Raymond Toy (RT/EUS) wrote:
> Perhaps the solution is to tell maxima to try the indefinite integral
> first, or at least once maxima has determined that there are no bad
> points within the limits of integration.
I made the simple change to try the antiderivative first. It seems to
work and the definite integral is computed in a second or two. It takes
this long because maxima is checking to see if the integrand has any
poles in the integration interval. You can disable that by setting
intanalysis:false, but you should know what you're doing.
Running the testsuite shows 2 new failures, but that failure is due to
the form of the answer, not the value.
Ray