This is not a hard integral at a first glance: it's not algebraic,   
   only one Exp function composed with rational function.   
      
   So, the simple case in Risch algorithm.   
      
   As the result shows, it doesn't have an elementary answer, but   
   an answer with Ei.  So this requires an extension of Risch algorithm,   
   and this is how FriCAS solved it.   
      
   How did MMA 2.2 solve it?  I can only guess.  It probably uses   
   heuristics, by breaking (a+b)/c into a/c and b/c, then try to   
   integrate the parts.  However, there are infinity ways to break   
   (a+b) into (a+d) and (b-d).   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)