Re: The pool system and debian-cd
On Sat, 16 Dec 2000, Philip Charles wrote:
> On Sat, 16 Dec 2000, jason andrade wrote:
> > On Sat, 16 Dec 2000, Philip Charles wrote:
> > > Does any one know if debian-cd will work with the incoming pool system?
Well, we won't know until someone tries ;-)
> > just to check.. i'm curious too.. are we both taking about
> > make-pseudo-image ?
> I hadn't considered pseudo-image. My understanding is that potato will
> still remain as such, but it will be made up of symlinks to pool, so I
> would expect speudo-image and debian-cd to work with potato (possibly
> needing a little tweeking).
The Pseudo-Image Kit itself will work regardless of the archive structure, but
creating the correct .list files may get a bit more tricky. (The easiest
solution would be to have pool/ available on the CDs, with packages either
hardlinked to binary-$ARCH/all or not, but I don't know if we want that.
I myself do like the alphabetic hashing better than section-based ordering
(guessing the correct section is hard in some cases), but then maybe we
shouldn't call it pool/ but have the alphabetic hashing right under
> Again my understanding, woody and sid will be a file lists that will
> extract the various packages from pool, and binary-all etc will not exist
> as such. The problem I am facing right now is that hurd is in sid and is
> accessed through woody. The future problem for debian-cd is very much my
> present problem with the hurd CD.
What you can do is grep Architecture:, Section: and Filename: from the
Packages file(s), and with that info create a second, hard-linked version of
the archive that is structured in the old-fashioned way. (Actually, Section:
can't always be trusted; extracting the section from Filename: works better.)