Git development
 help / color / mirror / Atom feed
From: Johannes Schindelin <Johannes.Schindelin@gmx.de>
To: Junio C Hamano <gitster@pobox.com>
Cc: Mike Coleman <tutufan@gmail.com>, git@vger.kernel.org
Subject: Re: bug: git-diff silently fails when run outside of a repository (v1.5.4.2)
Date: Wed, 30 Apr 2008 09:34:51 +0100 (BST)	[thread overview]
Message-ID: <alpine.DEB.1.00.0804300933570.17469@eeepc-johanness> (raw)
In-Reply-To: <7vwsmg2ks7.fsf@gitster.siamese.dyndns.org>

Hi,

On Tue, 29 Apr 2008, Junio C Hamano wrote:

> Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:
> 
> >> > Even so, this seems like a bug.  If I do this:
> >> >
> >> >     $ cd /
> >> >     $ git-diff
> >> >
> >> > there is no error message and no error status.  A diagnostic would be
> >> > very helpful.
> >> 
> >> Ah, that indeed is not very helpful.
> >> 
> >> Unfortunately, every time I look at this hack, I seem to find an unrelated
> >> bug in it.  Here is today's.
> >> 
> >> 	$ for i in 1 2 3; do >/var/tmp/$i; done
> >>         $ cd /
> >>         $ git diff /var/tmp/1
> >>         Segmentation Fault
> >> 
> >> When nongit is true, we know the user has to be asking --no-index diff, so
> >> perhaps we can fix it by doing something like this?
> >> 
> >> diff --git a/diff-lib.c b/diff-lib.c
> >> index 069e450..cfd629d 100644
> >> --- a/diff-lib.c
> >> +++ b/diff-lib.c
> >> @@ -264,6 +264,9 @@ int setup_diff_no_index(struct rev_info *revs,
> >>  			DIFF_OPT_SET(&revs->diffopt, EXIT_WITH_STATUS);
> >>  			break;
> >>  		}
> >> +	if (nongit && argc != i + 2)
> >> +		die("git diff [--no-index] takes two paths");
> >> +
> >>  	if (argc != i + 2 || (!is_outside_repo(argv[i + 1], nongit, prefix) &&
> >>  				!is_outside_repo(argv[i], nongit, prefix)))
> >>  		return -1;
> >
> > That looks to me as if the second if() should have triggered, and the 
> > caller of setup_diff_no_index() should have errored out.
> 
> I think the above three-liner fix is something we should have done when we
> added --no-index codepath.  Before the --no-index hack was introduced, we
> did not even got this far to the place the caller of this function is, if
> we are outside a repository.  By returning -1 from here instead of dying,
> this code is driving the codepath that has always expected to already be
> in a repository into a nonrepository, causing them to segfault because
> there is no git-dir or work-tree set up done yet as they expect.

Fair enough.

Ciao,
Dscho

      reply	other threads:[~2008-04-30  8:35 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-04-29 20:04 bug: git-diff silently fails when run outside of a repository (v1.5.4.2) Mike Coleman
2008-04-29 22:53 ` Junio C Hamano
2008-04-29 23:03   ` Mike Coleman
2008-04-29 23:37     ` Junio C Hamano
2008-04-30  0:56       ` Johannes Schindelin
2008-04-30  1:28         ` Junio C Hamano
2008-04-30  8:34           ` Johannes Schindelin [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=alpine.DEB.1.00.0804300933570.17469@eeepc-johanness \
    --to=johannes.schindelin@gmx.de \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=tutufan@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox