Re: [Sbcl-devel] Re: PPC bug in ASH?

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

Lars Brinkhoff <la...@no...> writes:

> Christophe Rhodes <cs...@ca...> writes:
>> Lars Brinkhoff <la...@no...> writes:
>> > Christophe Rhodes <cs...@ca...> writes:
>> >> > As far as I can tell, shifting by -32 or less will only shift right by
>> >> > 31 bits.
>> >> You're quite right.
>> > Same thing with constant shifts
>> Yup
>
> Unless someone else already did it, I plan to contribute a patch for
> this.  Shouldn't be too hard, once I know how to run the tests on a
> new build.

I think, actually, that I've fixed the problem on ppc, x86 and sparc.
If sourceforge CVS were less sucky, this would be obvious to
non-"developers"; as an interim solution, listening in on the
sbcl-commits mailing list might be a good way for non-"developers" to
see what is going on.  (I put a source tarball of sbcl-0.8.3.38 up at
<http://www-jcsu.jesus.cam.ac.uk/~csr21/sbcl-0.8.3.38-source.tar.gz>
since Rainer Joswig wanted access to current sources; that may be
useful to other people).

(I've got a build in progress for mips(el); I don't have easy access to
alpha, and our hppa port isn't complete.)

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)