Re: [PATCH] reset cached ident date before creating objects

From: Jeff King <peff@peff.net>
Date: Aug 1, 2016, 22:00 UTC
Message-ID: <20160801220054.jita2awuaoebjhhp@sigill.intra.peff.net>
In-Reply-To: <xmqqinvkgno8.fsf@gitster.mtv.corp.google.com>

On Mon, Aug 01, 2016 at 02:54:47PM -0700, Junio C Hamano wrote:

Show 14 quoted lines

> Jeff King <peff@peff.net> writes:
> 
> > I also didn't go looking for other spots which might want similar
> > treatment. Junio mentioned that the sequencer code still uses an
> > external commit process, so cherry-pick/revert are OK. git-fast-import
> > creates a stream of commits, but already deals with this issue in a
> > separate way. And I couldn't think of other programs that want to make a
> > string of commits.
> 
> Thanks.  I wonder if do_commit() may be a more appropriate place to
> reset this, but the difference only would matter in "am --resolved",
> and a call to do_commit() it has will always be the first commit in
> the process, so there would not be anything to clear, which would in
> turn mean that it would not make any difference.

Yeah, you may be right. I didn't actually find that spot. I was focused on the looping aspect. And seeing how complex the loop was, my thought was to just put it at the top, so we know it happens once per iteration. But I think do_commit() is a reasonble spot, too (as long as it comes before the call to fmt_ident(), of course).

-Peff