Grobner



Raymond Toy (RT/EUS) wrote:

> Michel Talon wrote:
>> Raymond Toy (RT/EUS) wrote:
>>> I don't see any issues with current CVS with cmucl, but I haven't
>>> finished running the entire demo yet.
>> 
> 
> Was the failure early in the demo or late?

Quite early. 

> 
> If you want to try cmucl, you can grab a binary from common-lisp.net.
> The latest snapshot is supposed to work on FreeBSD7, I think.  (I don't
> have FreeBSD.)  Installation is simple.  Untar the files anywhere.

Indeed i have downloaded cmucl-19d-x86-FreeBSD5.tar.bz2 and it works on
FreeBSD-7, at least my machine which has been upgraded from 5 and has the
old libraries. Anyways, i have recompiled maxima 5.14.0 with this version
of cmucl, and here package grobner works OK. So the problem is definitely
a bug related to sbcl.

The demo runs for around 7 mn on my machine
Note i still get just one an error with cmucl:
Maxima 5.14.0 http://maxima.sourceforge.net
Using Lisp CMU Common Lisp 19d Release (19D)
.....

                               3    4    5         2    3    3
(%i58) ev(poly_grobner([- 1 + z  + y  + x , - 1 + z  + y  + x ], [x, y, z]), 
                                                                   modulus =
3)
Maxima encountered a Lisp error:

 Type-error in KERNEL::OBJECT-NOT-LIST-ERROR-HANDLER:  T is not of type LIST

Other tests succeed.


>
> 
> I was mistaken.  It hasn't finished.  It takes a while to run....

I have recompiled the latest version of sbcl and maxima on FreeBSD-7 and
same problem. In fact even 
load("grobner")
produces errors, and almost all the programs in the demo fail.

niobe% maxima
Maxima 5.14.0 http://maxima.sourceforge.net
Using Lisp SBCL 1.0.11
...
(%i1) load("grobner");
....
; compilation unit finished
;   caught 7 ERROR conditions
;   caught 6 STYLE-WARNING conditions
;   printed 280 notes
Evaluation took 5.09 seconds (5.34 elapsed) using 137.764 MB.
(%o1)  /usr/local/share/maxima/5.14.0/share/contrib/Grobner/grobner.lisp
.....
.....


(%i39)            poly_grobner_member(y + x, [x, y], [x, y])
Maxima encountered a Lisp error:

 Error during processing of --eval option "(cl-user::run)":

  Execution of a form compiled with errors.
Form:
  (MAKE-MONOM NVARS INITIAL-ELEMENT 0)
Compile-time error:
  (in macroexpansion of (MAKE-MONOM NVARS :INITIAL-ELEMENT ...))
(hint: For more precise location, try *BREAK-ON-SIGNALS*.)
The value NVARS is not of type FIXNUM.

Automatically continuing.
To reenable the Lisp debugger set *debugger-hook* to nil.


I have taken just an example but similar errors occur before.
Basically everything fails.

> 
> Ray

-- 
Michel Talon