Bug #314247 “bzr tests fail on OpenBSD 4.4/amd64” : Bugs : Bazaar

Revision history for this message

Martin Pool (mbp) wrote on 2009-01-06:

#1

test.log Edit (2.7 MiB, text/plain)

Revision history for this message

Martin Pool (mbp) wrote on 2009-01-06:

#2

bzr-testing-tmp.cpio Edit (110.0 KiB, application/x-cpio)

contents of tmp after test failure

Revision history for this message

Martin Pool (mbp) wrote on 2009-01-06:

#3

bzr.log Edit (5.6 KiB, text/plain)

.bzr.log from Tobi

Revision history for this message

Martin Pool (mbp) wrote on 2009-01-06: Re: bzr 1.10 selftests

#4

Download full text (19.9 KiB)

On 2 Jan 2009, Toni Mueller <email address hidden> wrote:
>
> Hi Martin,
>
> On Mon, 29.12.2008 at 19:24:02 +0100, Toni Mueller <email address hidden> wrote:
> > I tried to run the self-test suite for bzr-1.10 on OpenBSD 4.4/amd64,
>
> I've re-tried, and got the results that you find attached. I didn't
> check whether all tests were actually performed, but simply ship
> everything that the tests yieled, to you. The test run (probably) ended
> when Python (2.5.2p4) segfaulted. I have the core lying around, so if
> you want me to do some surgery on it, just ask for it (I don't really
> know what to do). I created the "...cpio.." archive using "pax -x
> sv4cpio". Rolling a tarball was impossible because of the 100 char
> pathname limit in the tar archive format.

Hi,

Thanks for testing this. I opened
<https://bugs.edge.launchpad.net/bzr/+bug/314247> to make the results
available to other people.

The problems here seem to be:

Firstly, if bzr selftest doesn't run correctly without a terminal or in
nohup mode, that seems to be a bug in itself. Run under Linux with
nohup it does not write any output to the log file while it's running, I
guess because it's relying on a progress bar and that will be off if
there's no terminal, but I do get a list of failed, xfail etc tests at
the end of the run.

On Ubuntu Intrepid all the tests pass in a nohup environment pass except
for some problems in bzr-gtk.

bzr is supposed to delete temporary directories as it finishes each
test, so it shouldn't be using a large amount of disk space to run. It
is possible that particular tests will need several megabytes. As I'm
running the tests I see all of /tmp is about 12MB. If you can isolate
what is using too much space we may find a bug. Otherwise you can run
them with TMPDIR=~/tmp to put it on another filesystem.

So the substantial bugs in here are:

> blackbox.test_breakin.TestBreakin.test_breakin FAIL 619ms
> pattern "entering debugger" not found in
> """\
> No handlers could be found for logger "bzr"
> """

The logger message might be a problem with not having a console.
breakin.py seems to unconditionally, write to stderr so should not be
affected by that. It would be useful if you could try pressing C-\
while bzr is running normally, to see if there's a difference in sigquit
handling on openbsd.

> blackbox.test_check.ChrootedCheckTests.test_check_missing_branch
> ERROR 48ms
>
> (26, 'Failed to open/read local data from file/application')

> ^^^^[log from bzrlib.tests.blackbox.test_check.ChrootedCheckTests.test_check_missing_branch]
> ----------------------------------------------------------------------
> Traceback (most recent call last):
> File "/home/sw/local/bzr-1.10/bzrlib/tests/blackbox/test_check.py", line 104, in test_check_missing_branch
> 'check --branch %s' % self.get_readonly_url(''))
> File "/home/sw/local/bzr-1.10/bzrlib/tests/__init__.py", line 1486, in run_bzr
> working_dir=working_dir,
> File "/home/sw/local/bzr-1.10/bzrlib/tests/__init__.py", line 1400, in _run_bzr_autosplit
> encoding=encoding, stdin=stdin, w...

On  2 Jan 2009, Toni Mueller <support@oeko.net> wrote:
> 
> Hi Martin,
> 
> On Mon, 29.12.2008 at 19:24:02 +0100, Toni Mueller <support@oeko.net> wrote:
> > I tried to run the self-test suite for bzr-1.10 on OpenBSD 4.4/amd64,
> 
> I've re-tried, and got the results that you find attached. I didn't
> check whether all tests were actually performed, but simply ship
> everything that the tests yieled, to you. The test run (probably) ended
> when Python (2.5.2p4) segfaulted. I have the core lying around, so if
> you want me to do some surgery on it, just ask for it (I don't really
> know what to do). I created the "...cpio.." archive using "pax -x
> sv4cpio". Rolling a tarball was impossible because of the 100 char
> pathname limit in the tar archive format.

