[MERGE] [Bug 226769] Disable some strace tests

[John Arbash Meinel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Vincent Ladeuil wrote:
|>>>>> "vila" == Vincent Ladeuil <v.ladeuil+lp at free.fr> writes:
|
|>>>>> "robert" == Robert Collins <robertc at robertcollins.net> writes:
|     vila> <snip/>
|
|     robert> Lets merge a patch which collects statistics on leaked threads from
|     robert> tests, and reports at the end:
|     robert> - the number of tests where the thread count changed
|     robert> - the name of the first such test
|
|     vila> Ideally we need a setUp/tearDown on the whole test suite for that
|     vila> (and even more ideally some addCleanUp facility too).
|
|     vila> But I'd prefer not implementing that right now but instead go the
|     vila> _make_test_root way and use atexit too.
|
|     vila> Thoughts ?
|
| Patch :)
|
|       Vincent
|
|

I was going to say that we should hook into the same logic that we use for
cleaning up the test suite, rather than using 'atexit'. However, looking at the
code:


def _make_test_root(self):
~    if TestCaseWithMemoryTransport.TEST_ROOT is None:
~        root = osutils.mkdtemp(prefix='testbzr-', suffix='.tmp')
~        TestCaseWithMemoryTransport.TEST_ROOT = root

~        self._create_safety_net()

~        # The same directory is used by all tests, and we're not
~        # specifically told when all tests are finished.  This will do.
~        atexit.register(_rmtree_temp_dir, root)

~    self.addCleanup(self._check_safety_net)

So... atexit() it is.

+    # If the following tests are activated, selftest may hang (see bug
+    # #226769). This is due to strace strange behavior when required to trace
+    # its own parent in the presence of threads (or something like that). One
+    # strace is fixed, we may want to activate these tests again. Note: running
+    # these tests in isolation is still possible.
+

^- This comment is out of date. The tests are being run if the activeCount() ==
1. So the comment needs to be updated to describe the strace bug, and say why we
are XFAILing the tests.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkghvVwACgkQJdeBCYSNAAP4QgCfVndgAWPmd+VJXSYf7j7ZEnIT
D7wAnjTzaFqjLp1Fa4KJK6ArGfVaqJkT
=qMBI
-----END PGP SIGNATURE-----