Re: [PATCH] Export GIT_DIR after setting it

From: martin f. krafft <madduck@madduck.net>
Date: May 15, 2008, 10:15 UTC
Message-ID: <20080515101523.GA31719@lapse.madduck.net>
In-Reply-To: <7vod78i9r7.fsf@gitster.siamese.dyndns.org>

Thank you, Junio, for taking the time to reply to this!

also sprach Junio C Hamano <gitster@pobox.com> [2008.05.15.0325 +0100]:

> trying to solve like this patch does, but this change needs very
> careful vetting to make sure that you did not break other scripts
> with this change.

Absolutely agreed. It occured to me as I lied down to sleep that this fix could quite possibly have repercussions. And it's been in my head all the walk to my work this morning. I ended up thinking about it in this way:

If GIT_DIR is exported by git-sh-setup and we can assure that git-sh-setup gets it right, then it's effectively the same as if the user had set it explicitly, before calling the shell script: all external commands called by the shell script will have GIT_DIR set appropriately in all cases then.

The only problem I see now is when an external command (or the shell script) can't properly deal with GIT_DIR being set, but then that's a whole different bug.

I understand you're worried about this, but I can't really see specifics, now having thought about this for a bit.

> This arrangement predates separate work-tree by many months.  It
> could be that what needs fixing is the separate work-tree code.

Oh yeah, and I've been meaning to look into that for a long time. Sigh.

-- 
martin | http://madduck.net/ | http://two.sentenc.es/
 
"she was rather too intelligent and competent-looking to be
 considered entirely beautiful, but all the more attractive because
 of it."
                           -- george spencer-brown, "a lion's teeth"
 
spamtraps: madduck.bogus@madduck.net