Hi,

Thanks for testing this.  I opened
<https://bugs.edge.launchpad.net/bzr/+bug/314247> to make the results
available to other people.

The problems here seem to be:

Firstly, if bzr selftest doesn't run correctly without a terminal or in
nohup mode, that seems to be a bug in itself.  Run under Linux with
nohup it does not write any output to the log file while it's running, I
guess because it's relying on a progress bar and that will be off if
there's no terminal, but I do get a list of failed, xfail etc tests at
the end of the run.

On Ubuntu Intrepid all the tests pass in a nohup environment pass except
for some problems in bzr-gtk.

bzr is supposed to delete temporary directories as it finishes each
test, so it shouldn't be using a large amount of disk space to run.  It
is possible that particular tests will need several megabytes.  As I'm
running the tests I see all of /tmp is about 12MB.  If you can isolate
what is using too much space we may find a bug.  Otherwise you can run
them with TMPDIR=~/tmp to put it on another filesystem.

So the substantial bugs in here are:

> blackbox.test_breakin.TestBreakin.test_breakin                                                                       FAIL                 619ms
>     pattern "entering debugger" not found in
> """\
> No handlers could be found for logger "bzr"
> """

The logger message might be a problem with not having a console. 
breakin.py seems to unconditionally, write to stderr so should not be
affected by that.  It would be useful if you could try pressing C-\
while bzr is running normally, to see if there's a difference in sigquit
handling on openbsd.

> blackbox.test_check.ChrootedCheckTests.test_check_missing_branch
> ERROR                  48ms
> 
>    (26, 'Failed to open/read local data from file/application')

> ^^^^[log from bzrlib.tests.blackbox.test_check.ChrootedCheckTests.test_check_missing_branch]
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/home/sw/local/bzr-1.10/bzrlib/tests/blackbox/test_check.py", line 104, in test_check_missing_branch
>     'check --branch %s' % self.get_readonly_url(''))
>   File "/home/sw/local/bzr-1.10/bzrlib/tests/__init__.py", line 1486, in run_bzr
>     working_dir=working_dir,
>   File "/home/sw/local/bzr-1.10/bzrlib/tests/__init__.py", line 1400, in _run_bzr_autosplit
>     encoding=encoding, stdin=stdin, working_dir=working_dir,
>   File "/home/sw/local/bzr-1.10/bzrlib/tests/__init__.py", line 1430, in _run_bzr_core
>     args)
>   File "/home/sw/local/bzr-1.10/bzrlib/tests/__init__.py", line 1719, in apply_redirected
>     return a_callable(*args, **kwargs)
>   File "/home/sw/local/bzr-1.10/bzrlib/commands.py", line 912, in run_bzr_catch_user_errors
>     return run_bzr(argv)
>   File "/home/sw/local/bzr-1.10/bzrlib/commands.py", line 839, in run_bzr
>     ret = run(*run_argv)
>   File "/home/sw/local/bzr-1.10/bzrlib/commands.py", line 539, in run_argv_aliases
>     return self.run(**all_cmd_args)
>   File "/home/sw/local/bzr-1.10/bzrlib/builtins.py", line 2513, in run
>     check_dwim(path, verbose, do_branch=branch, do_repo=repo, do_tree=tree)
>   File "/home/sw/local/bzr-1.10/bzrlib/check.py", line 275, in check_dwim
>     BzrDir.open_containing_tree_branch_or_repository(path)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 919, in open_containing_tree_branch_or_repository
>     bzrdir, relpath = klass.open_containing(location)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 828, in open_containing
>     return BzrDir.open_containing_from_transport(transport)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 849, in open_containing_from_transport
>     result = BzrDir.open_from_transport(a_transport)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 803, in open_from_transport
>     redirected)
>   File "/home/sw/local/bzr-1.10/bzrlib/transport/__init__.py", line 1643, in do_catching_redirections
>     return action(transport)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 780, in find_format
>     transport, _server_formats=_server_formats)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 1638, in find_format
>     return format.probe_transport(transport)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 2621, in probe_transport
>     server_version = medium.protocol_version()
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 593, in protocol_version
>     client_protocol.query_version()
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/protocol.py", line 769, in query_version
>     self.call('hello')
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/protocol.py", line 617, in call
>     self._request.finished_writing()
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 421, in finished_writing
>     self._finished_writing()
>   File "/home/sw/local/bzr-1.10/bzrlib/transport/http/__init__.py", line 575, in _finished_writing
>     data = self._medium.send_http_smart_request(self._buffer)
>   File "/home/sw/local/bzr-1.10/bzrlib/transport/http/__init__.py", line 552, in send_http_smart_request
>     code, body_filelike = t._post(bytes)
>   File "/home/sw/local/bzr-1.10/bzrlib/transport/http/_pycurl.py", line 267, in _post
>     self._curl_perform(curl, header, ['Expect: '])
>   File "/home/sw/local/bzr-1.10/bzrlib/transport/http/_pycurl.py", line 341, in _curl_perform
>     curl.perform()
> error: (26, 'Failed to open/read local data from file/application')
>

