Subject: Re: [Gnu-arch-users] corrupt pristine
From: Zenaan Harkness <zen@cashpoolcomps.com>
To: arch <gnu-arch-users@gnu.org>
In-Reply-To: <20041030044950.GA26777@patrick.wattle.id.au>
References: <1099109410.17184.17.camel@localhost.localdomain>
	<20041030044950.GA26777@patrick.wattle.id.au>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Message-Id: <1099112449.17184.64.camel@localhost.localdomain>
Mime-Version: 1.0
Date: Sat, 30 Oct 2004 15:00:49 +1000
Precedence: list
Sender: gnu-arch-users-bounces+migo=homemail.com@gnu.org
Errors-To: gnu-arch-users-bounces+migo=homemail.com@gnu.org
Status: RO
Content-Length: 1300
Lines: 41
On Sat, 2004-10-30 at 14:49, Cameron Patrick wrote:
> Zenaan Harkness wrote:
> 
> > Can we simply delete a currupt (inodes due to old machine dying => new
> > disk) pristines tree, and all will continue to work as expected (no
> > revlib here)?
> 
> Yes.  Better yet, delete every pristine you can find and set up a
> revision library. :-)

Did work indeed. So, why have pristines at all?

> > Now, the really baffling thing for me (since I don't understand it) is
> > that we got the same error, when we renamed the old working dir tree to
> > a backup name, then tried to checkout (get) a new working tree ???
> 
> Arch looks in the current tree and in neighbouring directories to find
> a pristine, you really do need to delete it (or rename
> ++pristine-trees to ++not-pristine-trees or something).

Is that appropriate?

It wasn't a neighbouring tree, it was a subdirectory!

I wanted to checkout/get, not commit! As in, I wanted a new tree, with a
new pristine!!

This seems to me like a logically wrong thing to do - surely a bug in
the pristines-locating algorithm?

ta
zen


_______________________________________________
Gnu-arch-users mailing list
Gnu-arch-users@gnu.org
http://lists.gnu.org/mailman/listinfo/gnu-arch-users

GNU arch home page:
http://savannah.gnu.org/projects/gnu-arch/