Alexander E. Patrakov wrote:
OTOH, is it really any point in keeping non-i386 architectures in trunk?
1) the resulting CD is very different
2) a lot of ifdefs
3) x86_64 was just declared to be unmaintained
4) sparc64 CD simply doesn't build because of a forgotten "svn add silo".
For the record, this is not forgotten. Just never added because I
haven't had time to finish testing it. It's still there in my working
copy waiting to be added after I've tested it.
I don't mind at all the idea of simplifying the scripts. The idea
originally was that with minor fixes/upgrades, we wouldn't have to be
merging all the time to separate branches for the different architectures.
Here's what I'd like to propose, and tell me if you think this is
workable and beneficial:
We move out sparc64 and x86_64 to their own branch and leave x86 and ppc
together. Then we can start a re-write based on Alexander's POC. I think
this is fair, don't you? ppc, as you've already said, is similar enough
to x86 in the way it builds in that *very* minimal adjustments or
additions would need to be made. For that matter, due to the
improvements made with glibc and sparc I could work on adding that in
again, too. :)
Anyway, the biggest problems come from the cross-build techniques,
right? If we split trunk into 32-bit platforms and 64-bit platforms we
can simplify it a great deal wouldn't you say? And if you're worried
about breaking an ARCH, for now, don't. When we do the re-write, focus
on getting x86 done and the rest will follow as I, or anyone else that
wants to help, is able to test the other ARCHs.
--
JH
--
http://linuxfromscratch.org/mailman/listinfo/livecd
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