This occurs in many of the tests, and it looks like a problem in pycurl,
or in curl.  It might be best to uninstall it, or disable it by putting
a file named 'pycurl.py' containing 'raise ImportError()' in a directory
on your $PYTHONPATH.

> blackbox.test_diff.TestExternalDiff.test_external_diff                                                               FAIL                 495ms
>     not equal:
> a = ''
> b = "diff: unknown option -- binary\nusage: diff [-abdilpqTtw] [-I pattern] [-c | -e | -f | -n | -u]\n            [-L label] file1 file2\n       diff [-abdilpqTtw] [-I pattern] [-L label] -C number file1 file2\n       diff [-abdilqtw] [-I pattern] -D string file1 file2\n       diff [-abdilpqTtw] [-I pattern] [-L label] -U number file1 file2\n       diff [-abdilNPpqrsTtw] [-I pattern] [-c | -e | -f | -n | -u]\n            [-L label] [-S name] [-X file] [-x pattern] dir1 dir2\nbzr: ERROR: external diff failed with exit code 2; command: ['diff', '--label', u'goodbye\\t1970-01-01 00:00:00 +0000', '/tmp/bzr-diff-old-D3V3cy', '--label', u'goodbye\\t2008-12-31 00:10:49 +0000', '/tmp/bzr-diff-new-laJYMs', '--binary', '-ub']\n"
> 
>

We seem to be unnecessarily relying on gnu diff here in
bzrlib.diff.external_diff, and on bsd at least we should not.

> blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir
> ERROR                 344ms
>    [Errno 2] No such file or directory: '..'

