[Sbcl-devel] next steps [was RE: Titivillus Day ...]

[Martin A. <ma...@at...>]

William Harold Newman wrote:

[...]

> FYI, I plan to take a shot at the following reasonably soon, so if you
> are tempted to work on any of them, you might want to let me know so
> that I don't duplicate your efforts.
ok.

> 0.6.12.??:
>         tweaked SBCL along the lines of Raymond Toy patches from
>                 cmucl-imp:
>         ?? 2001-04-09: #'< for floating point infinities
[The real patch was on cmucl-imp 2001-04-12 (patch of 2001-04-09 was
broken)]
Actually, Looking at the code, this already seems to be implemented in
SBCL.
The test case, that you gave on 2001-03-22
(< single-float-positive-infinity 100)
also works.

The log entry for target-numbers says:
0.6.11.29:
        fixed bug 92: (< SB-EXT:SINGLE-FLOAT-POSITIVE-INFINITY 100)
                now works, thanks to Raymond Toy's patch to CMU CL.

Am I missing something?



>         ?? 2001-04-10: :EXTERNAL-FORMAT option for LOAD
I'll look into this, soon.



>         ?? 2001-04-12: various stuff related to COMPLEX and LOG
>         ?? ??: Wait for X86 floating point exception in ??
>         ?? 2001-04-30: printing FLOATs readably
I think I'll leave that to you/others. :)



> 0.6.12.??:
>         ?? deleted unneeded extra templates in src/pcl/walk.lisp, as per
>                 Fred Gilham's note cmucl-imp 2001-04-27
I looked at the CMUCL cvs: the extra templates were just commented out
there, which looks like:

;;;
;;; And the extra templates...
;;;
#|| These are macros in CMUCL
(define-walker-template DO      walk-do)
(define-walker-template DO*     walk-do*)
(define-walker-template PROG    walk-prog)
(define-walker-template PROG*   walk-prog*)
(define-walker-template COND    (NIL REPEAT ((TEST REPEAT (EVAL)))))
||#

They are not even present in the SBCL code. So this should be already
done, too.
Maybe the relevant walk-* functions can even be deleted, but I'm not
sure.



> 0.6.12.??:
>         ?? Tim Moore's improved disassembly patches 2001-05-07 cmucl-imp
I'll also try to port this to SBCL. If I can get it to work I'll send
you patches in a few days.

-- 
Martin Atzmueller <ma...@at...>