From: Dane Jensen <careo@fastmail.fm>
Date: Tue, 19 Feb 2008 23:18:51 GMT
Subject: Re: [Bug] Segfault in git status
Message-ID: <746D851F-CE49-4867-BB0B-D09615463244@fastmail.fm>
In-Reply-To: <7vfxvpn3da.fsf@gitster.siamese.dyndns.org>


On Feb 18, 2008, at 5:26 PM, Junio C Hamano wrote:

> Dane Jensen <careo@fastmail.fm> writes:
>
>> This was in OS X 10.5.2 with git compiled from cf5c51e. Falling back
>> to 1.5.4.2 cleared it up, so with the help of git bisect I've  
>> narrowed
>> the bug down to cf55870.
>
> Do you have this issue only on OSX?  Do you have the issue on
> any commit, or a particular commit?

It's not with any particular commit. It occurs with a fresh clone or  
at any other point, regardless of any changes I've made, what branch  
I'm on, or anything else I could think of testing.

> What does the change between HEAD, index and the work tree
> involve in the commit that causes the problem?  Can you make
> repository available to the public for diagnosis (if it is "Only
> on OSX", I may not be able to help much with, but never hurt to
> ask)?

Unfortunately not. I'll try to see if I can't cook up a test case that  
reproduces it however.

> Just to double check, can you see if reverting the commit (and
> another commit that immediately follows) fixes the issue for
> you?
>
> A patch to do the revert for cf5c51e is attached.

The patch does indeed fix it.

Additionally, now that I see the output of a successful run right  
after a failed one, it appears that git-status is dying before listing  
untracked files. It will output the "Changed but not updated:"  
section, but then die before getting to "Untracked files:".

-Dane