The bug also goes away if you set gcd to EZ,
and the bug STAYS when you set gcd to RED.
The reduced PRS GCD "RED" is about the simplest polynomial GCD that
one could reasonably use. The bug is not in the GCD program but
in the surroundings. Both EZ and SPMOD set "modulus" to a prime.
I don't know what else differs.
The definite integration program is really a hack. it uses numbers
like 10000000 to approximate infinity, I guess. It also uses
floating point numbers.
You can trace the polynomial gcd program by ?trace(?pgcd); and see that
the internal gcd program is being called on different inputs in the two
separate runs.
Barton Willis wrote:
> Here are two reported bugs that go away
> when gcd : 'spmod.
>
> bug 1020711
> http://sourceforge.net/tracker/index.php?func=detail&aid=1020711&group_id=4933&atid=104933
>
>
> bug 831445
> http://sourceforge.net/tracker/index.php?func=detail&aid=831445&group_id=4933&atid=104933
>
>
> The code for subres might be spotless, but the global
> environment might be messed up; or it might be that
> some code that subres calls is buggy; or ...
>
> Bug 1020711 is
>
> (%i1) integrate(sqrt(4*x^2 +4),x,0,1);
> QUOTIENT by ZERO
> -- an error. Quitting. To debug this try DEBUGMODE(TRUE);
> (%i2) integrate(sqrt(4*x^2 +4),x,0,1),gcd : 'spmod;
> (%o2) asin(1)+sqrt(2)
>
> Barton
>
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima