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

Re: Some observations regardig the progress towards Debian 3.1

On Thu, 20 Nov 2003, Graham Wilson wrote:

> On Wed, Nov 19, 2003 at 05:34:53PM +0200, Martin-Éric Racine wrote:
> > 1) experimental
> > 
> > The hacker playground. Dodgy uploads allowed. No guarantees to anyone  
> > on the sanity of anything there.
> > 
> > 2) unstable
> > 
> > * Whatever was thoroughly tested by developpers in experimental and is  
> > considered ready for the masses, is cautiously uploaded here. New glibc  
> > versions that have been thoroughly debugged would fit this category.
> > 
> > * Trusted packages entering Debian for the first time and which were  
> > built against Testing dependencies also enter the release chain here.  
> > New releases of Evolution based upon GTK2 would fit into this category.
> > 
> > Architectures don't have to be in sync, although an attempt is made to  
> > build on all of Debian's supported architectures, as a first exercise  
> > in QA on the package; RC bugs are fixed and a new build is uploaded,  
> > until the package has passed the usual 14-day rule and finally builds  
> > on all architectures, at which point it trickles down to Testing.
> >
> > [...]
> >
> > Hopefully, the above made sense.
> I don't know if I agree with the idea of building against testing
> (though I believe that is not exactly what you said). 

I indeed said exactly that:  always build against what has already made it to

> However, I really like your ideas about unstable and experimental. Maybe
> experimental should be used more often to decrease the churn in unstable. Of
> course, there are some issues with using testing, namely, the lack of
> autobuilders, and difficulty with pinning.

Lack of autobuilders can be remedied.  Pinning is of course difficult when
package versions are in constant flux.

> The first could be addressed by individuals committing time and machines
> to build for experimental. I can volunteer to build for alpha, i386,
> mips (soon), and powerpc.

Why can't everything just be cross-compiled on a _really_ fast 64-bit host
belonging to the Debian project?  Couldn't e.g. IBM or HP (who both support
Debian in some way or another) allocate some hosts for that, for instance?

> The second could be addressed by someone writing a short howto on how to set
> up the preferences file for pulling certain packages from experimental.

...or to safeguard oneself against unfortunate incidents.  I use the following
preferences, by default, on all my hosts and on those I install for others:

Package:  *
Pin:  release a=stable
Pin-Priority:  1001
Package:  *
Pin:  release a=testing
Pin-Priority:  101
Package:  *
Pin:  release a=unstable
Pin-Priority:  99
Package:  *
Pin:  release a=experimental
Pin-Priority:  9

This ensures that selected packages can be easily pulled from Testing and that
others can be also pulled from Unstable, if they don't require any dependencies
that cannot be satisfied by Testing.

IMHO, these defaults should be installed by debootstrap, on a virgin system.

Martin-Éric Racine, ICT Consultant

Reply to: