specint bug in 5.9.3



>>>>> "Richard" == Richard Fateman <fateman at cs.berkeley.edu> writes:

    Richard> 1.  (*) text/plain          ( ) text/html           
    Richard> %i1) specint(sqrt(t)*%e^(-p*t-a/t),t);

    Richard> Is  p  positive, negative, or zero?
    Richard> pos;
    Richard> Is  a  positive, negative, or zero?
    Richard> pos;
    Richard> Is  psey  positive, negative, or zero?

    Richard> What is "psey"???

A known issue for which I have no solution.  At this point, specint
has figured out the expression is of the form u*exp(a*x+e*f+c), and
has determined that a is negative.  In this particular case, a is -p.
It substitutes psey for -a and tries the transform again.

I think I tried a few things a while back, but never got it to work
right.

Ray