multiprocessing build fails on Solaris 10

[MrJean1]

BPO	3149
Nosy	@pitrou

^{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 = <Date 2008-08-26.16:11:55.811>
created_at = <Date 2008-06-20.01:38:31.999>
labels = ['build']
title = 'multiprocessing build fails on Solaris 10'
updated_at = <Date 2008-08-26.16:11:55.779>
user = 'https://bugs.python.org/MrJean1'

bugs.python.org fields:

activity = <Date 2008-08-26.16:11:55.779>
actor = 'pitrou'
assignee = 'jnoller'
closed = True
closed_date = <Date 2008-08-26.16:11:55.811>
closer = 'pitrou'
components = ['Build']
creation = <Date 2008-06-20.01:38:31.999>
creator = 'MrJean1'
dependencies = []
files = []
hgrepos = []
issue_num = 3149
keywords = []
message_count = 3.0
messages = ['68438', '68465', '71984']
nosy_count = 4.0
nosy_names = ['pitrou', 'MrJean1', 'roudkerk', 'jnoller']
pr_nums = []
priority = 'normal'
resolution = 'duplicate'
stage = None
status = 'closed'
superseder = None
type = 'compile error'
url = 'https://bugs.python.org/issue3149'
versions = ['Python 2.6']

[MrJean1]

The multiprocessing module fails to build on Solaris 10 when using the
Sun C compiler. The failure is due to one missing symbol
SEM_VALUE_MAX.

Defining that symbol (e.g. as _POSIX_SEM_VALUE_MAX) builds the
_multiprocessing module for both 32- and 64-bit.

However, for 32-bit, 5 tests from test_multiprocessing fail due to the
missing _ctypes module. For 64-bit, the first 3 tests fail (like for
32-bit) but the 4th test hangs. Here is a log:

./python Lib/test/test_multiprocessing.py
test_array (__main__.WithProcessesTestArray) ... ERROR
test_getobj_getlock_obj (__main__.WithProcessesTestArray) ... ERROR
test_rawarray (__main__.WithProcessesTestArray) ... ERROR
test_notify (__main__.WithProcessesTestCondition) ... Process Process-1:
Traceback (most recent call last):
  File "/..../64/Python-2.6b1/Lib/multiprocessing/process.py", line 237, 
in _bootstrap
    self.run()
  File "/.../64/Python-2.6b1/Lib/multiprocessing/process.py", line 93, 
in run
    self._target(*self._args, **self._kwargs)
  File "Lib/test/test_multiprocessing.py", line 610, in f
    sleeping.release()
ValueError: semaphore or lock released too many times

This is Solaris 10 on an Ultra20 (Opteron) machine:

uname -a
SunOS unknown 5.10 Generic_118855-14 i86pc i386 i86pc

cc -v
cc: Sun C 5.8 2005/10/13

[MrJean1]

One more comment on SEM_VALUE_MAX for Solaris. _POSIX_SEM_VALUE_MAX is
defined in <limits.h> on Solaris and that header file must be included.
Better might be to use _SEM_VALUE_MAX which is defined in <param.h> and
that header files needs to be included as well. Rerhaps, INT_MAX is
best.

/usr/include/limits.h:#define _POSIX_SEM_NSEMS_MAX 256
/usr/include/limits.h:#define _POSIX_SEM_VALUE_MAX 32767

/usr/include/sys/param.h:#define _SEM_NSEMS_MAX INT_MAX
/usr/include/sys/param.h:#define _SEM_VALUE_MAX INT_MAX

Btw, the value of _POSIX_SEM_MAX_VALUE is 32767.

[pitrou]

This is a duplicate of bpo-3110.