Re: [Sbcl-devel] MIPS compilation failures

[Christophe R. <cs...@ca...>]

Kevin Rosenberg <ke...@ro...> writes:

> I realize that mips/mipsel are not commonly used platforms for
> SBCL. If there are ideas on fixing this, I'll be glad to help as well
> as the Debian Project Leader who has a convertible mips/mipsel system.
> Alternately, so as to not hold up the progress of SBCL from Debian
> unstable to Debian testing on x86, powerpc, and sparc, I can ask the
> Debian ftpmasters to remove the SBCL mips/mipsel from the archive.

At a guess, these build failures are probably indicative of a backend
bug: somewhere, there isn't sufficient waiting for results to arrive
in memory or in registers.  Fixing this, if I'm right, will probably
involve finding people with time, detailed knowledge of the mips
platform, and motivation.  (However, I'm not ruling out something
simpler: maybe we're treading on a newly-placed library space, or
something; look at /proc/<pid>/maps for emacs, and look for suspicious
overlaps with the spaces in src/compiler/mips/vm.lisp).

Cheers,

Christophe
-- 
http://www-jcsu.jesus.cam.ac.uk/~csr21/       +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%")    (pprint #36rJesusCollegeCambridge)