git code import fails: ValueError: invalid literal for int() with base 10: '--700'

[Guillermo Gonzalez]

A code import failed with this traceback, but it also fails with tip of dulwich and bzr-git locally.

The branch is: https://code.launchpad.net/~verterok/boto/trunk-git

2011-01-05 19:14:09 INFO Starting job.
2011-01-05 19:14:09 INFO Getting exising bzr branch from central store.
2011-01-05 19:14:09 INFO [chan bzr SocketAsChannelAdapter] Opened sftp connection (server version 3)
2011-01-05 19:14:10 INFO [chan bzr SocketAsChannelAdapter] Opened sftp connection (server version 3)
2011-01-05 19:14:10 INFO 51 bytes transferred
2011-01-05 19:14:10 INFO Importing foreign branch.
2011-01-05 19:14:11 INFO Counting objects: 9947, done. 0
2011-01-05 19:14:12 INFO finding revisions to fetch:generating index 0/9947
2011-01-05 19:14:13 INFO finding revisions to fetch 1/2689
2011-01-05 19:14:13 INFO
Traceback (most recent call last):
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/scripts/code-import-worker.py", line 82, in <module>
    sys.exit(script.main())
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/scripts/code-import-worker.py", line 77, in main
    return import_worker.run()
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/lib/lp/codehosting/codeimport/worker.py", line 452, in run
    return self._doImport()
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/lib/lp/codehosting/codeimport/worker.py", line 590, in _doImport
    overwrite=True, **self.getExtraPullArgs())
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/optionalbzrplugins/git/branch.py", line 514, in pull
    stop_revision, overwrite=overwrite, graph=graph, limit=limit)
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/optionalbzrplugins/git/branch.py", line 463, in _update_revisions
    determine_wants, self.source.mapping, limit=limit)
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/optionalbzrplugins/git/fetch.py", line 571, in fetch_objects
    objects_iter, store, wants_recorder.wants, pb, limit)
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/optionalbzrplugins/git/fetch.py", line 413, in import_git_objects
    o = lookup_object(head)
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/optionalbzrplugins/git/fetch.py", line 399, in lookup_object
    return object_iter[sha]
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/lib/dulwich/pack.py", line 1488, in __getitem__
    return ShaFile.from_raw_string(type, uncomp)
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/lib/dulwich/objects.py", line 370, in from_raw_string
    obj.set_raw_string(string)
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/lib/dulwich/objects.py", line 257, in set_raw_string
    self.set_raw_chunks([text])
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/lib/dulwich/objects.py", line 261, in set_raw_chunks
    self._deserialize(chunks)
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/lib/dulwich/objects.py", line 958, in _deserialize
    parse_timezone(timezonetext)
  File "/srv/importd.launchpad.net/production/launchpad-rev-12138/lib/dulwich/objects.py", line 888, in parse_timezone
    offset = int(text)
ValueError: invalid literal for int() with base 10: '--700'
Import failed:
Traceback (most recent call last):
Failure: twisted.internet.error.ProcessTerminated: A process has ended with a probable error condition: process ended with exit code 1.

[Aaron Bentley (abentley)]

Failed code import is at: https://code.launchpad.net/~verterok/boto/trunk-git

[Jelmer Vernooij (jelmer)]

This is a broken git repository. It has a timezone specification "--700" when it should probably be "-0700".

charis:~/tmp/boto% git fsck --full
error in commit c1eddff4ee3f62b6039f1083651b9118883e7f07: invalid author/committer line - bad time zone

Dulwich should warn about it in check() rather than error out.

[Guillermo Gonzalez (verterok)]

Hi jelmer,
thanks for the quick reply.

Is there something (some magic git command) I can request to the owner of the repo in order to fix it?

regards,

[Jelmer Vernooij (jelmer)]

Fixing it will probably involve doing a fastexport + fastimport of the repository.

Since this changes the timestamp field it will affect the SHA1s of all commits after the problematic commit too, so they may not want to do this.

[Guillermo Gonzalez (verterok)]

Oh, right..

probably I'll patch dulwich to handle this case and convert the --700 to -0700 and do a one time import into bzr.

Thanks!

[Jason R. Coombs (jaraco)]

This issue has bitten me when trying to make an hg clone of the canonical boto repo (git://github.com/boto/boto.git) using hg-git. It would be nice to have a fix for this sooner than later.

[Jason R. Coombs (jaraco)]

I should mention that I'm using dulwich as bundled with TortoiseHg (Windows) as a binary, so I don't have the luxury of simply patching the source as a workaround.

[Jason R. Coombs (jaraco)]

I did some reading into the relevant code in cgit: https://github.com/gitster/git/blob/master/date.c#L604.

If I read the code correctly, the C code interprets "--700" as a double-negative, or equivalent to +0700. I believe that implementation is a mistake, but if copying that mistake into dulwich is the proper fix, I will put together a patch.

[George Kola (georgekola)]

I wanted to get the latest version of boto 2.3.0 (precise has 2.2.2) because of some bug fixes for EC2 oregon data center. I tried to build it on launchpad and ran into the above bug. Is there a timeframe of when launchpad would get this fix.

Thanks,
George

[Jelmer Vernooij (jelmer)]

This can be fixed in Launchpad by updating to the latest bzr-git, which would also require a newer version of bzr to be used.

[Launchpad QA Bot (lpqabot)]

Fixed in stable r15901 <http://bazaar.launchpad.net/~launchpad-pqm/launchpad/stable/revision/15901>.