> ======================================================================
> ERROR: test_remove_changed_files_from_child_dir (bzrlib.tests.blackbox.test_remove.TestRemove)
> 
> vvvv[log from bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir]
> 546.905  creating repository in file:///tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work/.bzr/.
> 546.942  creating branch <bzrlib.branch.BzrBranchFormat6 object at 0x21c09b750> in file:///tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work/.bzr/
> 546.993  trying to create missing lock '/tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work/.bzr/checkout/dirstate'
> 546.994  opening working tree '/tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work'
> 547.047  run bzr: ['commit', '-m', 'added files']
> 547.048  bzr arguments: ['commit', '-m', 'added files']
> 547.049  encoding stdout as sys.stdout encoding '646'
> 547.052  opening working tree '/tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work'
> 547.068  preparing to commit
>     INFO  Committing to: /tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work/
> 547.072  Selecting files for commit with filter []
>     INFO  added a
>     INFO  added b
>     INFO  added d
>     INFO  added b/c
>     INFO  Committed revision 1.
> 547.128  errors:
> 'Committing to: /tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work/\nadded a\nadded b\nadded d\nadded b/c\nCommitted revision 1.\n'
> 547.129  run bzr: ['remove', '../a', 'c', '.', '../d']
> 547.129  bzr arguments: ['remove', '../a', 'c', '.', '../d']
> 547.130  encoding stdout as sys.stdout encoding '646'
> 547.133  opening working tree '/tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work'
> 547.164  Traceback (most recent call last):
>   File "/home/sw/local/bzr-1.10/bzrlib/commands.py", line 912, in run_bzr_catch_user_errors
>     return run_bzr(argv)
>   File "/home/sw/local/bzr-1.10/bzrlib/commands.py", line 839, in run_bzr
>     ret = run(*run_argv)
>   File "/home/sw/local/bzr-1.10/bzrlib/commands.py", line 539, in run_argv_aliases
>     return self.run(**all_cmd_args)
>   File "/home/sw/local/bzr-1.10/bzrlib/builtins.py", line 1239, in run
>     force=file_deletion_strategy=='force')
>   File "/home/sw/local/bzr-1.10/bzrlib/mutabletree.py", line 52, in tree_write_locked
>     return unbound(self, *args, **kwargs)
>   File "/home/sw/local/bzr-1.10/bzrlib/workingtree.py", line 1932, in remove
>     raise errors.BzrRemoveChangedFilesError(tree_delta)
> BzrRemoveChangedFilesError: Can't safely remove modified or unknown files:
> modified:
>   a
>   b/c
> Use --keep to not delete them, or --force to delete them regardless.
> 
> 547.164  errors:
> "bzr: ERROR: Can't safely remove modified or unknown files:\nmodified:\n  a\n  b/c\nUse --keep to not delete them, or --force to delete them regardless.\n"
> 547.164  run bzr: ['remove', '--force', '../a', 'c', '.', '../d']
> 547.164  bzr arguments: ['remove', '--force', '../a', 'c', '.', '../d']
> 547.165  encoding stdout as sys.stdout encoding '646'
> 547.168  opening working tree '/tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir/work'
>     INFO  deleted d
>     INFO  deleted b/c
>     INFO  deleted b
>     INFO  deleted a
> 547.206  errors:
> 'deleted d\ndeleted b/c\ndeleted b\ndeleted a\n'
> 547.257  opening working tree '/tmp/testbzr-qdLfKY.tmp'
> 
> ^^^^[log from bzrlib.tests.blackbox.test_remove.TestRemove.test_remove_changed_files_from_child_dir]
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/home/sw/local/bzr-1.10/bzrlib/tests/blackbox/test_remove.py", line 202, in test_remove_changed_files_from_child_dir
>     os.chdir('..')
> OSError: [Errno 2] No such file or directory: '..'
> 
> 
>

This seems like a genuine portability bug, that once a directory has
been deleted on bsd you can no longer see it's .. entry.  On Linux I
think the directory remains alive in memory while it's in use.  We
should fix the test.

> blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh                                                        ERROR                 407ms
>     Connection closed: please check connectivity and permissions (and try -Dhpss if further diagnosis is required)

ERROR: test_bzr_connect_to_bzr_ssh
(bzrlib.tests.blackbox.test_serve.TestBzrServe)

