Open mgkurtz opened 1 year ago
mathematically, the gcd is well defined in ZZx as it is a UFD, but the gcdx is not possible in ZZx for this example, we have <x, x+2> cap ZZ = <2> the gcd computation in Nemo is done in C (flint). Due to Gauss, this is "correct" This cannot be detected easily as we don't seem to have a euclidean function as part of the ring interface
Accidentally somewhere in my code, I called
gcdx
for polynomials inZZ[:x]
instead ofQQ[:x]
. This led to my tests not terminating 😞 Here some example code:Is this non-terminating behavior intended?
In AbstractAlgebra instead, the call to
divrem(x, 2)
ingcdx
fails, while in Nemo it gives(0, x)
.Confusingly, both AbstractAlgebra and Nemo compute
gcd(x, x+2)
as1
.