[LFS Trac] #2094: Add a new section for build results

Ken Moffat ken at linuxfromscratch.org
Sun Oct 21 17:58:04 PDT 2007

On Sun, Oct 21, 2007 at 06:13:36PM -0600, Gerard Beekmans wrote:
> > of the differences in the build methods. I'm all for LFS having support 
> > for other architectures, but they need to have the resources available 
> I'm only intending to support x86_64 type stuff as it's the main stream 
> new computer systems these days. I definitely do not plan on doing 
> anything with PPC, Sparc, RISC or something "fancy" like that. We don't 
> have the resources, which pretty much ends that discussion.
 Hmm.  Experience (both questions asked on lfs-, and lack of clfs
testers) shows that some readers with other architectures ( I
remember particularly ppc and s390, but I'm sure there were others)
prefer native compilation and are reluctant to try pseudo
cross-compilation.  In other words, they prefer LFS to clfs.  That
isn't entirely surprising (clfs is targetted at experienced builders,
and in some ways is a successor to the lfs-hackers list with a
greater willingness to try alpha versions of things like 'patch').

 Personally, I've got fed up with parrotting the party line to
people with ppcs that they should use clfs, because some builders
plainly don't want to.

 But what sort of resources are you talking about ?  Building those
other architectures, or providing support ?  The thing about support
is that it's always a matter of luck whether anyone has already seen
the unusual problems (e.g. there was some problem on x86 related to
building from ubuntu, if I remember correctly, that took months to
get correctly diagnosed), and for arch-specific issues a few people
almost certainly will still ask on lfs-support.  But, if you *want*
to kill the discussion at an early stage...

 Of course, given that LFS is always constrained by limited
resources, the project could be concentrated on x86_64.  Anyone who
wants to build for x86 could be directed to clfs ;-)

das eine Mal als Tragödie, das andere Mal als Farce

More information about the lfs-dev mailing list