RE: [Gcl-devel] Re: [Axiom-developer] Minor bugs

[Page, Bill]

Camm, Tim,

I don't think these rounding and truncation problems are
a lisp issue as such. They are a result of an assumption
in Axiom about the behavour of the shift operation on
signed integers which is not compatible with the current
lisp standard. Basically Axiom expects an unsigned shift
function but the standard lisp function call is extending
the sign arithmetically instead of "logically". Details are
on Savannah bug and patch lists. There is some discussion
of the efficiency of the proposed patch in the axiom-
developer email list archive of a little more than a year
ago.

The patch I submitted (but haven't yet found time to move
to the axiom--main--1 archive, provides the logical shift
by check the sign and changing things accordingly. This is
check is pretty fast, but it occurs at such a deep level
that even a little inefficiency could affect computationally
intensive programs. In the medium/long term it would be more
efficient to directly map many of the numerical functions in
float.spad (and a few other places) to their high level
equivalents in lisp.

Regards,
Bill Page.

On Wednesday, December 15, 2004 4:21 PM Camm Maguire wrote:
> 
> Not sure what could be going on here.  The following results
> with 2.6.5 are unchanged on i386 and amd64 (Debian):
> 
> >(round -3.72)
> 
> -4
> 0.28000000000000025
> 
> >(truncate 9.2)
> 
> 9
> 0.19999999999999929
> 
> >(truncate -9.6)
> 
> -9
> -0.59999999999999964
> 
> >(by)
> 
> Perhaps the original poster could try this in their GCL lisp image.
> 
> Take care,
> 
> root <address@hidden> writes:
> 
> > Kathy,
> > 
> > Yes, this is the right list.
> > Thanks for the bug report. 
> > 
> > If I recall correctly this was a problem a long time ago
> > and resolved into an issue in lisp. Perhaps something got
> > broken in the GCL build?
> > 
> > Tim
> > 
> > ================================================================
> > 
> > Hi everyone -
> > 
> > I have been subscribed to this list for quite some time.  
> Usually, I am 
> > too busy or overloaded with other activities and pass on by.
> > 
> > However, I thought that the least I could do was to try to 
> do some basic 
> > testing ahead of your news announcement.
> > 
> > So already I see  problems in the Windows release.  Here 
> are the results:
> > 
> >                         AXIOM Computer Algebra System
> >              Version of Wednesday December 15, 2004 at 00:50:38
> > 
> --------------------------------------------------------------
> ---------------
> >    Issue )copyright to view copyright notices.
> >    Issue )summary for a summary of useful system commands.
> >    Issue )quit to leave AXIOM and return to shell.
> > 
> --------------------------------------------------------------
> ---------------
> > 
> >    Re-reading compress.daase   Re-reading interp.daase
> >    Re-reading operation.daase
> >    Re-reading category.daase
> >    Re-reading browse.daase
> > (1) -> round(-3.72)
> >    Loading C:/Program Files/axiom/mnt/windows/algebra/FLOAT.o for
> >       domain Float
> > 
> >    (1)  - 5.0
> >                                                             
>       Type: 
> > Float
> > (2) -> truncate(9.2)
> > 
> >    (2)  9.0
> >                                                             
>       Type: 
> > Float
> > (3) -> truncate(-9.6)
> > 
> >    (3)  - 10.0
> >                                                             
>       Type: 
> > Float
> > (4) -> fractionPart(-3.7)
> >    Loading C:/Program 
> Files/axiom/mnt/windows/algebra/FPS-.o for domain
> >       FloatingPointSystem&
> >    Loading C:/Program 
> Files/axiom/mnt/windows/algebra/RNS-.o for domain
> >       RealNumberSystem&
> > 
> >    Loading C:/Program Files/axiom/mnt/windows/algebra/UNISEG.o for
> >       domain UniversalSegment
> >    (4)  0.3
> >                                                             
>       Type: 
> > Float
> > (5) ->
> > 
> > 
> > 
> > I am running 32 bit XPSP2 on AMD 64 bit.
> > 
> > Please let me know if this is the incorrect list.
> > 
> > Thanks for all your hard work.
> > 
> > Kathy Gerber
> > 
> > 
> > _______________________________________________
> > Gcl-devel mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/gcl-devel
> > 
> > 
> > 
> 
> -- 
> Camm Maguire                                          
> address@hidden
> ==============================================================
> ============
> "The earth is but one country, and mankind its citizens."  -- 
>  Baha'u'llah
> 
> 
> _______________________________________________
> Axiom-developer mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/axiom-developer
>