Re: Testing migration doesn't check build-depends?
Le May 13, 2008 02:53:05 am Steve Langasek, vous avez écrit :
> On Mon, May 12, 2008 at 09:19:17PM -0400, Filipus Klutiero wrote:
> > > > > Sorry for a stupid question, but how is it ensured in the Debian
> > > > > project that all packages in stable are buildable in stable?
> > > >
> > > > The Debian project doesn't verify that. There are packages that FTBFS
> > > > in stable. However, it's rare, because there are never large numbers
> > > > of packages that FTBFS in testing to begin with.
> > >
> > > The Debian project does verify it. That doesn't mean there are no bugs,
> > > but they're clearly bugs.
> > There's a difference between having some release team members act at
> > times on debcheck and/or FTBFS bugs and having project infrastructure to
> > "ensure that all packages in stable are buildable in stable". I didn't
> > mean that nobody does anything to avoid these bugs or that FTBFS weren't
> > considered bugs.
> They're considered *RC* bugs, with all that implies. This is trivially
> verifiable from <http://release.debian.org/etch/rc_policy.txt> or
> <http://release.debian.org/lenny/rc_policy.txt>, either of which I would
> expect someone to be conversant with before responding to questions on
You must be conversant enough with release management to know that a bug being
considered RC doesn't make it block the release. Also, reports aren't filed
automatically, so only the bugs someone found the time to investigate are
> > > If you don't understand the way release management actually works (and
> > > it does work), please just don't answer on mail to debian-release.
> > If you prefer denigrating my contributions to answering questions to the
> > list, please do so more convincingly.
> I believe what he was denigrating were your research skills.
I'm less interested in speculating about what he was denigrating than to see
good reasons for denigrating whatever skill I would lack.