From: Eyvind Bernhardsen <eyvind.bernhardsen@gmail.com>
Date: Fri, 11 Jun 2010 08:47:44 GMT
Subject: Re: [RFC] ll-merge: Normalize files before merging
Message-ID: <4C11F830.3050404@gmail.com>
In-Reply-To: <20100611083641.GB31109@pvv.org>

On 11. juni 2010 10:36, Finn Arne Gangstad wrote:

> It would probably be more correct to do
>
> conert_to_git(convert_to_work_tree(x)) for the parts taken from base
> and theirs, in theory I guess that this can be true:
>
> convert_to_git(x) != convert_to_git(convert_to_work_tree(x))

Interesting idea, might be a bit slow perhaps?  I'll do some 
benchmarketing.  Incidentally, it has to convert_to_work_tree() "ours" 
as well, since the merge code doesn't read anything from the worktree.

I think convert_to_git(convert_to_work_tree(x)) is actually less likely 
to cause trouble than convert_to_git(convert_to_git(x)), which is what 
blind normalization would end up doing for at least one side of the merge.

- Eyvind