Re: [PATCH v5 15/15] fast-export: don't handle uninteresting refs

From: Felipe Contreras <felipe.contreras@gmail.com>
Date: Nov 22, 2012, 00:28 UTC
Message-ID: <CAMP44s2B2_htR8LFbHk99WaNUcaYJCxVJPdRdj5VQ0k+fB9NOg@mail.gmail.com>
In-Reply-To: <20121121194810.GE16280@sigill.intra.peff.net>

On Wed, Nov 21, 2012 at 8:48 PM, Jeff King <peff@peff.net> wrote:

Show 14 quoted lines

> On Tue, Nov 20, 2012 at 09:08:36PM -0800, Junio C Hamano wrote:
>
>> With such a one-sided discussion, I've been having a hard time
>> convincing myself if Felipe's effort is making the interface better,
>> or just breaking it even more for existing remote helpers, only to
>> fit his world model better.
>
> Felipe responded in more detail, but I will just add the consensus we
> came to earlier in the discussion: the series does make things better
> for users of fast-export that use marks, but does not make things any
> better for users of negative refs on the command line. However, I do not
> think that it makes things worse for them, either (neither by changing
> the behavior negatively, nor by making the code harder for a more
> complete fix later).

Patch 14 changes the behavior depending on the marks, patch 15 doesn't.

This patch is mostly orthogonal to marks.

Before without marks: % git branch unintresting master % git fast-export master ^uninteresting reset refs/heads/uninteresting from :0

Before with marks: % git fast-export --import-marks=marks master ^uninteresting reset refs/heads/uninteresting from :100

See? In both cases git is doing something the user doesn't want, nor specified. After my patch nothing gets updated, because nothing was specified to be updated.

I'm not going to bother explaining why other people objected to this patch (again), which is indeed related to marks, they should do it for themselves. Let me reaffirm that no valid reason has put forward to object to this patch.

> So while fixing everybody might be nice

I would like to understand that that even means. What behavior is currently broken? And for who? And how is this patch related to that?

> [1] There are other possible use cases for fast-export which might
>     benefit from negative refs working more sanely, but since they are
>     in the minority and are not being made worse, I think the partial
>     fix is OK.

Which ones? I don't think this is a partial fix.

Nobody has put forward such a use-case.

Cheers.

-- 
Felipe Contreras