Message 378155 - 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	ddddaaaa
Recipients	belopolsky, ddddaaaa, p-ganssle
Date	2020-10-07.10:08:31
SpamBayes Score	-1.0
Marked as misclassified	Yes
Message-id	<1602065311.95.0.279586156365.issue41966@roundup.psfhosted.org>
In-reply-to

Content
I've run into an issue pickling a datetime.time subclass in PyPy3. I believe it arises because PyPy uses the pure Python implementation of time and time.__reduce_ex__() returns (time, ...) on this line (https://github.com/python/cpython/blob/044a1048ca93d466965afc027b91a5a9eb9ce23c/Lib/datetime.py#L1551) rather than (self.__class__, ...) as it does for datetime.datetime (and in datetime.date.__reduce__()). So when pickle creates the dump it uses the time class rather than my subclass.

I've run into an issue pickling a datetime.time subclass in PyPy3. I believe it arises because PyPy uses the pure Python implementation of time and time.__reduce_ex__() returns (time, ...) on this line (https://github.com/python/cpython/blob/044a1048ca93d466965afc027b91a5a9eb9ce23c/Lib/datetime.py#L1551) rather than (self.__class__, ...) as it does for datetime.datetime (and in datetime.date.__reduce__()). So when pickle creates the dump it uses the time class rather than my subclass.

History
Date	User	Action	Args
2020-10-07 10:08:31	ddddaaaa	set	recipients: + ddddaaaa, belopolsky, p-ganssle
2020-10-07 10:08:31	ddddaaaa	set	messageid: <1602065311.95.0.279586156365.issue41966@roundup.psfhosted.org>
2020-10-07 10:08:31	ddddaaaa	link	issue41966 messages
2020-10-07 10:08:31	ddddaaaa	create