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

Re: Canonical and Debian

On 6/5/05, Steve Langasek <vorlon@debian.org> wrote:
> You can either step up and make sure the
> architectures you care about are in good shape for etch, or you can be a
> whiny brat expecting everything to be handed to you on a silver platter and
> accusing people of being members of a Canonical-controlled cabal when they
> do you the courtesy of informing you about their personal priorities for
> etch.  Your choice.

I am no fan of the Vancouver proposal, but Steve's got a point. 
Ensuring that packages build and run properly on a wide variety of
architectures is _work_.  I happen to think that it's worthwhile work,
and that it's the main factor that sets Debian apart from all the rest
and directly contributes to the superior quality of Debian relative to
other distros.  But if it isn't spread across a large number of
people, it's a crushing burden, and no one has a right to ask the
release team to shoulder it.

The mirror network is not the big issue, as I see it; I care more
about the question of whether the build procedures have adequate
conditional logic to handle the presence/absence of a native-code
compiler for language X, the existence or lack of an assembly-language
implementation of core routine Y, etc.  As I have argued previously,
the diversity of architectures is the best available proxy for the
evolution of platforms over time, and the packages which have a hard
time building on all arches are precisely those which it's a struggle
to maintain for the duration of a release cycle.

Steve's also right that buildds that have non-zero queue depths for
non-trivial lengths of time tend to expose fragility in build and
run-time dependencies, and so they get stuck in ugly ways that need
release team (or porter) attention.  So either Debian collectively is
willing to labor to maintain a high standard of portability and
stability, or we need to focus on a few arches and ignore
bugs-in-principle that don't happen to break on those systems.  I know
which one I'd like to see, but I have to admit that I have done a
lousy job of following through so far on things that I could help with
myself.  But at least I don't go around blaming the people who are
actually doing the work.  :-/

- Michael

Reply to: