From: Olivier Galibert <galibert@pobox.com>
Date: Tue, 03 Apr 2007 23:08:46 GMT
Subject: Re: Distribution of longest common hash prefixes
Message-ID: <20070403230846.GB8479@dspnet.fr.eu.org>
In-Reply-To: <alpine.LFD.0.98.0704031635100.28181@xanadu.home>

On Tue, Apr 03, 2007 at 04:39:02PM -0400, Nicolas Pitre wrote:
> On Tue, 3 Apr 2007, Junio C Hamano wrote:
> 
> > I stated it wrongly.  What I was getting at was that we might
> > want to consider an abbreviation that matches only a single
> > commit unambiguous even when there are ambiguous objects of
> > other kinds.
> 
> Maybe.  But by the time your object hash distribution starts showing 
> ambiguous objects with a given abbreviated name between a commit and a 
> non commit, I'll bet you'll start to see ambiguities between commits 
> soon enough as well.

Isn't the number of objects an order of magnitude bigger than the
number of commits?  Well, I guess that depends on your workflow...

  OG.