Xref: news2.mv.net alt.test:169406 comp.os.msdos.djgpp:575
From: pausch AT electra DOT saaf DOT se (Paul Schlyter)
Newsgroups: comp.os.msdos.djgpp,alt.test
Subject: Re: long double functions
Date: 23 Jan 1996 18:29:06 +0100
Organization: Svensk Amat|rAstronomisk F|rening
Lines: 31
Message-ID: <4e35t2$qas@electra.saaf.se>
References: <4dmqgv$3jf AT urca DOT nce DOT ufrj DOT br> <30FF05D1 DOT 9381AFB AT alcyone DOT com>
NNTP-Posting-Host: electra.saaf.se
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

In article <30FF05D1 DOT 9381AFB AT alcyone DOT com>,
Erik Max Francis  <max AT alcyone DOT com> wrote:
 
> Antonio Carlos Moreirao de Queiroz wrote:
> 
>> Are the math functions (sqrtl, sinl, cosl, etc.) for long doubles
>> available? I am not finding their definition in math.h or in other
>> files of the djgpp distribution.
>> I see that the long double type is apparently supported. But
>> without the functions...
> 
> Only the double-taking math functions are standard (with a few exceptions
> here and there).  Functions based on float or long double are not specified
> by the ANSI C standard.
 
They ARE specified by the numerical extension to the standard though,
and will very likely make it into the next version of Standard C,
which probably will appear in 4-5 years or so.
 
And many compilers support these functions already.
 
Another interesting question: djgpp may support "long double", but does
it have any higher precision than plain "double"?  The C standard allows
support of "long double" by keeping it to the same precision as plain
"double", but such "support" is of course not really very useful.
 
-- 
----------------------------------------------------------------
Paul Schlyter,  Swedish Amateur Astronomer's Society (SAAF)
Grev Turegatan 40,  S-114 38 Stockholm,  SWEDEN
e-mail:  pausch AT saaf DOT se        psr AT home DOT ausys DOT se