[M3devel] Open CM3 regression tests

[Olaf Wagner]

Quoting Henning Thielemann <lemming at henning-thielemann.de>:
> On Sun, 20 Jan 2008, Olaf Wagner wrote:
>
>> > no errors for p1 to p115
>>
>> --- p116b --- default IEEE floating point tests from Xerox PARC
>> --- p1/p116b/stderr.pgm	2008-01-20 00:47:20.000000000 +0100
>> +++ ../src/p1/p116b/stderr.pgm	2008-01-13 00:55:55.000000000 +0100
>
> My compiler complains about LONGINT type in src/Test.i3 . I assume that
> this is the 64 bit support. Seems that I have to upgrade my compiler.
>
> In general floating point is a strange issue. You cannot rely on that your
> machine supports IEEE format, although most machines do. The precision of
> the same float type can be different on different machines.  However, the
> Float modules shall reflect that. Computation with NaN and Infinity is
> broken, it should be avoided and in most cases continueing the computation
> with NaN or Infinity only hides the precise error location. It would be
> better to abort with an error for a division by zero, overflow and so on.
> I think one can enable this with FloatModes.

I don't really understand the consequences of what you say above:
are the tests wrong and should be adapted, or is the implementation
broken for IEEE-default (where most of the FloatMode procedures simply
raise exceptions; if so, can we fix that for these platforms), or do
we simply rely on broken hardware (x86 floating point processors)?

It seems that FloatMode is only implemented for very few targets:
DS3100 DS3100_OSF IRIX5 SOLsun SPARC SUN386 VAX. All the others use
the default. Will we simply have to implement the missing functionality
for all platforms? If so, can e.g. SUN386 be easily adapted to other
x86 platforms (FreeBSD4, LINUXLIBC6, I386_DARWIN, NetBSD2_i386)?
That would be a start.

As I said, I'm no expert in this domain, and would rather rely on others
knowledge there.

If you don't have a compiler that is up-to-date, you can always use
workspaces on birch.elegosoft.com for tests. The regression tests
on this machine use CM3 installations at

   m3 at birch:~/cm3$ ls /home/m3/work/cm3-inst/birch.elegosoft.com/
   current  last-ok  prev-ok  rel-5.4.0.

If you don't have login access but need it, just let me know. Of course
this applies to everybody wanting to work on the tests or other
improvements (as long as there aren't hundreds of requests at least :-).

Olaf
-- 
Olaf Wagner -- elego Software Solutions GmbH
                Gustav-Meyer-Allee 25 / Gebäude 12, 13355 Berlin, Germany
phone: +49 30 23 45 86 96  mobile: +49 177 2345 869  fax: +49 30 23 45 86 95
    http://www.elegosoft.com | Geschäftsführer: Olaf Wagner | Sitz: Berlin
Handelregister: Amtsgericht Charlottenburg HRB 77719 | USt-IdNr: DE163214194