Re: [PATCH 1/8] builtin/receive-pack: use check_signature from gpg-interface

From: Henning Schild <henning.schild@siemens.com>
Date: Jul 9, 2018, 08:18 UTC
Message-ID: <20180709101854.2777aae2@md1pvb1c.ad001.siemens.net>
In-Reply-To: <xmqqk1q8hwta.fsf@gitster-ct.c.googlers.com>

Am Fri, 6 Jul 2018 14:35:29 -0700 schrieb Junio C Hamano <gitster@pobox.com>:

Show 28 quoted lines

> Junio C Hamano <gitster@pobox.com> writes:
> 
> > Henning Schild <henning.schild@siemens.com> writes:
> >  
> >> The combination of verify_signed_buffer followed by
> >> parse_gpg_output is available as check_signature. Use that instead
> >> of implementing it again.
> >>
> >> Signed-off-by: Henning Schild <henning.schild@siemens.com>
> >> ---  
> >
> > Makes sense.  
> >
> > When d05b9618 ("receive-pack: GPG-validate push certificates",
> > 2014-08-14) implemented the check, there wasn't check_signature()
> > available.  The commit probably should have done what a4cc18f2
> > ("verify-tag: share code with verify-commit", 2015-06-21) later did
> > to introduce the check_signature() function by factoring it out of
> > commit.c::check_commit_signature() as a preparatory step.
> >
> > Will queue.  Thanks.  
> 
> Well, I guess I won't queue this version that would waste others'
> time, as you'd be rerolling to update variable names and such, so
> I'd wait for that (and you in turn would wait for the names and
> other discussions to settle).
> 
> Thanks anyway.

I think 1 and 2 can be seen as somewhat unrelated to the gpgsm feature, they are more general refactoring. So i think picking them is a good idea. It will make the series shorter and ease review in the next round.

Henning