[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian needs more buildds. It has offers. They aren't being accepted.

Ben Collins <bcollins@debian.org> writes:

> > > No, he said "you need to replace them".
> > 
> > He also said "cannot".
> If they can't put up a few buildd's, I'd hardly call that grounds for
> replacement. We have more pressing issues than setting up buildd's (like
> the security of our systems). And if they don't have the time to do that
> (currently, so consider that their time constraints may change), what
> makes anyone think they'll have time to train someone in doing the
> buildd setups?

They don't have time to add a user to the access list of wanna-build
(and the acceptes/autobuild repository, but thats less pressing).
Thats a 1 minute job anyone could do.

Its also nothing security relevant since all anyone could do with a
wanna-build access is disrupt the build process for a while and delay
a few packages from being build. It would affect Debian but in no way
endanger its security. If the argument against giving wanna-build
maintainance away is that it runs on the master archive then that has
to change. W-b can run everywhere that has _read_ access to the
Packages and Sources files. Also note that a team with spare time
maintaining wanna-build could far better watch thats its not being
abused in some way.

We have enough people willing and able to setup and maintain buildds
and the ones they have setup and configured idle because the last bit,
the wanna-build access, is missing.

They seem unable to do the job (for whatever and however valid
reasons) and thats all the reasons needed to add more people to the
job or replace them. The Wanna-build and buildd maintainance is
something that can be outsources wonderfully and can be managed by a
big team as m68k shows. Adding such a bottleneck to it is just
stupid. Why burden people that have too much work on their hands with
this too?

Whats anoying everyone (well, the people James just ignores) most is
that James is online (talking too) on irc and gets asked what the
problem is and he just ignores the questions, repeadately. Then 2
weeks later you have to be told by _someone_ _else_ that a mail that
was send probably used an old, retired gpg key of james and should be
resent. I can't say if Ryan behaves the same since I haven't seen him
on irc but both seem to ignore mails from people they don't
like. Thats fine personally but not when it affects the Debian
project like this.

A simple responce on irc that the mail was send wrong or that he
doesn't have time this week is all that it takes to keep the
blood pressure down. If his workload is to much (even only temporary)
for that one sentence its unlikely it will get low enough anytime
soon for him to do the job timely.

But there is an easy way out of this. We can setup an alternative
wanna-build and setup a bunch of mips and mipsel buildds as well as
reconfigure the m68k buildds to use it. After that the mips/mipsel
buildd pool from us would compete against Ryans buildds and he, as well
as we, would get a lot of "Rejected" mails from katie due to duplicate

But is it neccessary to go that far? A little cooperation would turn
those potential REJETED mails into more ACCEPTED mails.


PS: Maybe its just IJ and me that get ignored but we both feel
mistreated. We spend our time and money for debian and get neither
recognition nor respect for it.

Reply to: