Re: More problems...

From: Ryan Anderson <ryan@michonline.com>
Date: Apr 29, 2005, 19:50 UTC
Message-ID: <20050429195055.GE1233@mythryan2.michonline.com>
In-Reply-To: <20050429182708.GB14202@pasky.ji.cz>

On Fri, Apr 29, 2005 at 08:27:08PM +0200, Petr Baudis wrote:

Show 16 quoted lines

> Dear diary, on Fri, Apr 29, 2005 at 06:01:27PM CEST, I got a letter
> where Russell King <rmk@arm.linux.org.uk> told me that...
> > rmk@dyn-67:[linux-2.6-rmk]:<1049> cg-update origin
> > `../linux-2.6/.git/objects/00/78aeb85737197a84af1eeb0353dbef74427901' -> `.git/objects/00/78aeb85737197a84af1eeb0353dbef74427901'
> > cp: cannot create link `.git/objects/00/78aeb85737197a84af1eeb0353dbef74427901': File exists
> > 
> > By that time, the object files in the reference tree appear to have
> > a newer timestamp than the corresponding ones in my local tree, and
> > cp -lua fails.
> 
> I'm now away ,unfortunately, and no immediate idea stems to my mind on
> how to fix it. Ideas welcomed - I need to hardlink missing entries from
> one tree to another; it would be enough to be able to just tell cp to
> ignore already present files.
> 
> Could you please try to give cp the -f flag?

Why not just use "rsync" for both remote and local synchronization, and provide a "relink" command to scan two .git/objects/ repositories and hardlink matching files together?

With the SHA1 hash, you can even have a --unsafe option that just compares the has names and does a link based purely off of that and the stat(2) results of both files. (I'd expect that a ... safer variant would extract both files and compare them, but the --unsafe should be sufficient, in practice, I would think.)

-- 
Ryan Anderson
  sometimes Pug Majere