Re: What's cooking in git.git (Jan 2010, #06; Wed, 20)

From: Johan Herland <johan@herland.net>
Date: Jan 21, 2010, 09:18 UTC
Message-ID: <201001211018.10085.johan@herland.net>
In-Reply-To: <7vvdeww187.fsf@alter.siamese.dyndns.org>

On Thursday 21 January 2010, Junio C Hamano wrote:

> Johan Herland <johan@herland.net> writes:
> > On Thursday 21 January 2010, Sverre Rabbelier wrote:
> >> On Thu, Jan 21, 2010 at 01:52, Junio C Hamano <gitster@pobox.com>

wrote:

Show 20 quoted lines

> >> > * jh/notes (2010-01-17) 23 commits
> >> > [...]
> >> > Updated with a re-roll.
> >>
> >> Just checking, you reverted all from next (with exception of the first
> >> three), and now re-queued it to pu, with the first three still in
> >> next? Or did I mis-remember and did only the first three make it to
> >> next in the first place?
> >
> > You misremembered. Only the three first were merged to 'next'. Junio
> > was about to merge the rest, but I asked him to hold until I had
> > produced the current iteration.
> 
> I've been meaning to merge the first three to 'master', as many people
> have been running 'next' and new features tend to be exercised less by
> those on 'master' than on 'next', and it would be beneficial to make
> 'master' at 1.7.0-rc0 as close to what we have had in 'next' for a long
> time.
> 
> Worries?

Maybe. I've got a maybe-todo entry about (yet again) rewriting the fast- import handling of notes to use the updated notes API (which now can handle non-notes, and therefore should be better suited for fast-import's use case). However, the result is only worth submitting if it _both_ decreases code duplication between fast-import.c and notes.c, _and_ decreases the overall impact on fast-import.c. I don't know when I'll get around to doing this experiment, so feel free to ignore it for now.

In any case, I totally understand if you'd rather want to hold the entire notes series in 'pu', and instead focus on fixing the bugs in what's already been released.

...Johan

-- 
Johan Herland, <johan@herland.net>
www.herland.net