> vvvv[log from
> bzrlib.tests.blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh]
> 611.394  creating repository in
> file:///tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh/work/a_branch/.bzr/.
> 611.430  creating branch <bzrlib.branch.BzrBranchFormat6 object at
> 0x21c005d50> in
> file:///tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh/work/a_branch/.bzr/
> 611.466  failed to load system host keys: [Errno 2] No such file or
> directory:
> '/tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh/home/.ssh/known_hosts'
> 611.466  failed to load bzr host keys: [Errno 2] No such file or
> directory:
> '/tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh/home/.bazaar/ssh_host_keys'
> 611.466  creating config directory:
> '/tmp/testbzr-qdLfKY.tmp/bzrlib.tests.blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh/home/.bazaar'
>    DEBUG  starting thread (client mode): 0x1c1baa90L
>     INFO  Connected (version 2.0, client paramiko_1.6.2)
>    DEBUG  kex algos:['diffie-hellman-group1-sha1'] server
> key:['ssh-rsa'] client encrypt:['aes128-cbc', 'blowfish-cbc',
> 'aes256-cbc', '3des-cbc'] server encrypt:['aes128-cbc', 'blowfish-cbc',
> 'aes256-cbc', '3des-cbc'] client mac:['hmac-sha1', 'hmac-md5',
> 'hmac-sha1-96', 'hmac-md5-96'] server mac:['hmac-sha1', 'hmac-md5',
> 'hmac-sha1-96', 'hmac-md5-96'] client compress:['none'] server
> compress:['none'] client lang:[''] server lang:[''] kex follows?False
>    DEBUG  Ciphers agreed: local=aes128-cbc, remote=aes128-cbc
>    DEBUG  using kex diffie-hellman-group1-sha1; server key type ssh-rsa;
> cipher: local aes128-cbc, remote aes128-cbc; mac: local hmac-sha1,
> remote hmac-sha1; compression: local none, remote none
>    DEBUG  Switch to new keys ...
>  WARNING  Adding ssh-rsa host key for localhost:
> 60733844CB5186657FDEDAA22B5A57D5
>    DEBUG  userauth is OK
>     INFO  Authentication (password) successful!
>    DEBUG  Max packet in: 34816 bytes
>    DEBUG  Max packet out: 34816 bytes
>     INFO  Secsh channel 1 opened.
>    DEBUG  Sesch channel 1 request ok
>    DEBUG  EOF received
>    DEBUG  EOF sent
> 611.925  opening working tree '/tmp/testbzr-qdLfKY.tmp'
> 
> ^^^^[log from
> bzrlib.tests.blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh]
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/home/sw/local/bzr-1.10/bzrlib/tests/blackbox/test_serve.py",
> line 213, in test_bzr_connect_to_bzr_ssh
>     'bzr+ssh://fred:secret@localhost:%d%s' % (port, path_to_branch))
>   File "/home/sw/local/bzr-1.10/bzrlib/branch.py", line 123, in open
>     possible_transports=possible_transports)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 766, in open
>     return BzrDir.open_from_transport(t, _unsupported=_unsupported)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 808, in
> open_from_transport
>     return format.open(transport, _found=True)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 1770, in open
>     return self._open(transport)
>   File "/home/sw/local/bzr-1.10/bzrlib/bzrdir.py", line 2646, in _open
>     return remote.RemoteBzrDir(transport)
>   File "/home/sw/local/bzr-1.10/bzrlib/remote.py", line 93, in __init__
>     response = self._call('BzrDir.open', path)
>   File "/home/sw/local/bzr-1.10/bzrlib/remote.py", line 51, in _call
>     return self._client.call(method, *args)
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/client.py", line 122, in
> call
>     result, protocol = self.call_expecting_body(method, *args)
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/client.py", line 135, in
> call_expecting_body
>     method, args, expect_response_body=True)
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/client.py", line 83, in
> _call_and_read_response
>     expect_body=expect_response_body)
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/message.py", line 258, in
> read_response_tuple
>     self._wait_for_response_args()
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/message.py", line 224, in
> _wait_for_response_args
>     self._read_more()
>   File "/home/sw/local/bzr-1.10/bzrlib/smart/message.py", line 247, in
> _read_more
>     "(and try -Dhpss if further diagnosis is required)")
> ConnectionReset: Connection closed: please check connectivity and
> permissions (and try -Dhpss if further diagnosis is required)
>

It's not obvious to me why this failed; perhaps an unexpected error on
the server side.

> blackbox.test_serve.TestBzrServe.test_bzr_serve_port_readonly                                                        FAIL                 140ms
>     not equal:
> a = 'bzr: interrupted\n'
> b = '<bzrlib.smart.medium.SmartServerSocketStreamMedium object at 0x209201590> terminating on exception (4, \'Interrupted system call\')\nException in thread smart-server-child:\nTraceback (most recent call last):\n  File "/usr/local/lib/python2.5/threading.py", line 486, in __bootstrap_inner\n    self.run()\n  File "/usr/local/lib/python2.5/threading.py", line 446, in run\n    self.__target(*self.__args, **self.__kwargs)\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 202, in serve\n    server_protocol = self._build_protocol()\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 217, in _build_protocol\n    bytes = self._get_line()\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 161, in _get_line\n    line, excess = _get_line(self.read_bytes)\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 99, in _get_line\n    new_bytes = read_bytes_func(1)\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 148, in read_bytes\n    return self._read_bytes(bytes_to_read)\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 277, in _read_bytes\n    return self.socket.recv(_MAX_READ_SIZE)\nerror: (4, \'Interrupted system call\')\n\nbzr: interrupted\n'
> 
> blackbox.test_serve.TestBzrServe.test_bzr_serve_port_readwrite                                                       FAIL                 270ms
>     not equal:
> a = 'bzr: interrupted\n'
> b = '<bzrlib.smart.medium.SmartServerSocketStreamMedium object at 0x20f639a50> terminating on exception (4, \'Interrupted system call\')\nException in thread smart-server-child:\nTraceback (most recent call last):\n  File "/usr/local/lib/python2.5/threading.py", line 486, in __bootstrap_inner\n    self.run()\n  File "/usr/local/lib/python2.5/threading.py", line 446, in run\n    self.__target(*self.__args, **self.__kwargs)\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 202, in serve\n    server_protocol = self._build_protocol()\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 217, in _build_protocol\n    bytes = self._get_line()\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 161, in _get_line\n    line, excess = _get_line(self.read_bytes)\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 99, in _get_line\n    new_bytes = read_bytes_func(1)\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 148, in read_bytes\n    return self._read_bytes(bytes_to_read)\n  File "/home/sw/local/bzr-1.10/bzrlib/smart/medium.py", line 277, in _read_bytes\n    return self.socket.recv(_MAX_READ_SIZE)\nerror: (4, \'Interrupted system call\')\n\nbzr: interrupted\n'
>

