Re: Freeze exceptions: parmetis, ccc, babel, illuminator; please be considerate to busy developers
On Wed, 2005-05-18 at 13:43 +0200, Luk Claes wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Adam C Powell IV wrote:
> > Greetings,
>
> Hi Adam
>
> [...]
> > I must say, as a Debian developer at a US university, I feel this freeze
> > process has treated me very inconsiderately. We were told for about
> > eight months that a freeze was "on the way", which after four, five, six
> > and seven of those months didn't sound credible. Then suddenly in one
> > of the two busiest times of year at US universities (final exams in May
> > and December), we're given an ultimatum: get your packages done in two
> > weeks or they're out of sarge. The pressure has forced me to divert time
> > from writing and grading exams (babel in particular took the better part
> > of a whole day) and I am not happy about this at all. Because of the
> > propensity of academics (students, faculty and staff) to be free
> > software developers, and the fraction of Debian developers in the US, I
> > am certain I'm not alone in this.
>
> You had 7 months to make sure the packages you uploaded are ready for
> sarge... You only have 2 weeks to fix RC bugs that *still* don't make
> your package ready for sarge...
Okay. Perhaps it's instructive to contrast this with AJT's
"woody-sucks" practice during the woody freeze, when RC and non-RC bugs
-- even new upstream versions -- were accepted. This time, we had
"sarge is gonna freeze someday", "sarge will be ready in a month or
two", then three months later, "we're getting closer to the sarge
freeze", then all of a sudden, "the security infrastructure works, sarge
is frozen, no more non-RC bugs can get in, period". Can you understand
why after seven months of "crying wolf", some developers would want a
bit more time to finalize recent fixes?
> > Given that processing of the NEW queue has been extremely slow over the
> > past several months (e.g. mpich lib restructuring sat there from Dec 22
> > to Feb 19), which only picked up *after* the freeze, how can the release
> > team treat developers this way? "Your packages are not important enough
> > to process in less than two months, but you *must* turn around your last
> > fixes for sarge with just two weeks' notice during the busiest time of
> > year for a significant fraction of you."
>
> This problem has nothing to do with the release per se and has been
> solved as good as possible for the future.
Right, but that has nothing to do with the fact that its recent
brokenness has impacted sarge directly, effectively shortening that
"seven month freeze" time enormously. When I had the time to put into
new petsc and illuminator packages (January), I could not do anything
because mpich was stuck in NEW. I finally got time to work on petsc in
April, and struggled to upload illuminator 0.9.1-1 just after the freeze
started.
And a broken sparc buildd is now keeping illuminator out. Should that
not have been fixed before the freeze? Can we be sure that the security
infrastructure will build similar packages properly?
> PS: Please don't see this as an attack or something to start a flame
> for. I can understand that you think that the freeze came suddenly, but
> on the other hand there was in fact already for 7 months a pseudo freeze...
I understand this is not an attack or flame, but a response in a
civilized debate. :-)
So, Colin Watson said in March 2004 that d-i should be done by the end
of May. Suppose that you had been prevented from making progress on the
release by infrastructure breakage during May and June, and d-i had
missed deadline after deadline. Suppose that with no other freeze
announcement, someone had said in September 2004, "Okay, d-i is
finished, the release team now has two weeks to fix the security
infrastructure, make sure upgrades work smoothly, and get the release
out the door". Finally, suppose that you and a significant fraction of
the release team were caught in an unrelated corporate restructuring, in
which any deficiency in your performance would impact your future career
(less-than-perfect analogy to final exams at US universities).
How would that announcement have made you feel?
As I said, I'm just looking for a bit of leniency and patience in this
process.
Thank you,
-Adam
--
GPG fingerprint: D54D 1AEE B11C CE9B A02B C5DD 526F 01E8 564E E4B6
Welcome to the best software in the world today cafe!
http://www.take6.com/albums/greatesthits.html
Reply to: