Python 3.13.1 test_datetime failures

[karel1980]

Bug report

Bug description:

Hi. I'm trying to build from source on a raspberry pi, on a fresh and fully updated debian bookworm install.

test_datetime was failing so I tried to get more information using make test TESTOPTS="-v test_datetime" as suggested by the readme, but that just runs all tests again without additional output for test_datetime (see terminal output pastedbelow.

I hope that's enough information. I'll try to look a bit further, this is just where I got the feeling it might not be 100% my own fault anymore ;)

karel@homeassistant:~/Python-3.13.0 $ lsb_release -a
No LSB modules are available.
Distributor ID:	Debian
Description:	Debian GNU/Linux 12 (bookworm)
Release:	12
Codename:	bookworm
karel@homeassistant:~/Python-3.13.0 $ uname -a
Linux homeassistant 6.6.51+rpt-rpi-2712 #1 SMP PREEMPT Debian 1:6.6.51-1+rpt3 (2024-10-08) aarch64 GNU/Linux

Test output

karel@homeassistant:~/Python-3.13.1 $ make test TESTOPTS="-v test_datetime"
Running code to generate profile data (this can take a while):
# First, we need to create a clean build with profile generation
# enabled.
make profile-gen-stamp
make[1]: Entering directory '/home/karel/Python-3.13.1'
make[1]: 'profile-gen-stamp' is up to date.
make[1]: Leaving directory '/home/karel/Python-3.13.1'
# Next, run the profile task to generate the profile information.
./python -m test --pgo --timeout=
Using random seed: 1333715147
0:00:00 load avg: 0.33 Run 44 tests sequentially in a single process
0:00:00 load avg: 0.33 [ 1/44] test_array
0:00:02 load avg: 0.38 [ 2/44] test_base64
0:00:02 load avg: 0.38 [ 3/44] test_binascii
0:00:02 load avg: 0.38 [ 4/44] test_binop
0:00:02 load avg: 0.38 [ 5/44] test_bisect
0:00:03 load avg: 0.38 [ 6/44] test_bytes
0:00:08 load avg: 0.43 [ 7/44] test_bz2
0:00:09 load avg: 0.43 [ 8/44] test_cmath
0:00:10 load avg: 0.43 [ 9/44] test_codecs
0:00:11 load avg: 0.48 [10/44] test_collections
0:00:13 load avg: 0.48 [11/44] test_complex
0:00:14 load avg: 0.48 [12/44] test_dataclasses
0:00:15 load avg: 0.48 [13/44] test_datetime
test test_datetime failed
0:00:20 load avg: 0.52 [14/44] test_decimal -- test_datetime failed (4 failures)
0:00:26 load avg: 0.67 [15/44] test_difflib
0:00:27 load avg: 0.67 [16/44] test_embed
0:00:38 load avg: 0.72 [17/44] test_float
0:00:38 load avg: 0.72 [18/44] test_fstring
0:00:42 load avg: 0.74 [19/44] test_functools
0:00:43 load avg: 0.74 [20/44] test_generators
0:00:43 load avg: 0.74 [21/44] test_hashlib
0:00:44 load avg: 0.74 [22/44] test_heapq
0:00:45 load avg: 0.74 [23/44] test_int
0:00:46 load avg: 0.76 [24/44] test_itertools
0:00:53 load avg: 0.78 [25/44] test_json
... rest omitted...

CPython versions tested on:

3.13

Operating systems tested on:

Linux

Linked PRs

• gh-129483: Make TestLocalTimeDisambiguation's time format locale independent #142193
• [3.14] gh-129483: Make TestLocalTimeDisambiguation's time format locale independent (GH-142193) #142258
• [3.13] gh-129483: Make TestLocalTimeDisambiguation's time format locale independent (GH-142193) #142259

[karel1980]

I ended up running ./python -mtest -v test_datetime and got the results listed below.
Looks like the test_vilnius_1949_* tests fail with something date formatting related. I suspect I have to configure a locale somewhere. I'll keep looking

======================================================================
FAIL: test_vilnius_1941_fromutc (test.datetimetester.TestLocalTimeDisambiguation_Pure.test_vilnius_1941_fromutc)
----------------------------------------------------------------------
Traceback (most recent call last):
 File "/home/karel/Python-3.13.1/Lib/test/datetimetester.py", line 5808, in test_vilnius_1941_fromutc
 self.assertEqual(ldt.strftime("%c %Z%z"),
 ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^
 'Mon Jun 23 23:59:59 1941 MSK+0300')
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: 'Mon 23 Jun 1941 23:59:59 CET MSK+0300' != 'Mon Jun 23 23:59:59 1941 MSK+0300'
- Mon 23 Jun 1941 23:59:59 CET MSK+0300
? --- ^^^^ ^^^
+ Mon Jun 23 23:59:59 1941 MSK+0300
? ^^ ^^^^


