I've just checked in two patches to the integrator (files driver.red and
isolve.red) that are supposed to solve some problems in conjunction with algint.
algint uses the fluid intvar to store the integration variable, but the standard
integrator didn't always update this when it performs a substitution before
(I would appreciate if someone would look over these patches.)
In particular, this improves the result for the integral of
as reported in
Note: run it with "on ezgcd", otherwise it takes at least 10 times as long to compute.
You still get the "Zero divisor" error messages, though.
Another example that went very wrong is
kk := 1/(sqrt((a-x)*(b-x)*(x-1)*x));
ll:=kk/(x*(a*b*x - a*b - a*x**2 + a*x - b*x**2 + b*x + x**3 - x**2));
This one gives the correct result only when precise is off, and the final result
with precise on suggests to me that there is a problem with a sign of a sqrt