viktorin, pitrou, tim.peters, vstinner
Date 2019-09-30.16:51:19
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1569862280.07.0.0751067457009.issue38006@roundup.psfhosted.org>
In-reply-to
Content
> Why setting it back to release blocker? As far as I recall, this issue is not
> a Python 3.8 regression. The regression which triggered this old and existing
> bug has been fixed, see previous comments.

I leave it up to our glorious release manager to decide how to proceed.  IMHO,
GH-15826 (revert the addition of tp_clear for functions) and GH-15641 (Avoid
closure in weakref.WeakValueDictionary) are two cases of not fixing the real
bug, just hiding symptoms.  However, since this bug has existed for decades, I
wouldn't fault him for taking the conservative approach and not trying to
address the real bug in 3.8.

I've created a new PR (GH-16083) that I think is the correct fix.  If that PR
is applied, I think we should also restore tp_clear for functions (revert
GH-15826).  GH-15641 can probably stay.


Tim on my rough patch:
> It's semantically correct since we never wanted to execute a callback from a
> trash weakref to begin with

Thanks for your help Tim.  I think my PR is a cleaner fix that does essentially
the same thing.  Just call _PyWeakref_ClearRef() on weakrefs that are in
'unreachable'.
History
Date User Action Args
2019-09-30 16:51:20naschemesetrecipients: + nascheme, tim.peters, pitrou, vstinner, christian.heimes, petr.viktorin, inada.naoki, lukasz.langa, Mark.Shannon, jdemeyer, pablogsal, miss-islington
2019-09-30 16:51:20naschemesetmessageid: <1569862280.07.0.0751067457009.issue38006@roundup.psfhosted.org>
2019-09-30 16:51:20naschemelinkissue38006 messages
2019-09-30 16:51:19naschemecreate