HELP needed: Taylor bugs



I just submitted two fixes for a trivial and an easy taylor bug I found. 
Unfortunately, the ONE taylor bug I tried to fix I couldn't, but I have
a clue now:

taylor(sum(M^K*sum(L^K,L,1,D),K,1,INF),M,0,2),simpsum;

gives a nearly correct answer, only the L's should be replaced by D's...


It appears that the error occurs when fpolysum in combin.lisp is called.
fpolysum in turn is

(defun fpolysum (e)     ;returns *ans*
   (let ((a (fpoly1 (setq e ($expand ($ratdisrep ($rat e *var*))))))

and within the ratdisrep rat stuff the following happens:

usually a rat expression is

 ((MRAT SIMP <varlist> <genvar> <tlist> trunc) <poly-form>)

where varlist and genvar have the same length.

Here they don't, quite on the contrary, genvar contains quite some rubbish 
I do not REALLY understand where it could come from.

My question is:

should varlist and genvar really always have the same length or is this 
not required?

where should I look to now?

Martin