[Aboriginal] Random notes.

[Rob Landley]

1) I've gone back to blogging extensively about working on this since
nobody's replying here.  (Instead, people are emailing me directly, as
usual.  Sigh.)  Feel free to reply to stuff I blog about here on the
list, quoting as much as you like.  It's just easier for me to grep my
notes files than try to search the mailing list archives, and if
nobody's going to engage I'll do what's easy for me.

2) I may wind up becoming part of the funtoo core team, which is the
Gentoo restart Daniel Robbins did when he wandered back and found Gentoo
had been buried alive by Debian developers who brought their flamewars
with them when they looked around for alternatives during the five year
"debian stale" period where debian was too constipated to get a release out.

Daniel agrees that Funtoo's portage tree should be as target-agnostic as
possible, and the roadblock I hit trying to bootstrap gentoo was that
every single PACKAGE had a list of targets it was verified to build on,
so adding a new architecture meant touching every single package file in
the entire tree, meaning nothing it did could NEVER be target agnostic.
 Funtoo is fixing that, and if I can make it work Aboriginal Linux may
become an official part of the Funtoo bootstrapping process as a sort of
stage 0.  That would be way cool.

3) I got lfs-bootstrap to run on all the arm targets, both mips
endiannesses, powerpc, and all the x86/x86-64 variants.  This means
every target that hasn't got something majorly wrong with it (such as
armeb or m68k having no qemu support, or sh4 only allocating 64 megs of
ram no matter what -m says) is now building Linux From Scratch 6.7.

That last one calls for a release, so I'm polishing.  I get on a plane
Tuesday to hang out with my apprentice for a week and may be
incommunicado for a chunk of that, so said release might be tomorrow.
Or it might be at the end of the month.

My pending biggish todo items after the release are:

1) funtoo-bootstrap

2) update LFS to 6.8.

3) Do "infrastructure", "networking", and "multimedia" BLFS bootstrap
control images as described in
http://linuxfromscratch.org/pipermail/blfs-dev/2011-July/021073.html

4) Break down and do more current toolchain build scripts, even if I
don't distribute the resulting binaries because they're not under a
license compatible with the Linux kernel.

I'm waffling between working on #2 and #3 (both low-hanging fruit and
"shortest job first" scheduling choices with no obvious unknowns and
nice payoff) vs working on #1 on the theory that actual humans are
waiting for the result and would help me over the hard bits.

Currently, #1 is winning.  (And #4 is on the list because Daniel Robbins
wants it, so it hangs off of #1...)

Rob