Re: What's cooking in git.git (Jul 2016, #03; Fri, 8)

From: Junio C Hamano <gitster@pobox.com>
Date: Jul 11, 2016, 17:13 UTC
Message-ID: <xmqqzipo9jyx.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <alpine.DEB.2.20.1607110821500.6426@virtualbox>

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

> What about the am-call-merge-recursive-directly patch series? As I
> demonstrated by rebasing it to `pu`, it is actually not butchering the
> smudge/clean pathway as you suggested.

What I said was it seemed to conflict with something else, which butchered that codepath that needed further work. It seems I was reading the conflicts incorrectly and the conflicts were coming from other topics like i18n and oid changes.

> I am a bit at a loss here: what can I do to get this picked up?

You can do one of three things, and they apply not specifically to this case but in general when working with others.

 - You can help other topics that collide with what you do to move
   forward by helping their reviews.  Instead of leaving them
   something the still need polishing and requires your topic to get
   adjusted every time they change, help them be polished earlier
   and become part of the solid foundation you build on top.

 - You can wait until that polishing happens to the other topics
   that block you.

 - You can shoot down the other topics that block you, e.g. "these
   are not good ideas", "it aims to do a good thing, but its
   implementation is far from ready--it misses this and that cases
   among others. I'd recommend ejecting the topic for now and have
   it redesigned from scratch", etc., which would shift the issue of
   conflicting change to their problem as a side effect.

Rebasing on 'pu' essentially is the second course; you are explicitly making your topic depend on the others (which creates a bit more work to me to identify exactly which topics in 'pu' you absolutely need to decide where to queue your patches, compared to the case in which you explicitly say "these patches apply to a merge of topic X and Y on top of v2.9", though).

Also, without _any_ conflicts with other topics, the above three points apply well when working with others. There are tons of topics that are marked as "Needs review", and they will not advance until that happens. When the set of "needs review" topics balloon, I have to stop picking up non-trivial topics to make time to review them myself.

Thanks.