bzr log crash

[Peter Amstutz]

I created a fairly large branch (actually an import of boost) and committed a couple of changes. Now when I try to do bzr log, I get the following error. This problem also seems to be affecting my ability to use tailor to export changesets from bzr to cvs. "bzr check" does not report any problems, and this error occurs in two different repositories with copies of this branch. This branch is available publically at http://interreality.org/home/bzroot/boost

First the output:

$ bzr log
------------------------------------------------------------
revno: 2
committer: Peter Amstutz <email address hidden>
branch nick: boost
timestamp: Mon 2006-05-15 00:36:27 -0400
message:
  eliminated unused libraries
------------------------------------------------------------
revno: 1
committer: Peter Amstutz <email address hidden>
branch nick: vos-boost
timestamp: Sat 2006-05-13 19:38:29 -0400
message:
  sending to server
bzr: ERROR: exceptions.KeyError:
'<email address hidden>'
  at /usr/lib/python2.4/site-packages/bzrlib/log.py line 256
  in _show_log

What shows up in .bzr.log:

bzr 0.8.2 invoked on python 2.4.2.final.0 (linux2)
  arguments: ['/usr/bin/bzr', 'log']
  working dir: u'/home/bzroot/boost'
looking for plugins in /usr/lib/python2.4/site-packages/bzrlib/plugins
Plugin name __init__ already loaded
Plugin name __init__ already loaded
add plugin name launchpad
load plugin (None, '/usr/lib/python2.4/site-packages/bzrlib/plugins/launchpad', ('', '', 5))
registered plugin command register-branch
loaded succesfully
looking for plugins in /home/tetron/.bazaar/plugins
got branch format Bazaar-NG Metadir branch format 5
encoding log as 'ANSI_X3.4-1968'
cache hit for revision-history in BzrBranch5(u'/home/bzroot/boost/')
readv coalesced 1 reads.
readv coalesced 1 reads.
readv coalesced 1 reads.
[20095] Mon 23:55:05.366 ERROR: exceptions.KeyError: '<email address hidden>'
  at /usr/lib/python2.4/site-packages/bzrlib/log.py line 256
  in _show_log
Traceback (most recent call last):
  File "/usr/lib/python2.4/site-packages/bzrlib/commands.py", line 650, in run_bzr_catch_errors
    return run_bzr(argv)
  File "/usr/lib/python2.4/site-packages/bzrlib/commands.py", line 612, in run_bzr
    ret = run(*run_argv)
  File "/usr/lib/python2.4/site-packages/bzrlib/commands.py", line 246, in run_argv_aliases
    return self.run(**all_cmd_args)
  File "/usr/lib/python2.4/site-packages/bzrlib/commands.py", line 622, in ignore_pipe
    result = func(*args, **kwargs)
  File "/usr/lib/python2.4/site-packages/bzrlib/builtins.py", line 1262, in run
    search=message)
  File "/usr/lib/python2.4/site-packages/bzrlib/log.py", line 161, in show_log
    start_revision, end_revision, search)
  File "/usr/lib/python2.4/site-packages/bzrlib/log.py", line 256, in _show_log
    lf.show(rev_nos[rev_id], rev, delta)
KeyError: '<email address hidden>'

return code 3

[Lalo Martins (lalo.martins)]

some additional info from dissecting the branch with the failure:

lalo@bugg:/home/bzroot/boost/.bzr$ zcat repository/revisions.knit | grep revision_ref
<revision_ref <email address hidden>" />
<revision_ref <email address hidden>" />
lalo@bugg:/home/bzroot/boost/.bzr$ cat branch/revision-history
<email address hidden>
<email address hidden>

So as you can see, revision 2 has the same id on both places, but revision 1 has different ids. How could that possibly happen?

[John A Meinel (jameinel)]

revision_ref is actually the parent revision. Indicating what the previous revision was. What you really want to look for is the line with "<revision.*revision_id"

Though by the looks of it, it seems you have a revision which isn't properly referenced.

John

[John A Meinel (jameinel)]

Still waiting on more information about the cause of this bug.

If it is still causing problems, then we could possibly meet on IRC sometime next week to help diagnose the problem.
Otherwise this bug report should probably be closed.

[Aaron Bentley (abentley)]

Not apparently a bug in log, and no response from reporter.

[Lalo Martins (lalo.martins)]

Er. Sorry. This wasn't related to log at all, it was something weird to do with missing revisions, a bad combination of push and checkouts, I don't remember the details anymore. Either way, I did the same mistake on purpose on 0.14, and it doesn't crash log anymore, so yeah, it's closed.