======================================================================
FAIL: test_vilnius_1941_toutc (test.datetimetester.TestLocalTimeDisambiguation_Pure.test_vilnius_1941_toutc)
----------------------------------------------------------------------
Traceback (most recent call last):
 File "/home/karel/Python-3.13.1/Lib/test/datetimetester.py", line 5832, in test_vilnius_1941_toutc
 self.assertEqual(gdt.strftime("%c %Z"),
 ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
 'Mon Jun 23 19:59:59 1941 UTC')
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: 'Mon 23 Jun 1941 19:59:59 UTC' != 'Mon Jun 23 19:59:59 1941 UTC'
- Mon 23 Jun 1941 19:59:59 UTC
? --- ^^^^
+ Mon Jun 23 19:59:59 1941 UTC
? ^^ ++++


======================================================================
FAIL: test_vilnius_1941_fromutc (test.datetimetester.TestLocalTimeDisambiguation_Fast.test_vilnius_1941_fromutc)
----------------------------------------------------------------------
Traceback (most recent call last):
 File "/home/karel/Python-3.13.1/Lib/test/datetimetester.py", line 5808, in test_vilnius_1941_fromutc
 self.assertEqual(ldt.strftime("%c %Z%z"),
 ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^
 'Mon Jun 23 23:59:59 1941 MSK+0300')
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: 'Mon 23 Jun 1941 23:59:59 CET MSK+0300' != 'Mon Jun 23 23:59:59 1941 MSK+0300'
- Mon 23 Jun 1941 23:59:59 CET MSK+0300
? --- ^^^^ ^^^
+ Mon Jun 23 23:59:59 1941 MSK+0300
? ^^ ^^^^


======================================================================
FAIL: test_vilnius_1941_toutc (test.datetimetester.TestLocalTimeDisambiguation_Fast.test_vilnius_1941_toutc)
----------------------------------------------------------------------
Traceback (most recent call last):
 File "/home/karel/Python-3.13.1/Lib/test/datetimetester.py", line 5832, in test_vilnius_1941_toutc
 self.assertEqual(gdt.strftime("%c %Z"),
 ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
 'Mon Jun 23 19:59:59 1941 UTC')
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: 'Mon 23 Jun 1941 19:59:59 UTC' != 'Mon Jun 23 19:59:59 1941 UTC'
- Mon 23 Jun 1941 19:59:59 UTC
? --- ^^^^
+ Mon Jun 23 19:59:59 1941 UTC
? ^^ ++++


----------------------------------------------------------------------
Ran 1036 tests in 4.967s

FAILED (failures=4, skipped=31)
test test_datetime failed
test_datetime failed (4 failures)

== Tests result: FAILURE ==

1 test failed:
 test_datetime

Total duration: 5.1 sec
Total tests: run=1,036 failures=4 skipped=31
Total test files: run=1/1 failed=1
Result: FAILURE
karel@homeassistant:~/Python-3.13.1 $ 

[karel1980]

Fixed by setting my locale to 'C'. It was set to en_GB.UTF-8
Command used to fix it: export LC_ALL=C

I feel like either I missed an instruction or it would be an improvement to the documentation if it's not there.
Let me know and I'll send a PR to add it to README.rst (or somewhere else if you prefer somewhere else)

[gkirchou]

Hit this issue today.

I am not sure if we should fix this under Makefile while I think we can program this with locale (and it's more reasonable to do this way unless we expect LC_ALL=C to always be the case). Added a PR.

[vstinner]

I cannot reproduce this issue on Fedora 43:

$ LC_ALL=en_GB.UTF-8 make test TESTOPTS="-v test_datetime"
(...)
Result: SUCCESS

pythoninfo:

$ ./python -m test.pythoninfo|grep -E 'libc|CC.version|uname|platform\.|os.environ'
CC.version: gcc (GCC) 15.2.1 20251111 (Red Hat 15.2.1-4)

os.environ[DISPLAY]: :0
os.environ[HOME]: /home/vstinner
os.environ[LANG]: fr_FR.UTF-8
os.environ[MAKEFLAGS]: -j14
os.environ[PATH]: /home/vstinner/.local/bin:/usr/local/bin:/usr/bin:/home/vstinner/.npm-global/bin
os.environ[SHELL]: /bin/bash
os.environ[TERM]: xterm-256color
os.environ[WAYLAND_DISPLAY]: wayland-0

os.uname: posix.uname_result(sysname='Linux', nodename='mona', release='6.17.7-300.fc43.x86_64', version='#1 SMP PREEMPT_DYNAMIC Sun Nov 2 15:30:09 UTC 2025', machine='x86_64')

platform.architecture: 64bit ELF
platform.freedesktop_os_release[ID]: fedora
platform.freedesktop_os_release[NAME]: Fedora Linux
platform.freedesktop_os_release[VARIANT_ID]: workstation
platform.freedesktop_os_release[VERSION]: 43 (Workstation Edition)
platform.freedesktop_os_release[VERSION_ID]: 43
platform.libc_ver: glibc 2.42
platform.platform: Linux-6.17.7-300.fc43.x86_64-x86_64-with-glibc2.42
platform.python_implementation: CPython

[gkirchou]

My environment's LC_ALL is default to unset.

kirchou@kirchou:~$ echo $LC_ALL

Based on Victor's comment, maybe LC_ALL=C is not the only one to solve this?

If that's the case, my first commit may not be the most accurate fix (depends on the original intention of the test)...

I changed this by explicitly changing %c to the exact expected time format.