[Coco] Converting ANSI-C to K&R(Microware) - Ansifront didn't work

[gene heskett]

On Sunday, January 09, 2011 07:47:23 pm Willard Goosey did opine:

> On Sat, Jan 08, 2011 at 09:06:50PM -0500, gene heskett wrote:
> > Ansifront012, AIUI it, should not ever have its mind blown as it is
> > essentially a void to int translator, just treating the serial text
> > stream that goes through it, making the appropriate substitutions.
> 
> Actually, Ansifront does a lot of name hashing.  It hashes any name
> that's more than 8 characters long, and it also hashes all struct and
> union element names, enforces "unsigned char", and I don't even know
> what all else.
> 
> > If you are going to be doing transcendental math functions in C, I
> > have the trig.l stuff in that same directory.  The .l stuff goes in
> > /dd/LIB so the compiler and cc can find it.  This does trig functions
> > to about 16.5 digits accuracy over an EE range of +-EE38 or so. There
> > is a src package and a read.me.re.trig.l text file too
> 
> How does this compare to Krieder's clibt?
> 
> Willard

This has all its transcendentals in full double precision.  I believe 
Kreiders is single where there are dups, but the proto's won't match 
anyway.  Carls was a heck of an improvement over the regular c.lib, 
particularly in speed & bug freedom.  I don't think my eclipse proggy used 
any of Carls code though as I wasn't fixated on speed, but in accuracy, at 
least back to where the julian calendar dies in -4713BC.  It doesn't 
underflow "gracefully" in any math system I've tried.  But I'll make zero 
claims about my math abilities. I once had a calculator that said 2+2 was 
3.9999999.  It didn't last long.

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
<Endy> taniwha: Have you TESTED this one? :)
<taniwha> Endy: of course not