What should the default GCD algorithm be in Maxima?
Right now, it is subres, but it appears that spmod works in more cases.
See bug reports 696818 (Bad: subres, red, algebraic), 711871 (Bad: ez,
mod); 716508 (Bad: subres, red). So that seems to exclude subres, red,
ez, mod, and algebraic, and leaves spmod and mod; the others on the
*gcdl* list (sphen, eez) don't actually exist. I don't know of any bugs
in spmod or mod, but presumably they have not been exercised as heavily
as subres.
Wolfgang Jenkner posted a patch to subres for the algebraic case, but if
I'm not mistaken, spmod still works better in general.
Is there any reason the default shouldn't be changed to spmod? Of
course, if bugs show up with it under intensive use, we'll have to
reconsider. Wolfgang, Richard, do you have any sort of database of hard
cases for gcd that we can use for tests?
-s