bzr send tracebacks when used outside of a branch

[Matt Nordhoff]

It tracebacks on branch.unlock() when used outside of a branch/tree/whatever.

bzr: ERROR: exceptions.UnboundLocalError: local variable 'branch' referenced before assignment

Traceback (most recent call last):
  File "/.../bzr.dev/bzrlib/commands.py", line 802, in run_bzr_catch_errors
    return run_bzr(argv)
  File "/.../bzr.dev/bzrlib/commands.py", line 758, in run_bzr
    ret = run(*run_argv)
  File "/.../bzr.dev/bzrlib/commands.py", line 492, in run_argv_aliases
    return self.run(**all_cmd_args)
  File "/.../bzr.dev/bzrlib/builtins.py", line 4024, in run
    kwargs.get('from', '.'), mail_to, message)
  File "/.../bzr.dev/bzrlib/builtins.py", line 4124, in _run
    branch.unlock()
UnboundLocalError: local variable 'branch' referenced before assignment

bzr 0.93.0.dev.0 on python 2.5.1.final.0 (linux2)
arguments: ['/home/me/local/bin/bzr', 'send']
encoding: 'ANSI_X3.4-1968', fsenc: 'ANSI_X3.4-1968', lang: 'C'
plugins:
  bzrtools /.../bzr.dev/bzrlib/plugins/bzrtools [0.92.1]
  launchpad /.../bzr.dev/bzrlib/plugins/launchpad [unknown]
  multiparent /.../bzr.dev/bzrlib/plugins/multiparent.pyc [unknown]
*** Bazaar has encountered an internal error.
    Please report a bug at https://bugs.launchpad.net/bzr/+filebug
    including this traceback, and a description of what you
    were doing when the error occurred.

(Latest bzr.dev, r3109.)

[Lukáš Lalinský (luks)]

http://bundlebuggy.aaronbentley.com/request/%3C1196457323.6389.25.camel@laptop%3E

[Matt Nordhoff (mnordhoff)]

Oh, perfect. :-)

Shouldn't that have been merged by now?

[Lukáš Lalinský (luks)]

Dunno, it's been approved, but not merged, for about two weeks now...

[John A Meinel (jameinel)]

Yeah it should have been merged.
The general rule is that the last person who BB:approve is supposed to merge it.
BB is designed to hide things others have approved so you don't have to worry about it.

But that means that in the normal views *I* don't see all the things that Robert or Alexander or Martin have approved but not merged, and Robert went on vacation.

So even with BB meaning we don't fully drop patches, things can still slide. So if this happens again, a reminder is perfectly appropriate.

In the meantime, I'll merge this.

[Alexander Belchenko (bialix)]

John A Meinel пишет:
> Yeah it should have been merged.
> The general rule is that the last person who BB:approve is supposed to merge it.
> BB is designed to hide things others have approved so you don't have to worry about it.
>
> But that means that in the normal views *I* don't see all the things
> that Robert or Alexander or Martin have approved but not merged, and
> Robert went on vacation.
>
> So even with BB meaning we don't fully drop patches, things can still
> slide. So if this happens again, a reminder is perfectly appropriate.

Yes, short ping for merge request or the last person approved the patch
directly will not harm. So, please, ping us, when patch too long sit in
the queue.

[Lukáš Lalinský (luks)]

On Pi, 2007-12-14 at 13:33 +0000, John A Meinel wrote:
> So even with BB meaning we don't fully drop patches, things can still
> slide. So if this happens again, a reminder is perfectly appropriate.

Well, it depends on the patch. Patches for bug fixes like this are not
the ones I particulary care about. I usually run bzr.dev and this bug
happened to me once when I typed 'bzr send' in a wrong directory, I had
some free time, fixed it, sent a patch and forgot about it.

But new bug reports are always good as reminders. :)

Lukas