Two warnings from pyflakes:

diff -r 72889bf8531d Lib/multiprocessing/spawn.py
--- a/Lib/multiprocessing/spawn.py      Tue Mar 18 13:21:29 2014 +0100
+++ b/Lib/multiprocessing/spawn.py      Wed Mar 19 11:49:24 2014 +0100
@@ -64,6 +64,7 @@ def freeze_support():
     Run code for process object if this in not the main process
     '''
     if is_forking(sys.argv):
+        # FIXME: main() is undefined
         main()
         sys.exit()
 
diff -r 72889bf8531d Lib/multiprocessing/synchronize.py
--- a/Lib/multiprocessing/synchronize.py        Tue Mar 18 13:21:29 2014 +0100
+++ b/Lib/multiprocessing/synchronize.py        Wed Mar 19 11:50:50 2014 +0100
@@ -51,7 +51,6 @@ class SemLock(object):
     _rand = tempfile._RandomNameSequence()
 
     def __init__(self, kind, value, maxvalue, *, ctx):
-        ctx = ctx or get_context()
         ctx = ctx.get_context()
         unlink_now = sys.platform == 'win32' or ctx._name == 'fork'
         for i in range(100):

Serhiy suggested (msg214136):

"Lib/multiprocessing/synchronize.py:
Add at the start of the file:
from . import get_context"

It looks like multiprocessing.get_context() exists, but I don't see how it is defined.

For main(), it should maybe be replaced with _main()?

> It looks like multiprocessing.get_context() exists, but I don't see how it
> is defined.

It is defined by:

globals().update((name, getattr(context._default_context, name))
                 for name in context._default_context.__all__)

> For main(), it should maybe be replaced with _main()?

_main() has different signature.

New changeset 619331c67638 by Richard Oudkerk in branch '3.4':
Issue #20990: Fix issues found by pyflakes for multiprocessing.
http://hg.python.org/cpython/rev/619331c67638

Would it be possible to write a test?

Le dimanche 23 mars 2014, Roundup Robot <report@bugs.python.org> a écrit :

>
> Roundup Robot added the comment:
>
> New changeset 619331c67638 by Richard Oudkerk in branch '3.4':
> Issue #20990: Fix issues found by pyflakes for multiprocessing.
> http://hg.python.org/cpython/rev/619331c67638
>
> ----------
> nosy: +python-dev
>
> _______________________________________
> Python tracker <report@bugs.python.org <javascript:;>>
> <http://bugs.python.org/issue20990>
> _______________________________________
>

New changeset bb6377db0a9e by Richard Oudkerk in branch '3.4':
Issue #20990: Correction for 619331c67638.
http://hg.python.org/cpython/rev/bb6377db0a9e

Can we close this issue? Or Richard wants to write a test?

Testing the is_forking() requires cx_freeze or something similar, so it really cannot go in the test suite.

I have tested it manually (after spending too long trying to get cx_freeze to work with a source build).

It should be noted that on Unix freezing is currently only compatible with the default 'fork' start method.