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

From: Felipe Contreras <felipe.contreras@gmail.com>
Date: Nov 22, 2012, 00:15 UTC
Message-ID: <CAMP44s0Sd9V+GCqxid_rCwNH49-+dzmreg9zwPgxoZb1hQkb1A@mail.gmail.com>
In-Reply-To: <7vmwya3h0x.fsf@alter.siamese.dyndns.org>

On Wed, Nov 21, 2012 at 7:14 PM, Junio C Hamano <gitster@pobox.com> wrote:

Show 12 quoted lines

> Felipe Contreras <felipe.contreras@gmail.com> writes:
>
>> They have been marked as UNINTERESTING for a reason, lets respect that.
>> ...
>> The current behavior is most certainly not what we want. After this
>> patch, nothing gets exported, because nothing was selected (everything
>> is UNINTERESTING).
>
> The old behaviour was an incorrect "workaround" that has been
> superseded by your 14/15 "make sure updated refs get updated", no?
> Mentioning that would help people realize that this patch would not
> cause regression on them, I would think.

This particular patch is not getting rid of that "workaround", if you can call it that, it's just making it work correctly.

There's absolutely no possibility of regression (that is known or anybody has mentioned).

The only argument that was put forward was that 'git fast-export ^master master' should throw: from :0

As it does now, because in the future, with another patch (that nobody is pursuing), it might do: from 8c7a786

Which as I have tried to explain; is equally useless.

There's no regression, nobody would be affected negatively by this because when there are no marks, nobody expects a 'from :0'; it's totally useless, and when there are marks, nobody expects an update when the user does '^uninteresting master' for the 'uninteresting' ref. And not even potential future users would be affected, because 'from 8c7a786' is not helpful either, even if the user wanted '^uninsteresting' to be updated (which they won't), the git SHA-1 is useless to a remote helper without marks.

Cheers.

-- 
Felipe Contreras