Re: powerpc 2.2r4
Raphael Hertzog <email@example.com> writes:
> Le Fri, Nov 16, 2001 at 05:57:45AM -0900, Ethan Benson écrivait:
> > > If so, any ideas what I did wrong?
> > its not your fault its mine, for one i was under the apparently
> > mistaken impression that older versions of debian-cd were uses for
> > potato anyway (someone i think debian-cd's maintainer told me this).
> Yes, and it was so for 2.2r3. Anne had setup a CVS branch (off22r3) for
> building potato images ... i've even been bashed because I
> broke debian-cd CVS without tagging a good version that was usable to
> build potato CDs.
> For this time, it will be ok. But we'll try to do better for woody ...
> using debian-cd CVS for building an already released Debian version will
> always be problematic. Documentation included changes, CD characteristics
> also, even boot images used can change ...
To make this more problematic, there is the fact that I've upgraded
open.hands.com (cdimage.d.o) to woody, in order to be in a position to
build woody CDs. This means that the old debian-cd doesn't like the
new options for mkisofs, among other things.
Perhaps I should set up a chroot environment for building CDs, so that
they are always built using tools available in the release, rather
than whatever happens to be on open at the time.
On the other hand, if mkisofs or some other tool has a bug fixed that
makes the resulting images better in some way, is there really any
point in producing images that are not as good as they might be.
Also, I get the impression that the changes like those that move junk
off CD1 are only in the latest branch, which would mean either
duplicating the patching & testing effort to get it back into the old
debian-cd, or that the r4 CDs would have had as many copies of the
kernel on CD1 as the r3 ones.