Re: build system work

[Graham Percival]

On Fri, Mar 04, 2011 at 10:23:36PM -0700, Colin Campbell wrote:
> Actually, and before I go stepping on toes in all directions, I
> wouldn't mind some guidance on a couple of things, please.  CG 12.2
> Meisters describes the Frog Meister as being responsible for code
> patches from (relatively) inexperienced contributors, basically
> shepherding the patches through the review and acceptance process
> and pushing when finally accepted.

Yes.  We're not really doing that.

> However, CG 8.6 Patch Handling
> talks about a Patch Meister with possible helpers; is this what you
> were talking about, Graham, and is there any duplication here?

CG 8.6 takes precedence.

At the moment, I am the Patch Meister.  Marek adds new issues with
the tag patch-new.  You will check those patch-new issues, and any
updated versions of issues with patch-needs_work.


> I've cleaned out a lot of old -devel mail, so I don't recall the
> results of a discussion a year or so ago about patch handling,
> and I've no interest in adding more bureaucracy.

There were no results, and for that reason I am do not propose to
open the question again unless we have solid reason to believe
that the discussion will really change stuff.

> Another possible discussion would centre on the link between the
> bug squad and patches: in many ways, patches arise in response
> to bug squaddies creating issues, so perhaps Phil Holmes needs
> to have input here.

Yes, but such discussions should take place during GOP.  We are
one issue away from having a new release candidate, and I think
we've had more reviewers and a healthier development community in
the past few weeks than we've had in a long time.  I don't want to
open up a huge discussion right now.

> I also don't think a patch meister would need help, given the
> role of the bug squad and the devel community:

Checking that patches don't break regtests is a task that
currently occupies about 2 hours per week.  If you do that for me,
I can spend those 2 hours doing something else.  End of story.
It's not going to change the world, but it's a measureable
improvement to the current state of affairs.

> it's mostly a nanny function, just making sure we all play nice
> and look after each other.

Yes and no -- long-time developers will probably check the
regtests themselves, although we all forget from time to time.
The main facet is checking patches from new contributors.

> At any rate, how do you feel about creating a new mail account
> called address@hidden which would be forwarded to whomever has
> the Patch-Meister role.

That's not a bad idea, but I don't think we should be changing the
development process at the moment.  We can discuss this during the
Grand Organization Project.

I'm not certain if you saw this part of the CG; if not, it may
clarify my reasoning for postponing such discussions:
http://lilypond.org/doc/v2.13/Documentation/contributor/policy-decisions

Cheers,
- Graham