> Sorry for having been this silent. In the past few days I've spent many hours
> on getting debian-cd ready for 2.2 rev3 (issues you mentioned, updated/
> redesigned README (matching www.d.o but actually better code) and the
> long-promised "make-a-useful-CD1" project which involved lots of test runs and
> a new unexclude feature). I'm mostly finished now, a few things still need
> some tweaking, then I'll leave it to Phil to generate the new .iso's.
>
> The unexclude will be used to move any number of kernel sources to CDx, x>=2;
> the source of only 2.2.19 will be on CD1. However because we currently have
> 10 (!!) kernel sources in potato, it looks like we're going to have a 4th CD
> with only 15MB on it. So, unless anyone objects, I think we'd better NOT
> have the sources of 2.2.10 and possibly 2.2.12 on any CD (which still leaves
> us 2.2.13, 2.2.15, 2.2.17, 2.2.18pre21, 2.2.18, 2.2.19pre17 and 2.2.19 -- and
> 2.0.36, 2.0.38 of course).
AFAIK, the only kernels worth having in either binary or source
are (in release order) 2.0.36, 2.0.17, 2.0.19, 2.4.2. All others are
either junk and pretty unstable or useless (at least IMHO).
So, I've never really seen any use in having ancient sources in
the *distribution* cds and I fear things will become worse with
woody (5 CDs binary installation ?).
Perhaps it would make more sense to build two different "master" tasks -
one for building the installation cds (with everything *necessary* and
stable on them), then some kind of task for "other" optional Debian
CDs - having things like old kernel sources, ancient compilers, compat-libs
not depending on the stable stuff etc. etc.
Just my $0.02 ;)
Bernd
--
ixsoft Softwareentwicklung und -vertrieb Bernd Hentig
Unter den Buchen 22E, D-16244 Altenhof, Germany
Voice ++49 333 63 46 100 Fax ++49 333 63 46 109
Email [EMAIL PROTECTED]
