Integration bug. A lot of them!!!



As I reported maxima gives wrong answer on integrands like
x^n/sqrt(1-x^2)/(1+a^2*x^2)^m with even powers n,m.
Integrands like asin(x)*x/(1+a^2*x^2)^2 are in close connection with them.
Maxima fails on it too. You can substitute acos in place of  asin

(C7) factor(defint(2*x*asin(x)/(1+x^2)^2,x,-1,1));
       %PI
(D7) - ---                 (lose contibution sqrt(2)/2)
        2
(C8) factor(defint(2*x*acos(x)/(1+x^2)^2,x,-1,1));
     %PI
(D8) ---                   (lose contibution -sqrt(2)/2)
      2
(C9) factor(defint(2*x*atan(x)/(1+x^2)^2,x,-1,1));
      1
(D9) -
      2
The last one is correct. Probably due different procedure to integrate (by parts)
However 
(C10) factor(defint(2*x*acot(x)/(1+x^2)^2,x,-1,1));
        1
(D10) -      (lose contibution -%pi/2)
        2

rgds,
v