axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Axiom-developer] source access


From: Juergen Weiss
Subject: RE: [Axiom-developer] source access
Date: Tue, 22 Jul 2003 11:17:18 +0200

Hi,


> -----Original Message-----
> From: Camm Maguire [mailto:address@hidden 
> Sent: Tuesday, July 22, 2003 2:17 AM
> To: Juergen Weiss; address@hidden
> Subject: Re: [Axiom-developer] source access
> 
> 
> Greetings!
> 
> "Juergen Weiss" <address@hidden> writes:
> 
> > Unfortunately it seems, that there are two problems. The
> > first one is insufficient value stack size for Axiom
> > (the rs6000 port of gcl had a higher limit for quite a while).
> 
> I can't find any settings for the r6000.  Might you know where to
> look? 

rios.h or rios-aix3.h

> > I think sometimes the function |Join| is called with
> > quite a few arguments. Setting the value to the value
> > used for rs6000 should fix that problem. For the second 
> 
> Don't see what this has to do with the stack.  But there are limits to
> 64 arguments in places inside GCL.  Overflows trigger reported
> errors.  Do you feel this could be an issue?

Probably not, at least the problems we see now is caused by
infinite recursion, not by stacks which are too small. I think
there was a problem with the number of arguments to a function,
at least in compiled code. There was a comment in the axiom
code at some place. But sorry, I do not remember the details.
 
> > one I posted an example on 06/10. There is an infinite
> > recursion. Reason is unclear. Could be different sharing
> > of cells in cmu and gcl in destructive calls, that is a
> > coding error in Axiom. Change of semantics of a lisp
> > function. Or the gcl compiler has generated some bad code
> > (I already found a compiler error in cmu cl at another 
> > place).
> 
> So far, I know that compForMode on first input (|CommutativeRing|)
> gives a vmode of (|Ring|), and compMakeCategoryObject on (|Ring|)
> returns a 'catlist' containing (|has| R (CommutativeRing|)).  Might
> you know if either of these might be in error?  (One of them certainly
> should be at least, I think.) (See knownInfo in info.clisp).

I have to look at this.


Juergen Weiss

Juergen Weiss     | Universitaet Mainz, Zentrum fuer Datenverarbeitung,
address@hidden| 55099 Mainz, Tel: +49(6131)39-26361, FAX:
+49(6131)39-26407




reply via email to

[Prev in Thread] Current Thread [Next in Thread]