-----BEGIN PGP SIGNED MESSAGE-----
sommerfeld@orchard.medford.MA.US (Bill Sommerfeld) writes:
> ,open floatnums
> I don't know how buggy that code is...
Well, it passes my (simplistic) sanity checking, but as I
noted to you in email things like (/ (log 1000) (log 10))
will give correct results, but also a warning:
Scsh 0.4
> ,open floatnums
Load structure floatnums (y/n)? y
[floatnums
/usr/local/lib/scsh/rts/floatnum.scm
Analyzing...
Calls will be compiled in line: (floatnum-datum make-float-datum exp log sin
cos tan asin acos atan sqrt)
.......................................................
]
> (/ (log 1000) (log 10))
Warning: invalid arguments
(/ (log 1000) (log 10))
(procedure wants: (&rest :number))
(arguments are: (:extended-number :extended-number))
3.
However, I also noted that I thought that it might have to
do with more-packages.scm's structure definition for
usual-features declaration, and in particular:
;; Choose either innums, floatnums, or neither
innums ;Silly inexact numbers
;; floatnums ;Still don't print correctly
Commenting out innums, uncommenting floatnums, and
rebuilding now does:
Scsh 0.4
> (/ (log 10000) (log 10))
4.
which, except for the trailling dot, is pretty much
perfect.
So far, my simplistic testing has been limited to the
nifty crypt.scm code from http://www.c2.org/~um/crypt/scheme.html
tested against applications in C, but since that's more or
less what I wanted to be running anyway, I'm very happy.
Thanks, Bill.
Sean.
-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
Comment: PGP Public Key in ftp://ftp.sprintlink.net/engineer/smd/pgpkey
iQCVAwUBMgj0aUSWYarrFs6xAQF8bwP+KVonjpDKbJ9rKODGs3WLd1a5cp49x/jc
Wdk/TPSGEQAX7DnUxdXOcm76U5MIcghq8YuWDPMlliMjzou80uIWQJSSJ0byFYtH
0gq3ViUZi5o6O1DR/QenTuv9aitwY4J2XAuaO+9ZW3uFSijL4SZUJyuUvjgXVuAm
GxFNx7ytMcQ=
=qs7n
-----END PGP SIGNATURE-----
|