test_sigdumpmem causes test suite to hang

[Aaron Bentley]

When I run "bin/test -t test_sigdumpmem -t testAppServerIsAvailable", the second test fails and the test suite hangs. This is on my Karmic VM. The hang is obviously related to bugs in meliae, which is presumably triggered by test_sigdumpmem.

I cannot reproduce this bug on my normal machine (non-VM) running lucid, but I can't run the test suite on lucid because of the problems with pygpgme.

[Aaron Bentley (abentley)]

$ bin/test -v --layer AppServerLayer
Running tests at level 1
Running canonical.testing.layers.AppServerLayer tests:
  Set up canonical.testing.layers.BaseLayer in 0.012 seconds.
  Set up canonical.testing.layers.DatabaseLayer in 4.721 seconds.
  Set up canonical.testing.layers.LibrarianLayer in 11.588 seconds.
  Set up canonical.testing.layers.MemcachedLayer in 0.196 seconds.
  Set up canonical.testing.layers.LaunchpadLayer in 0.000 seconds.
  Set up canonical.testing.layers.FunctionalLayer in 6.327 seconds.
  Set up canonical.testing.layers.LaunchpadFunctionalLayer in 0.000 seconds.
  Set up canonical.testing.layers.AppServerLayer in 15.839 seconds.
  Running:
...................................................................................................................................

Error in test testAppServerIsAvailable (canonical.testing.ftests.test_layers.LayerProcessControllerInvariantsTestCase)
Traceback (most recent call last):
  File "/usr/lib/python2.5/unittest.py", line 260, in run
    testMethod()
  File "/home/abentley/launchpad/recipe-index/lib/canonical/testing/ftests/test_layers.py", line 367, in testAppServerIsAvailable
    home_page = urlopen(mainsite.rooturl).read()
  File "/usr/lib/python2.5/urllib.py", line 82, in urlopen
    return opener.open(url)
  File "/usr/lib/python2.5/urllib.py", line 190, in open
    return getattr(self, name)(url)
  File "/usr/lib/python2.5/urllib.py", line 325, in open_http
    h.endheaders()
  File "/usr/lib/python2.5/httplib.py", line 860, in endheaders
    self._send_output()
  File "/usr/lib/python2.5/httplib.py", line 732, in _send_output
    self.send(msg)
  File "/usr/lib/python2.5/httplib.py", line 699, in send
    self.connect()
  File "/usr/lib/python2.5/httplib.py", line 683, in connect
    raise socket.error, msg
IOError: [Errno socket error] (111, 'Connection refused')

Traceback (most recent call last):
  File "bin/test", line 259, in <module>
    result = testrunner.run([])
  File "/home/abentley/launchpad/stable/eggs/zope.testing-3.8.1-py2.5.egg/zope/testing/testrunner/__init__.py", line 32, in run
    failed = run_internal(defaults, args, script_parts=script_parts)
  File "/home/abentley/launchpad/stable/eggs/zope.testing-3.8.1-py2.5.egg/zope/testing/testrunner/__init__.py", line 45, in run_internal
    runner.run()
  File "/home/abentley/launchpad/stable/eggs/zope.testing-3.8.1-py2.5.egg/zope/testing/testrunner/runner.py", line 136, in run
    self.run_tests()
  File "/home/abentley/launchpad/stable/eggs/zope.testing-3.8.1-py2.5.egg/zope/testing/testrunner/runner.py", line 216, in run_tests
    setup_layers, self.failures, self.errors)
  File "/home/abentley/launchpad/stable/eggs/zope.testing-3.8.1-py2.5.egg/zope/testing/testrunner/runner.py", line 374, in run_layer
    return run_tests(options, tests, layer_name, failures, errors)
  File "/home/abentley/launchpad/stable/eggs/zope.testing-3.8.1-py2.5.egg/zope/testing/testrunner/runner.py", line 306, in run_tests
    test(result)
  File "/usr/lib/python2.5/unittest.py", line 281, in __call__
    return self.run(*args, **kwds)
  File "/usr/lib/python2.5/unittest.py", line 278, in run
    result.stopTest(self)
  File "/ho...

[Guilherme Salgado (salgado)]

Before disabling a test that might be causing problems on Karmic VMs but
not on Hardy/Lucid, I think we should get the pygpgme tests to pass on
Lucid because we're going to need that before we can upgrade our
ec2/buildbot AMIs to Lucid.

In fact, someone has done that already, and we just need to find out why
it's stalled. https://code.edge.launchpad.net/~jelmer/pygpgme/bug452194

[Aaron Bentley (abentley)]

On 04/07/2010 03:40 PM, Guilherme Salgado wrote:
> Before disabling a test that might be causing problems on Karmic VMs but
> not on Hardy/Lucid, I think we should get the pygpgme tests to pass on
> Lucid because we're going to need that before we can upgrade our
> ec2/buildbot AMIs to Lucid.

I disagree. I think that having a working test suite is essential. We
know how to get that right away. We don't know how to get pygpgme
working right away.

> In fact, someone has done that already, and we just need to find out why
> it's stalled. https://code.edge.launchpad.net/~jelmer/pygpgme/bug452194

I am all in favour of fixing pygpgme, but not *before* I can have a
working development environment. The fact is that there's been plenty
of time for someone to apply jelmer's fix, and it hasn't happened yet.

Aaron

[Edwin Grubbs (edwin-grubbs)]

This also breaks for me running in a VM, but I'm running Lucid instead of Karmic. You probably know this already, but it looks like the problem is caused by id() returning a number that doesn't fit in a signed 4-byte integer and meliae stores it in a C python module so it can't handle python long ints. I imagine that this test could break on other systems at random in the future.

[Guilherme Salgado (salgado)]

The test has been disabled, so we now need to find out what's needed to make it pass on Karmic VMs so that it can be re-enabled.

[Ursula Junque (ursinha)]

Fixed in stable r10649 <http://bazaar.launchpad.net/~launchpad-pqm/launchpad/stable/revision/10649>