Re: Catching up: git-pasky-0.6.2 broken?

From: Russell King <rmk@arm.linux.org.uk>
Date: Apr 23, 2005, 14:12 UTC
Message-ID: <20050423151238.C32116@flint.arm.linux.org.uk>
In-Reply-To: <20050423124758.B32116@flint.arm.linux.org.uk>

On Sat, Apr 23, 2005 at 12:47:58PM +0100, Russell King wrote:

Show 20 quoted lines

> I've been away from git stuff since about Wednesday due to eye problems.
> Yesterday and today, I've been trying to catch up, but I'm running into
> problems.
> 
> I grabbed git-pasky-0.6.2, and followed Linus' message about converting
> repositories.  I've updated all the "heads", and cleaned out all the old
> sha1 files.  After updating, I did a read-tree for the tracked head,
> followed by checkout-cache -f -a and update-cache --refresh.
> 
> Therefore, in theory, everything should be in sync.
> 
> However, git pull now has nasty side effects.  At first I thought this
> was down to something still being out of sync.  However, on the second
> pull, it's still producing the same complaints.
> 
> It's almost although git has a single patch file subversely written into
> it to apply lpfc and qla changes which it's trying to apply irrespective
> of the objects downloaded.
> 
> Any ideas what's going on?

Could the problem be related to some random garbage left in .git/add-queue and .git/rm-queue?

If so, how did these files get generated in the first place, and why weren't they removed when they were finished with?

IMHO updating a repository from an external source should _NOT_ be affected by the presence (or absense) of these two files, which contain only _local_ state information.

-- 
Russell King