This seems to be because we have some io code that should retry on
eintr, but that does not.  This could be a problem on any unix platform.

That's as far as I've got through the log, down to about line 800.

-- 
Martin

Revision history for this message

ToniMueller (support-oeko) wrote on 2009-01-06:

#5

Sorry for the confusion. This error:

> blackbox.test_breakin.TestBreakin.test_breakin FAIL 619ms
> pattern "entering debugger" not found in
> """\
> No handlers could be found for logger "bzr"
> """

and, in general, all error logs I sent you, were conducted in a second run where I called 'script' before, and just locked the terminal while the tests were running. IOW, a terminal should have been available at all times.

The problem with filling up /tmp I could not reproduce, but I ran the first test, where it happened, as user 'root', and both the ~/.bzr.log (how can I switch this off, please?) plus /tmp ate some 345 MB on my 400MB root partition (currently, there are only some 52 MB out of 400MB used). Also, tests showed a different failure pattern in that case, but I'm hesitant to do it again on that machine.

As should be obvious from the attached files, bzr fails to remove the temporary files and directories for tests that failed.

Revision history for this message

Martin Pool (mbp) wrote on 2009-01-07: Re: [Bug 314247] Re: bzr tests fail on OpenBSD 4.4/amd64

#6

On 6 Jan 2009, ToniMueller <email address hidden> wrote:
> The problem with filling up /tmp I could not reproduce, but I ran the
> first test, where it happened, as user 'root', and both the ~/.bzr.log
> (how can I switch this off, please?) plus /tmp ate some 345 MB on my
> 400MB root partition (currently, there are only some 52 MB out of 400MB
> used).

ln -s /dev/null ~/.bzr.log

should work

If that machine has more than 400MB disk in total, I suggest you set
TMPDIR to run it on a different partition.

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

Revision history for this message

ToniMueller (support-oeko) wrote on 2009-01-17:

#7

If you make any progress and can tell me what to test (eg. new code), I'd be happy to do so, and report back.

Revision history for this message

Martin Pool (mbp) wrote on 2009-03-24:

#8

This bug is languishing in part because it's too large and not a single problem. So if we can split it apart to individual classes of failure it should be easier.

The problem with eintr should now be fixed in 1.13.1 by Andrew - we retry on eintr.

The problem with options to diff is bug 314254.

Revision history for this message

Martin Pool (mbp) wrote on 2009-03-24:

#9

Please run the 1.13.1 tests, in a plain terminal, and let's see what happens.

Changed in bzr:
importance:	Undecided → Low

Martin Pool (mbp) on 2009-04-21

Changed in bzr:
importance:	Low → Medium
status:	New → Confirmed

Revision history for this message

Robert Collins (lifeless) wrote on 2009-09-25:

#10

Are there any remaining components to this bug?

Revision history for this message

Martin Pool (mbp) wrote on 2009-09-25:

#11

2009/9/25 Robert Collins <email address hidden>:
> Are there any remaining components to this bug?

There may well still be failures there but I don't know if retaining
this large bug will help fix them.

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

Vincent Ladeuil (vila) on 2012-03-12

tags:

added: selftest
removed: test

Jelmer Vernooij (jelmer) on 2017-11-08

tags:

added: check-for-breezy

Jelmer Vernooij (jelmer) on 2018-05-18

tags:

removed: check-for-breezy

Bazaar

bzr tests fail on OpenBSD 4.4/amd64

Bug Description

Other bug subscribers

Bug attachments

Remote bug watches