RE: [cc65] Floating point support (was: C++ compiler) - IEEE 32bit compliance

Date view Thread view Subject view

From: Greg Long (cc65_at_maneuveringspeed.com)
Date: 2003-03-07 00:37:04


I would suggest this compliance is an important goal, even at the cost
of additional memory consumption.  I say this because you guys seem to
have worked hard for as much standard compliance as possible, seems that
using a different standard for different machines would cause more
problems than saving  the 1k, 2k, or even 4k or more would be worth.

On the c64 specifically, using BASIC ROM routines would be problematic,
since the interpreter would have to be switched back on, or the routines
would have to be copied to another location, which negates any memory
saving anyways.

Using the same code from platform to platform would be easier to develop
and debug. (surely the routines for IEEE float and double support have
been written, wouldn't they "merely" (ha, ha) need to be translated into
6502 code?

Greg

-----Original Message-----
From: owner-cc65_at_musoftware.de [mailto:owner-cc65_at_musoftware.de] On
Behalf Of Groepaz
Sent: Thursday, March 06, 2003 3:21 PM
To: cc65_at_musoftware.de
Subject: Re: [cc65] Floating point support (was: C++ compiler)


On Thursday 06 March 2003 23:49, Ullrich von Bassewitz wrote:
> On Thu, Mar 06, 2003 at 11:43:49PM +0100, Groepaz wrote:
> > btw...(you know that i want floats aswell =P)....would steve woz's 
> > float routines be sufficient? would they be "licence compatible" ?
>
> I've never seen them, so I cannot tell. Did he use the IEEE 32 bit 
> floating point format?

i am talking about the apple2 rom routines :o) the sources are out on
the web 
(i couldnt find them with a quick check at google though mmmh...maybe
someone 
of the apple2 guys has an url handy :=)) one nice thing with this code
is 
that is nicely documented (from woz himself) and that it obviously
proved to 
be working :)

about the format i dont know, i dont remember seeing any format
mentioned :=P

> Just the list of math functions in the C99 standard is more than four 
> pages long (function prototypes, line after line). I'm pretty sure 
> that most of these functions are missing in the floating point package

> you're talking
> about:-)

uh yes ofcoz :=) i am not talking about math.h or whatever routines,
just the 
few things required to get basic support running.

gpz



----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo_at_musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.

----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo_at_musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.


Date view Thread view Subject view

This archive was generated by hypermail 2.1.3 : 2003-03-07 00:37:49 CET