rootsof problem -- design doesn't have enough info!

Raymond Toy wrote:
> Richard Fateman wrote:
>   
>> Who added integrate_use_rootsof  etc?  The essential idea is OK in 
>> there, but is problematical.
>>   
>>     
> I think Bill Schelter added that long ago.
>
> Ray
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>   
I have a fix for this, which is in one function in sinint.  To test it 
try, before and after loading it,
integrate_use_rootsof:true$
integrate(1/(a*x^3+x+1),x);

maybe someone with access to source can insert it?
thnx
RJF


(defun integrate-use-rootsof (f q variable &aux qprime ff qq
                  (dummy (make-param)) lead)
  ;; this program is used if $integrate_use_rootsof is true,
  ;; and rational function integration of f/q wrt variable
  ;; can make use of it.
  ;;
  ;; fixed by RJF 5/14/09 to work on non-monic denoms and also
  ;; add 2nd arg to rootsof, so it makes sense for multivariate polys:
  ;; the second arg of rootsof is the main variable in the poly wrt 
defining roots.
 
  (setq lead (pdisrep(p-lc q)))  ;; inserted pdisprep; rjf
  (setq qprime (disrep (pderivative q (p-var q))))
  (setq ff (disrep f)
        qq (subst dummy variable (disrep q)));put correct variable in poly
 
    `(($lsum) ((mtimes) ;removed factor of lead, which was just wrong. rjf
         ,(div*  (subst dummy variable ff)
            (subst dummy variable qprime))
         ((%log) ,(sub* variable  dummy)))  ,dummy
    (($rootsof) ,qq ,dummy))) ; use 2nd arg to rootsof to indicate main var.