[Python-checkins] cpython: Issue #15794: Relax a test case due to the deadlock detection's

[Brett Cannon]

Should there be a Misc/NEWS entry since we are in rc mode?

On Tue, Aug 28, 2012 at 2:13 PM, antoine.pitrou
<python-checkins at python.org>wrote:

> http://hg.python.org/cpython/rev/454dceb5fd56
> changeset:   78790:454dceb5fd56
> parent:      78788:06497bbdf4fe
> user:        Antoine Pitrou <solipsis at pitrou.net>
> date:        Tue Aug 28 20:10:18 2012 +0200
> summary:
>   Issue #15794: Relax a test case due to the deadlock detection's
> conservativeness.
>
> files:
>   Lib/test/test_importlib/test_locks.py |  22 ++++++++++++--
>   1 files changed, 18 insertions(+), 4 deletions(-)
>
>
> diff --git a/Lib/test/test_importlib/test_locks.py
> b/Lib/test/test_importlib/test_locks.py
> --- a/Lib/test/test_importlib/test_locks.py
> +++ b/Lib/test/test_importlib/test_locks.py
> @@ -1,4 +1,5 @@
>  from importlib import _bootstrap
> +import sys
>  import time
>  import unittest
>  import weakref
> @@ -41,6 +42,17 @@
>  @unittest.skipUnless(threading, "threads needed for this test")
>  class DeadlockAvoidanceTests(unittest.TestCase):
>
> +    def setUp(self):
> +        try:
> +            self.old_switchinterval = sys.getswitchinterval()
> +            sys.setswitchinterval(0.000001)
> +        except AttributeError:
> +            self.old_switchinterval = None
> +
> +    def tearDown(self):
> +        if self.old_switchinterval is not None:
> +            sys.setswitchinterval(self.old_switchinterval)
> +
>      def run_deadlock_avoidance_test(self, create_deadlock):
>          NLOCKS = 10
>          locks = [LockType(str(i)) for i in range(NLOCKS)]
> @@ -75,10 +87,12 @@
>
>      def test_deadlock(self):
>          results = self.run_deadlock_avoidance_test(True)
> -        # One of the threads detected a potential deadlock on its second
> -        # acquire() call.
> -        self.assertEqual(results.count((True, False)), 1)
> -        self.assertEqual(results.count((True, True)), len(results) - 1)
> +        # At least one of the threads detected a potential deadlock on its
> +        # second acquire() call.  It may be several of them, because the
> +        # deadlock avoidance mechanism is conservative.
> +        nb_deadlocks = results.count((True, False))
> +        self.assertGreaterEqual(nb_deadlocks, 1)
> +        self.assertEqual(results.count((True, True)), len(results) -
> nb_deadlocks)
>
>      def test_no_deadlock(self):
>          results = self.run_deadlock_avoidance_test(False)
>
> --
> Repository URL: http://hg.python.org/cpython
>
> _______________________________________________
> Python-checkins mailing list
> Python-checkins at python.org
> http://mail.python.org/mailman/listinfo/python-checkins
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-checkins/attachments/20120828/f557eb62/attachment.html>