From: Jeff King <peff@peff.net>
Date: Thu, 26 Feb 2009 14:40:52 GMT
Subject: Re: [PATCH 08/21] remote: let guess_remote_head() optionally return all matches
Message-ID: <20090226144052.GB9693@coredump.intra.peff.net>
In-Reply-To: <20090226143729.GA9693@coredump.intra.peff.net>

On Thu, Feb 26, 2009 at 09:37:29AM -0500, Jeff King wrote:

> Hmm. This should probably be:
> 
>   dst->peer_ref = src->peer_ref ? copy_ref(src->peer_ref) : NULL;
> 
> (or copy_ref should return NULL when given NULL). I also wonder if the
> copied ref's peer_ref should be explicitly NULL'd.

BTW, all of my "probably" and "I wonder" here are because I think the
"peer ref" pointer is a little vague as a concept. E.g., I think in most
cases src->peer_ref->peer_ref != src.

Rather than having ref structs with "next" and "peer" pointers, I think
a more natural data structure would be a list (or array) of "ref pairs".

But you didn't create that with this series, and I don't think it is
worth the major surgery to change it now.

-Peff