Repository navigation
.replace functions in datetime do not call __new__ #89039
Description
| BPO | 44876 |
|---|---|
| Nosy | @joelgibson |
Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
assignee = None
closed_at = None
created_at = <Date 2021-08-09.22:58:31.112>
labels = ['type-bug', 'library']
title = '.replace functions in datetime do not call __new__'
updated_at = <Date 2021-08-09.22:58:31.112>
user = 'https://github.com/joelgibson'bugs.python.org fields:
activity = <Date 2021-08-09.22:58:31.112>
actor = 'joelgibson'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Library (Lib)']
creation = <Date 2021-08-09.22:58:31.112>
creator = 'joelgibson'
dependencies = []
files = []
hgrepos = []
issue_num = 44876
keywords = []
message_count = 1.0
messages = ['399295']
nosy_count = 1.0
nosy_names = ['joelgibson']
pr_nums = []
priority = 'normal'
resolution = None
stage = None
status = 'open'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue44876'
versions = []Linked PRs
- gh-89039: call subclass constructors in datetime.*.replace #114780
- [3.12] gh-89039: Call subclass constructors in datetime.*.replace (GH-114780) #131239
Rationale for not backporting: #89039 (comment).
Metadata
Metadata
Assignees
Labels
Projects
- Status
Activity
joelgibson commentedon Aug 9, 2021
I've come here while investigating a segfault when datetime.replace(...) was called with a Pandas Timestamp object:
pandas-dev/pandas#42305
The Python implementation of datetime.replace
cpython/Lib/datetime.py
Line 1823 in 03e5647
is polymorphic, in the sense that if a custom object like pd.Timestamp is passed in, a pd.Timestamp will come out rather than a datetime. This works just fine (copying and using Python code makes this segfault disappear).
The C implementation is also polymorphic
cpython/Modules/_datetimemodule.c
Line 5845 in 03e5647
but I think that something in the C implementation is wrong: eventually tp_alloc gets called for the passed type, but never tp_new, and then afterwards the custom type is treated just like a regular datetime. In the pd.Timestamp case, there are some extra fields that only get set in the __new__ method, leading to a segfault later when they're accessed.
I'm not familiar enough with the C/CPython interface (especially object setup and initialisation) to tell where a fix for this should go, I would assume that the line
cpython/Modules/_datetimemodule.c
Line 5872 in 03e5647
should be replaced by a call to PyObject_new(Py_TYPE(self), ...), or something similar.
This also affects date.replace and time.replace.
pythongh-89039: call subclass constructors in datetime.*.replace
pythongh-89039: call subclass constructors in datetime.*.replace
gh-89039: Call subclass constructors in datetime.*.replace (GH-114780)
pythongh-89039: Call subclass constructors in datetime.*.replace (pyt…
StanFromIreland commentedon Mar 12, 2025
Can this be closed? It doesn't seem its going to be back ported any time soon.
picnixz commentedon Mar 13, 2025
I think it's better not to backport it now. 3.11 is security-only and 3.12 is nearing its last bugfix and since the ABI seemed to have changed, and we needed the RM approval, it's better to be on the safe side.
[3.12] pythongh-89039: Call subclass constructors in datetime.*.repla…
serhiy-storchaka commentedon Mar 14, 2025
There was a small discussion about backporting in the PR. It seems there were no serious objections.
I created a backport to 3.12. It does not change the ABI. Previous complains were likely caused by using Argument Clinic in the backport.
In any case, I leave this on the release manager, @Yhg1s.
[3.12] gh-89039: Call subclass constructors in datetime.*.replace (GH…
Python: fix Python 3.13 incompatibility
Python: Fix datetime.replace(tzinfo=None) with Python 3.13
Fix datetime.replace(tzinfo=None) with Python 3.13
Fix datetime.replace(tzinfo=None) with Python 3.13