From mboxrd@z Thu Jan  1 00:00:00 1970
From: larsh at hjemli.net (larsh at hjemli.net)
Date: Thu, 21 Jul 2011 10:34:30 +0000
Subject: [PATCH] guess default branch from HEAD
In-Reply-To: <loom.20110719T120142-766@post.gmane.org>
References: <loom.20110610T051936-229@post.gmane.org>
 <20110613101621.GA29265@hjemli.net>
 <BANLkTinOiwZy6Tn4nN7NoJGnEkXqFgHKZw@mail.gmail.com>
 <20110616145921.GA5875@plenz.com>
 <BANLkTim+jC6A1zcx6Qepc43q=KEypsTSQA@mail.gmail.com>
 <loom.20110709T095610-848@post.gmane.org> <20110719063058.GA7524@hjemli.net>
 <loom.20110719T112847-313@post.gmane.org> <20110719095607.GC16136@hjemli.net>
 <loom.20110719T120142-766@post.gmane.org>
Message-ID: <20110721103430.GD8377@hjemli.net>

On Tue, Jul 19, 2011 at 10:13:54AM +0000, chris wrote:
>  <larsh at ...> writes:
> > On Tue, Jul 19, 2011 at 09:35:42AM +0000, chris wrote:
> > > Should guess_defbranch() be called here instead of hard-coding "master"?
> > 
> > I don't think so, due to performance concerns (as explained earlier in this
> > thread). It's a compromise, favoring performance over correctness.
> 
> Well, it sounds like premature optimization to me.

That may be true.


> I wonder how much 
> performance impact that would really have, and if cgit's caching support doesn't 
> negate any potential issue on busy servers anyway.

I'll try to do some benchmarking with different
setups and post some numbers.

--
larsh