#2231 Integral with float exponent goes wrong

[Stavros Macrakis]

integrate(x^(1/2)*%e^-x,x,0,inf) => sqrt(%pi)/2 OK

integrate(x^.5*%e^-x,x,0,inf) => 8545/9642 !!! This is rat(float(sqrt(%pi)/2)) -- why?!

There is some argument that the second result should be a float (i.e. approximate number) to reflect that the input included an approximate number, but I see no reason that it should be a rat.

Maxima 5.23.2 GCL 2.6.8 Windows 7