-----maxima-bounces at math.utexas.edu wrote: -----
>============= 8< ===============
>(%i1) display2d:false;
>
>(%o1) false
>(%i2) sqrt((f - z^2/4/f)^2 + (d-z)^2) + sqrt((f - z^2/4/f)^2 + (zp-z)^2);
>
>(%o2) sqrt((zp-z)^2+(f-z^2/(4*f))^2)+sqrt((f-z^2/(4*f))^2+(d-z)^2)
>(%i3) taylor(%, [z, zp, d], 0, 4);
>
>(%o3) 2*f+(z^2+(-2*zp-2*d)*z+zp^2+d^2)/(2*f)
> +((2*zp+2*d)*z^3+(-5*zp^2-5*d^2)*z^2+(4*zp^3+4*d^3)*z-zp^4-d^4)
> /(8*f^3)
>(%i4) diff(%, z);
>
>Maxima encountered a Lisp error:
>
> Error in PROGN [or a callee]: Bad plist (($ZP ((4 . 1)) 0
This is a bug; I'll report it to our bug list. Thanks for bringing it to
our attention.
There are several other known bugs that involve 'bad plist' messages, so
fixing this bug might
fix several other bugs.
Welcome to Maxima; if you have other questions, send us a note.
Barton
P.S. A possible fix is to change old-get to
(defun old-get (plist tag)
(oldget plist tag))
Currently old-get is a macro: `(getf (cdr ,plist) ,tag)); here I changed
it to a function.
After rebuilding, the bug is gone and the testsuite reports no errors.