From: Junio C Hamano <junkio@cox.net>
Date: Sun, 15 May 2005 21:46:11 GMT
Subject: Re: git repository for net drivers available
Message-ID: <7vacmwqi1o.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <20050515200514.GA31414@pasky.ji.cz>

>>>>> "PB" == Petr Baudis <pasky@ucw.cz> writes:

PB> Dear diary, on Fri, May 13, 2005 at 05:29:30PM CEST, I got a letter
PB> where Jeff Garzik <jgarzik@pobox.com> told me that...
>> Looks like cogito is using $repo/heads/$branch, whereas my git repo is 
>> using $repo/branches/$branch.

PB> Would it be a big problem to use refs/heads/$branch? That's the
PB> currently commonly agreed convention about location for storing branch
PB> heads, not just some weird Cogito-specific invention. And it'd be very
PB> nice to have those locations consistent across git repositories.

Since Jeff brought up $repo/branches/$branch, you may also want
to add that $repo/branches/$branch is used to record the URL of
the remote $branch (the information used to be in a flat file
$repo/remotes, branch name and URL separated by shell $IFS, one
record on each line), and is quite different from those 40-byte
SHA1 plus LF files you see in $repo/refs/*/ directory.

I think it is a reasonable one, I also follow the
$repo/branches/$branch convention Cogito uses, and I would
encorage other Porcelain implementations to follow suit.