Hom between modules over different rings

[mahrud]

Currently the implementation of Hom(M, N) convert both modules to the same ring in an ambiguous way, causing incorrect results:

i1 : K = ZZ/11;

i2 : S = K[x,y];

i4 : Hom(S^1, K^1)

      1
o4 = S

o4 : S-module, free

i5 : Hom(S^1, coker vars S)

o5 = cokernel | y x |

                            1
o5 : S-module, quotient of S

[mahrud]

Maybe Hom should give an error if the rings aren't the same?

[Devlin-Mallory]

I didn't know you could even run Hom(M,N) if M,N weren't modules over the same ring! I think that having it throw an error in that case makes sense - Hom(S^1, K^1) is mathematically meaningless, so I'd rather have it not return anything.

(As far as I can tell, adding this error doesn't break any tests...)