Author jednaszewski
Recipients csernazs, dagobert, doughellmann, gregory.p.smith, jednaszewski, loewis, pitrou, twouters
Date 2010-02-26.20:27:32
SpamBayes Score 1.8436e-11
Marked as misclassified No
Message-id <1267216055.13.0.620310385603.issue7242@psf.upfronthosting.co.za>
In-reply-to
Content
I spent some time working on and testing a unit test as well.  It's the same basic idea as Zsolt Cserna's, but with a slightly different approach.  See 7242_unittest.diff.  My unittest fails pre-patch and succeeds post-patch.

However, I still have reservations about the patch.  The existing test test_threading.ThreadJoinOnShutdown.test_3_join_in_forked_from_thread hangs with the patch in place.

Vanilla 2.6.2 - test passes
Vanilla 2.6.4 - test fails
Patched 2.6.4 - test hangs

Note: the code of the test_threading test is identical in all 3 cases.  I'd feel more confident about the patch if this test didn't hang with the patch in place.
History
Date User Action Args
2010-02-26 20:27:35jednaszewskisetrecipients: + jednaszewski, loewis, twouters, gregory.p.smith, csernazs, pitrou, doughellmann, dagobert
2010-02-26 20:27:35jednaszewskisetmessageid: <1267216055.13.0.620310385603.issue7242@psf.upfronthosting.co.za>
2010-02-26 20:27:33jednaszewskilinkissue7242 messages
2010-02-26 20:27:33jednaszewskicreate