Re: [PATCH] Add test for symlinked .git/packed-refs

From: Junio C Hamano <junkio@cox.net>
Date: Apr 19, 2007, 06:06 UTC
Message-ID: <7vabx499u2.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <20070419053518.GK5913@xp.machine.xx>

Peter Baumann <waste.manager@gmx.de> writes:

Show 15 quoted lines

> Signed-off-by: Peter Baumann <waste.manager@gmx.de>
> ---
> On Wed, Apr 18, 2007 at 02:31:29PM -0700, Junio C Hamano wrote:
>> 
>> Oh, I never questioned that you made that basic case work.  I
>> was worried about not making sure the symlink we are looking at
>> really is the case we are willing to handle, and not erroring
>> out if that is not the case, perhaps like the attached patch on
>> top of yours.
>> 
>> An additional test or two in t/t3210 would be nice to accompany
>> this change.
>> 
>
> Something like this?

That's a good start, but I expected to see at least tests for two cases: a case in which .git/packed-refs symlink points at an actual file (i.e. the original repository has run pack-refs) and another case in which .git/packed-refs symlink is dangling (i.e. the original repository hasn't run pack-refs). I understand that the borrower "worktree" can have .git/packed-refs symlink pointing at the repositories .git/packed-refs yet to be born.