NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/alt/lucidem/help/431

Internet Message Format | 1992-09-14 | 1.6 KB

Path: sparky!uunet!mcsun!sunic!corax.udac.uu.se!woody.csd.uu.se!hb
From: [e-mail] (Henrik B}kman  CSD)
Newsgroups: alt.lucid-emacs.help
Subject: floating-point functions
Date: 15 Sep 1992 09:19:50 GMT
Organization: Uppsala University
Lines: 25
Distribution: world
Message-ID: <[e-mail]>
NNTP-Posting-Host: woody.csd.uu.se


I'am looking on some floating-point functions, especially the
missing ones in hpux, and I have some questions.

Some of the floating-point functions isn't defined for all values
and returns the special IEEE-754 value 'Nan' and on overflow they
return '+-(infinit)'. There is two C-functions, isnan() and isinf()
which you can check for this special values but I can't find any
such functions in Lucid-emacs. I wonder if there is some functions
like this in Lucid-emacs or if I haven't looked enough or if Lucid-
emacs is handling such math-errors in another way or if Lucid-emacs
doesn't handling such errors at all?

If I want to implement the missing functions what should I go for.
Fast and eaisy but with some round-errors or should I make them
return more correct values? That is, can I use already defined functions
or should I use "pure" numerical methods like Newton-Raphson and such?
(I don't want to implement them in PA-RISC assembler :-) )

--
----------------------------------------------------------------------
Henrik Bakman           Adress: Box 520          Tel: [phone]
System Manager                  751 20 UPPSALA   Fax: [phone]
Computing Science Dep.          SWEDEN           Email: [e-mail]
Uppsala University