Re: [gentoo-dev] Herds suck, fix them

[Chris Gianelloni]

On Thu, 2006-06-15 at 12:34 +0200, Jakub Moc wrote:
> Marcus D. Hanwell wrote:
> > I don't know if this is a really unpopular viewpoint, but for a lot of stuff I
> > maintain I put myself as maintainer and the herd I am acting as part of in
> > herd. My intention there is to say primarily I am taking care of this and
> > have taken responsibility but if I disappear, am slow or someone else just
> > wants to bump it etc in that herd then they are also free to do so.
>
> Well yeah, that's how I read the metadata.xml in such cases... but since
> some people are suggesting that <herd> is not relevant info wrt
> maintainership, this attempt for clarification has been proposed.

*sigh*

Who said that?

I have not seen that said, *at all* in this thread.  What I *have* seen
said is that whomever *maintains* the herd is the package's maintainer,
except in the case where an explicit maintainer is listed.  In almost
all of the cases, this is the same email alias as the name of the herd.
When you assign bugs to games@g.o it isn't because the people on
the alias are a part of the games herd, it is because the people on the
alias *maintain* the games herd.  It is a subtle distinction, and one
that actually doesn't change your practices, since you've already been
doing everything "correctly" this entire time.  An example of this *not*
being the case is apache.  You still give the apache bugs to the right
people, right?  Of course you do, because apache-bugs is listed as the
maintainer.

> > May be it would be more correct for me to add the herd alias as a second
> > maintainer? I think it is good for people to take responsibility for what
> > they add to the tree and that is my intention there...
>
> :=) If a general consent is (games left apart ;) that <herd> is a backup
> for cases when maintainer is unavailable/goes MIA, and a primary
> maintainer if there's no <maintainer> tag in metadata.xml, let's just
> leave it at that, be done with it and save ourselves the hassle...

Why is it that you are restating *exactly* what I am saying, but then
trying to pretend like I'm not saying what you are saying?  The games
team has nothing to do with this, because I am saying that we work
*exactly* as you expect us to.  What I am saying is that the *herd* is a
collection of packages.  The email alias associated with the herd, goes
to *people* not packages.  The alias could have any number of people, or
other aliases, as its contacts.

I guess I need to spell this out so there's no more ambiguity.

A package belongs to a herd.  The herd is listed in metadata.xml, like
it should be.  If there is nobody listed in maintainer explicitly, you
email the alias associated with the herd.  This email goes to the
*maintainers* not the *herd*, since the maintainers are people, and the
herd is packages.  Each herd has *at least* one maintaining project,
team, or group of people responsible for it.

> If we can't agree upon this, then we probably should stick herd alias
> into <maintainer> tag when that herd _is_ actually willing to act as a
> maintainer.
>
> More clear now? :)

No.  You've gone and changed the practices we have in place now to make
it more complicated.

Say it with me.

Herd == packages
Team == people

Good.  Now say "This requires me to make no changes to how I operate."

Very good.

Somebody give this man a cookie!

-- 
Chris Gianelloni
Release Engineering - Strategic Lead
x86 Architecture Team
Games - Developer
Gentoo Linux