New submission from Thomas Klausner <[email protected]>:
On NetBSD by default, the following tests do not finish in > 1h:
1:07:13 load avg: 0.00 running: test_compileall (1 hour 7 min),
test_multiprocessing_fork (1 hour 7 min), test_concurrent_futures (1 hour 6 min)
Defining HAVE_BROKEN_POSIX_SEMAPHORES fixes this, and they finish:
0:00:32 load avg: 10.63 [408/427/17] test_compileall passed ...
...
0:02:37 load avg: 3.04 [427/427/22] test_concurrent_futures passed (2 min 33
sec)
The last one fails:
test_multiprocessing_fork
with most of the subtests failing like this:
ERROR: test_shared_memory_SharedMemoryServer_ignores_sigint
(test.test_multiprocessing_fork.WithProcessesTestSharedMemory)
----------------------------------------------------------------------
Traceback (most recent call last):
File
"/scratch/lang/python310/work/Python-3.10.1/Lib/test/_test_multiprocessing.py",
line 4006, in test_shared_memory_SharedMemoryServer_ignores_sigint
sl = smm.ShareableList(range(10))
File
"/scratch/lang/python310/work/Python-3.10.1/Lib/multiprocessing/managers.py",
line 1372, in ShareableList
sl = shared_memory.ShareableList(sequence)
File
"/scratch/lang/python310/work/Python-3.10.1/Lib/multiprocessing/shared_memory.py",
line 327, in __init__
self.shm = SharedMemory(name, create=True, size=requested_size)
File
"/scratch/lang/python310/work/Python-3.10.1/Lib/multiprocessing/shared_memory.py",
line 92, in __init__
self._fd = _posixshmem.shm_open(
OSError: [Errno 86] Not supported: '/psm_b1ec903a'
I think this is a separate issue, so I'd like to define
HAVE_BROKEN_POSIX_SEMAPHORES for now.
This has been done in pkgsrc since at least python 2.7 (in 2011), I haven't dug
deeper.
----------
components: Interpreter Core
messages: 408291
nosy: wiz
priority: normal
severity: normal
status: open
title: NetBSD: do not use POSIX semaphores
type: behavior
versions: Python 3.11
_______________________________________
Python tracker <[email protected]>
<https://bugs.python.org/issue46045>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
