ccreweb.org fsl modules update
  Home FAQ Contact Sign in
comp.lang.forth only
 
Advanced search
POPULAR GROUPS

more...

comp.lang.forth Profile…
 Up
ccreweb.org fsl modules update         


Author: Krishna Myneni
Date: Sep 19, 2008 06:06

An update on the status of our experimental FSL modules may be found at

ftp://ccreweb.org/software/fsl/README

The set of modules has been tested now under Gforth and VFXforth for linux
(Alpha 8). The causes of various test failures have not yet been identified, but
I would greatly appreciate any feedback on the problems seen thus far.

Krishna
3 Comments
Re: ccreweb.org fsl modules update         


Author: David N. Williams
Date: Sep 21, 2008 16:20

Krishna Myneni wrote:
> An update on the status of our experimental FSL modules may be found at
>
> ftp://ccreweb.org/software/fsl/README
>
> The set of modules has been tested now under Gforth and VFXforth for
> linux (Alpha 8). The causes of various test failures have not yet been
> identified, but I would greatly appreciate any feedback on the problems
> seen thus far.

I just want to say, thanks for doing this! I'm a fan of
portable libraries and portable tests. :-)

-- David
no comments
Re: ccreweb.org fsl modules update         


Author: Krishna Myneni
Date: Sep 22, 2008 05:30

Charles Davis wrote:
> Krishna Myneni wrote:
>> Krishna Myneni wrote:
>>
>>> An update on the status of our experimental FSL modules may be found at
>>>
>>> ftp://ccreweb.org/software/fsl/README
>>>
>>> The set of modules has been tested now under Gforth and VFXforth for
>>> linux (Alpha 8). The causes of various test failures have not yet
>>> been identified, but I would greatly appreciate any feedback on the
>>> problems seen thus far.
>>>
>>> Krishna
>>
>>
>> VFXforth for linux fails the lufact.fs tests for the following reason:
>>
>>
>> VFX Forth for Linux IA32 ...
Show full article (1.37Kb)
no comments
Re: ccreweb.org fsl modules update         


Author: Thomas Pornin
Date: Sep 22, 2008 05:52

According to Charles Davis gamewood.net>:
> I seem to remember, the 80387 chip is the 'Floating Point' chip that
> goes with the 80386 family CPU? If 387 support is being loaded, then
> 387 operations will fail if that chip isn't physically there.

Not with Linux. Using FPU opcodes when the FPU is not physically present
triggers exceptions, but the OS traps them and emulates them.
Computations will be slow, but should not fail.

Besides, it takes some effort to find an i386-compatible platform which
runs Linux but does not feature a FPU (either you dig out a 15-year old
486SX PC, or you go for an embedded system with a low-end Intel clone
such as the Vortex86SX).

--Thomas Pornin
no comments