Cygwin-bug#20050820T1947: [serious] IndexError list index out of range

[Jonathan Lange]

On 23/08/05, John A Meinel <john at arbash-meinel.com> wrote:
> Jari Aalto wrote:
> > | > Version: 0.0.6
> > | >
> > | > I get this error under Cygwin. Please take a look.
> > | >
> > |
> > | (The weird traceback is because on an exception we log the state at the
> > | time of the error. I really don't know why that code is failing, since
> > | it seems to be giving a line number error.)
> >
> > I just wanted confirmation that it is normal behavior to get traceback
> > even for cases like this. I would have expected simple:
> >
> 
> Well, that's what it was trying to do. Usually it adds a bunch of stuff
> to ~/.bzr.log for debugging purposes. So that when someone posts a
> message saying, "Hey, this is broken", we can have a good traceback of
> what went wrong.
> 
> What is weird is that the traceback code was trying to read some context
> lines, and didn't seem to find them.
> I think the code needs a try/catch (or actually better to just check
> first, because you don't want to throw another exception, and lose track
> of the current one).
> But I'm not sure if this is bzr code, or the built-in python libraries.
> It certainly should be investigated.
> 

I get the same stack trace.  It happens pretty much any time an
exception is raised (I guess), but a solid way to reproduce is: run
bzr selftest; press Ctrl-C while its actually running the test
(printing '~'s).   I'm using Mac OS X 10.3, Python 2.4.1 and bzr
0.0.6.

This appears to be fixed in the latest bzr.