Failed selftest 1.13rc1

[Martin Pool]

2009/3/13 Matt Nordhoff <mnordhoff at mattnordhoff.com>:
> Müller-Reineke wrote:
>> Dear bzr list,
>>
>> "bzr selftest" told me "FAILED (failures=26, errors=3, known_failure_count=27)" on Ubuntu 8.10 and
>> "FAILED (failures=3, errors=1, known_failure_count=14)" on SuSE SLES 9 (full output is attached to this mail).
>> Is that OK or is it dangerous to use bzr-1.13rc1? (Thats no rhetorical question.)
>> This confuses me because
>> http://doc.bazaar-vcs.org/bzr.dev/en/developer-guide/HACKING.html#bazaar-development-in-a-nutshell
>> claims "automated validation that all tests still pass before code is merged into the main code branch.".

So we do have automated validation, but with such a large test suite,
there are some variations depending on the environment in which it is
run.  Some at least of these can be considered bugs in the test suite.

Some of the tests assume they're run in an English environment because
they examine error messages.  It would be better if this requirement
was explicit in the tests that needed it.   However, if there are
really so few tests failing when it's run in german, maybe we should
just teach the tests to use strerror and therefore pass.

FAIL: test_osutils.TestWalkDirs.test_walkdirs_os_error
    not equal:
a = "[Errno 13] Keine Berechtigung: './test-unreadable'"
b = "[Errno 13] chdir: Permission denied: './test-unreadable'"

This is the locale thing, plus a spurious failure that was just fixed
when the extensions have not been built.  (Run 'make'.)

======================================================================
ERROR: test_send_broken_body_stream
(bzrlib.tests.test_smart_transport.TestResponseEncodingProtocolThree)

vvvv[log from bzrlib.tests.test_smart_transport.TestResponseEncodingProtocolThree.test_send_broken_body_stream]

^^^^[log from bzrlib.tests.test_smart_transport.TestResponseEncodingProtocolThree.test_send_broken_body_stream]
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/local/Python-2.5/lib/python2.5/site-packages/bzrlib/tests/test_smart_transport.py",
line 2831, in test_send_broken_body_stream
    encoder.send_response(response)
  File "/usr/local/Python-2.5/lib/python2.5/site-packages/bzrlib/smart/protocol.py",
line 1166, in send_response
    error_struct = request._translate_error(exc_info[1])
  File "/usr/local/Python-2.5/lib/python2.5/site-packages/bzrlib/smart/request.py",
line 348, in _translate_error
    trace.log_exception_quietly()
  File "/usr/local/Python-2.5/lib/python2.5/site-packages/bzrlib/trace.py",
line 354, in log_exception_quietly
    mutter(traceback.format_exc())
  File "/usr/local/Python-2.5/lib/python2.5/traceback.py", line 236,
in format_exc
    return ''.join(format_exception(etype, value, tb, limit))
  File "/usr/local/Python-2.5/lib/python2.5/traceback.py", line 145,
in format_exception
    list = list + format_exception_only(etype, value)
  File "/usr/local/Python-2.5/lib/python2.5/traceback.py", line 176,
in format_exception_only
    stype = etype.__name__
AttributeError: 'NoneType' object has no attribute '__name__'

This seems to be a new bug but I can't reproduce it.  Thanks for reporting it.

<https://bugs.edge.launchpad.net/bzr/+bug/342140>

-- 
Martin <http://launchpad.net/~mbp/>