From: Nguyen Thai Ngoc Duy <pclouds@gmail.com>
Date: Mon, 09 Jun 2008 04:24:19 GMT
Subject: Re: [PATCH] fix attribute handling in bare repositories
Message-ID: <fcaeb9bf0806082124m604e5280o55bd0fef39345ad1@mail.gmail.com>
In-Reply-To: <7vskvnirjm.fsf@gitster.siamese.dyndns.org>

On Mon, Jun 9, 2008 at 3:46 AM, Junio C Hamano <gitster@pobox.com> wrote:
> René Scharfe <rene.scharfe@lsrfire.ath.cx> writes:
>
>> Nguyen Thai Ngoc Duy schrieb:
>>> Hi,
>>>
>>> Currently attr.c will read .gitattributes on disk no matter there is a
>>> real worktree or not. This can lead to strange behavior.
>>
>> Yes, it probably shouldn't do that.  What about this patch?
>
> Hmm.  I do not know if it breaks anything, but if you are indeed in a bare
> repository, the files the codepaths affected try to read would not exist
> anyway, so I am not sure what this would fix, other than changing the
> behaviour of check-attr from noticing that it was asked for nonsense and
> bail out to not noticing nor saying anything useful.

Yes. But just in case those .gitattributes exist by accident (for
example you access a bare repository from a worktree of another
repository). I don't want to check cwd for .gitattributes everytime I
want to use git-archive :)

Anyway if I do "git-archive <commit>" it should read .gitattributes
from that commit, not from working directory.
-- 
Duy