I have found a difference in the handling of an integral between Maple versions 10 and 12. It bothers me since it was working in version 10, but now it gives a numeric exception... I'm not sure whether the replacement is valid or not, so I hope someone can validate it for me.
Suppose I have the function definition
> f := (x,a,b) -> int( t^(a-1) * (1-t)^(b-1), t=0..x );
As an example, consider
This gives -0.8 both in Maple version 10 and 12.
However, if I replace the previous statement with
> subs( x=2., a=3., b=4., % );
> evalf( % );
Then I still get -0.8 in Maple version 10, but in version 12 this results in "Error, (in GAMMA), numeric exception: division by zero"!
The problem seems to be in the first line, f(x,a,b), which in Maple version 10 gives back the same/unchanged integral from the function definition. In Maple version 12 however, this integral is replaced with an expression involving GAMMA and hypergeom. (The error is generated since there is a factor of the form GAMMA(-b+1) which gives a pole for positive values of b.)
I am wondering whether replacing this integral is valid, or whether this is a wrong replacement or even bug in Maple 12 (the same problem seems to appear in Maple version 11 as well).
Many thanks for any reply.