Re: `Every package must have exactly one maintainer'
Christian Schwarz <schwarz@monet.m.isar.de> writes:
> > Questions instead go to the new-maintainer alias and one of us
> > answers it depending on who isn't busy at the time. This is, I
> > would have thought, obviously a better situation.
>
> Note, that we are talking about `package maintenance', not other
> developer's duties. Facts are different if we talk about
> `webmasters,' `listmasters', etc.--this is not what this discussion
> is about.
What's the difference between maintaining a package and processing new
maintainer requests when it comes to the benefits and problems of
multiple maintainers?
Is it that it's convenient for you to pretend there's a difference
because then you can ignore all the examples of multiple
maintainership where things *work*?
It comes down to this: there are currently two packages in the
distribution which are maintainer by multiple maintainers: dpkg and
boot-floppies. There will soon be a third (debian-keyring). You were
and are wrong about boot-floppies (IMO), and *no one* has even begun
to prove that dpkg's is in the state it's in because it has multiple
maintainers. Yes, dpkg's maintenance \begin{litotes}isn't
great\end{litotes} but you are scape-goating multiple maintainers by
blaming it on that. Please show me some evidence that had dpkg
continued to be maintained by Ian alone, it would be in better shape
now than it currently is.
> > > 2. Having only one person listed in the "Maintainer:" field
> > > does not mean that only one person works on a package! It
> > > only means, that there is a unique person who coordinates
> > > all changes.
> >
> > So if that person gets busy no changes can be coordinated... baz
> > bat bamus batis bant.
>
> In which case we could define a `Backup-Maintainer'.
And if he's busy? Sorry, but your arguments are circular and bogus.
If there is a single person responsible for coordinating changes he
becomes the critical link in the chain, if he becomes busy, nothing
happens. If you define a backup, it only takes the primary and the
backup to become busy before again output stops. What next? Make a
Backup-Backup-Maintainer:?
With multiple maintainers there is no critical link in the chain,
anyone person can step forward to take up the slack. Output stops
only when all maintainers are busy, this is no loss over single
maintainership or single person responsible for coordinating changes.
> (What does the `baz bat...' mean?? This doesn't look like you are
> taking the discussion seriously.)
Please don't try to be so condescending, it doesn't work.
> > > Unfortunately, we still have communication problems with the two
> > > multi-maint packages we have (dpkg and boot-floppies).
> >
> > You're:
> >
> > a) ignoring the complexities of the two packages and the fact that
> > they're native Debian packages and thus have no upstream
> > authors to fix complex bugs.
>
> Fixing the `dpkg-dev is missing the debian-changelog-mode.el file'
> is trivial--and the bug is very annoying.
It could have been fixed by either Miquel or Juan, and they didn't
(one assumes because they didn't know/forgot about the bug), maybe
that indicates it's not quite the disaster for humanity you paint it
to be.
> I'd agree to you if dpkg would only have complex bugs--but that's
> not the case.
Obviously it's not the case, but the point is you aren't taking into
account all the complex bugs dpkg does have. (Have you actually gone
through dpkg's bug list or are your comments about as well informed as
they are WRT to boot floppies?)
> > b) picking on boot-floppies for no good reason (what exactly is
> > wrong with it? Okay, there are a lot of bugs, but there's a
> > lot of weirdo hardware out there. There is a lot of active and
> > good development, people responding to bugs, new versions
> > etc. etc., it's no comparison to dpkg).
>
> Maybe, I was wrong. I just remember that a) you (as m68k-porter)
IHYM, "as *an* m68k-porter".
> mentioned coordination problems with new features on the bootdisks
That was an offhand comment that I made *months* ago, in private mail,
which you misinterpreted and has zero relevance to this discussion[1].
> and b) every few months someone asks who is responsible for the
> installation manual.
The installation manual is a minor part of bootfloppies; or would you
orphan emacs if the manpage wasn't uptodate?
[1] If, IIRC, I said something along the lines of that bootfloppies
was hard to track because of the rapid changes. There were no
coordination problems and there were no communication problems,
bootfloppies was simply developing quite fast at the time, making it
hard to keep up on m68k. I'll dig out my mail archives if you don't
believe me.
--
James
--
To UNSUBSCRIBE, email to debian-policy-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: