Normally gcd() returns a positive integer, and gcd(a,0) = gcd(a,a) = abs(a). But since abs(INT_MIN) cannot be represented as a 32-bit integer, both those cases should throw an integer-out-of-range error.
I'm also in favor of that option, rather than sending a negative result as a result.
About lcm(a, b): a / gcd(a, b) * b, at least if a & b are positive. If not, some thoughts are needed:-)
Returning a NUMERIC as suggested by Tom would solve the overflow problem by sending it back to the user who has to cast. This looks ok to me.
Maybe we could provide "int4 lcm(int2, int2)", "int8 lcm(int4, int4)", as ISTM that there cannot be overflows on those (eg for the later: lcm <= a*b, a & b are 31 non-signed bits, 62 bits are needed, 63 are available).
-- Fabien.
