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

Re: Debian 2.2 Release.



On Tue, Jun 27, 2000 at 11:51:38PM +1000, Anthony Towns wrote:
> On Tue, Jun 27, 2000 at 08:58:33AM -0400, Daniel Burrows wrote:
> > > What about freezing woody one week after releasing potato
> > > and releasing it after no more than one month of freeze ???
> 
> Heh.
> 
> Note that powerpc and sparc autobuilders have deliberately been ignoring
> woody (and I think arm autobuilding doesn't entirely exist yet, so they're
> out of date too); note that there are large numbers of bugs being ignored
> pending woody, and large numbers of new packages that haven't been tested
> for woody yet. Compare the uninstallables counts:
> 
>                   potato  woody
> 	alpha        30    338
> 	arm          85   2783
> 	hurd-i386          586
> 	i386          3     97
> 	m68k         40    331
> 	sparc        21    409
> 	sparc64            183
> 	powerpc      28    430

hurd-i386 isn't intended to be released with woody.
And both your and I know that 95% of above are trivial.

> You really think we'll get all this fixed and spotless in a month?

95-99% packages fixed, 1-5% packages not released.
Have we ever released 100% packages ?

> It is to laugh.
> 
> Personally, I think Debian's outgrown the "develop / debug / test /
> release" model at this point. There are too many arches to try to keep
> in sync, too many packages to ensure are bugfree, too many independent
> goals to work on, too many configurations to support if we try to do it
> all in series: as soon as we've got B settled down and try to move onto
> C, A breaks. We need, IMHO, more parallelisation, not just more pressure
> to get things done faster and better and what not.
> 
> I'd strongly encourage people to hold their thoughts on how we should
> handle releases until we've actually finished this one, ie, when we can
> actually take some of the ideas everyone has into account. Write your
> messages and all, but lets postpone them for a month and actually do
> something rather than just talk about it this time.
> 
> >   Another idea might be ajt's "testing" distribution.  From what I've heard,
> > that has a lot of promise for alleviating some of the problems; I'm interested
> > to see what happens when it's integrated into the archives (I've heard this
> > will happen after the woody release?) 
>                         ^^^^^
> current or potato. ie, it's ready and waiting to be in place *for* woody.

Of course there are better ways.
Better ways were already proposed 666 times.

Is any change planned for woody ?



Reply to: