From: Jeff King <peff@peff.net>
To: Chris Patti <cpatti@gmail.com>
Cc: "Carlos Martín Nieto" <cmn@elego.de>, git@vger.kernel.org
Subject: Re: Odd issue - The Diffs That WILL NOT DIE.
Date: Wed, 7 Dec 2011 18:03:29 -0500 [thread overview]
Message-ID: <20111207230329.GB21852@sigill.intra.peff.net> (raw)
In-Reply-To: <CAJ8P3RA48W7ZiABvjkn_KkU-JPnCnaF_X_WK0wPtToph3DGDvg@mail.gmail.com> <CAJ8P3RB=Gj-QCe6meqXSZ7N8+PnfNxSD8omUxT6dDh00bUf0QQ@mail.gmail.com>
On Wed, Dec 07, 2011 at 05:24:01PM -0500, Chris Patti wrote:
> Yup, you nailed it. The files in question are CloudSponge.php
> (deleted) and Cloudsponge.php (still being actively maintained).
> [...]
> Actually I'm wrong on that count, but in an interesting way.
>
> Both CloudSponge.php and Cloudsponge.php exist and are *not* deleted
> in the remote repository, but on OSX only Cloudsponge.php shows up on
> the filesystem, yet CloudSponge.php is being reported as modified.
>
> Turns out two of our other developers are also seeing this behavior.
OK. Then it's not a bug[1], but rather an incompatibility issue. Your
repository stores two files which differ only in case, but your
underlying case-insensitive filesystem is incapable of storing both of
them simultaneously.
There's really no way around it except to either use a case-sensitive
filesystem, rename your files (or if one if obsolete, get rid of it).
[1] Arguably git could be better at noticing and reporting this issue,
but that wouldn't change the fundamental problem.
> I am seeing the same behavior with 1.7.7.4 which I backrevved to
> yesterday while troubleshooting this issue. Can you suggest an older
> version for me to try next?
No, you'll see the same behavior in all versions of git (I was mostly
concerned that it was a regression new in v1.7.8, as I fixed some
case-sensitivity bugs recently).
> I'm not sure how I would git bisect in this case, I'd need to have all
> the different git revs installed in order to do that right? (I'm
> relatively new to git bisect, just figured it out the other day).
You can't bisect, because you don't have a version that actually worked.
But if you had, you would do it by cloning and building the git
repository itself, then starting a bisection like:
git clone git://git.kernel.org/pub/scm/git/git.git
git bisect start
git bisect bad v1.7.8
git bisect good v1.7.7.4 ;# if it had turned out to be OK
and then repeatedly build and try your test with each version of git,
until if finally narrowed it down to a single offending commit.
-Peff
prev parent reply other threads:[~2011-12-07 23:03 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-12-06 21:43 Odd issue - The Diffs That WILL NOT DIE Chris Patti
2011-12-06 21:47 ` Frans Klaver
2011-12-06 21:51 ` Carlos Martín Nieto
2011-12-07 16:54 ` Chris Patti
2011-12-07 17:20 ` Philippe Vaucher
2011-12-07 18:05 ` Chris Patti
2011-12-07 18:16 ` Chris Patti
2011-12-07 22:03 ` Jeff King
2011-12-07 22:24 ` Chris Patti
2011-12-07 22:30 ` Chris Patti
2011-12-07 22:58 ` Carlos Martín Nieto
2011-12-07 23:06 ` Jeff King
2012-09-05 19:38 ` Robin Rosenberg
2011-12-07 23:03 ` Jeff King [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=20111207230329.GB21852@sigill.intra.peff.net \
--to=peff@peff.net \
--cc=cmn@elego.de \
--cc=cpatti@gmail.com \
--cc=git@vger.kernel.org \
/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;
as well as URLs for NNTP newsgroup(s).