Re: [Maxima-discuss] ratsimp not working as expected

[andre m. <and...@gm...>]

On 03/07/2014 07:20 PM, andre maute wrote:
> On 03/07/2014 05:57 PM, andre maute wrote:
>> Conclusion at the moment:
>> No working Maxima 5.31.3 with CLISP, GCL(extra segfault), SBCL
>> Working Maxima 5.28.0 with SBCL
> I have now compiled (advice from Thomas D. Dean) Maxima 5.32.1
> again with CLISP, GCL, SBCL the behavior is the same as with Maxima 5.31.3
> So this is somehow contradictory to the report from Thomas.
>
> @Thomas: what sort of linux distribution do you have?
>
> Conclusion at the moment:
> No working Maxima 5.31.3 with CLISP, GCL(extra segfault), SBCL
> No working Maxima 5.32.1 with CLISP, GCL(extra segfault), SBCL
> Working Maxima 5.28.0 with SBCL
Robert Dodier suggested to use load("invert")
which uses the adjoint version of matrix inversion.

Now 'Quotient by a polynomial of higher degree' doesn't appear anymore
for SBCL, CLISP and ratsimp-bug1.max

The segfault  doesn't appear anymore.
for GCL and ratsimp-bug3.max

BUT: for GCL and ratsimp-bug1-adjoint.max gives nevertheless a segfault 
for the last ratsimp

Regards Andre

--- ratsimp-bug1-adjoint.max -----------
load("invert");
display2d : false;
d : 4;
A : genmatrix(lambda([k,l],v[[k,l]]-v[[0,l]]),d,d);
detA : factor(determinant(A));
AinvT : transpose(invert(A));
/* AinvT : ratsimp(AinvT); /* error trigger */ */
zz : [0,-1,0,0];
h1 : transpose(matrix(zz));
h2 : AinvT.h1;
h3 : sum(h2[m,1]^2,m,1,d)*detA^2;
res : ratsimp(h3); /* chokes if trigger is commented */
--- ratsimp-bug1-adjoint.max -----------