Problem with emaxima in CVS



I'm using the current CVS maxima with GNU Emacs 21.3.1.  There seems
to be a problem with the output formatting.  Here's an example:

kill(all);
ratvars(kbt,mu);
f(x):=(kbt*x/mu+1)**(3/2);
order:4;
aa:ratdisrep(taylor(f(x),x,0,order));
bb:integrate(exp(x)/(exp(x)+1)^2*aa,x,minf,inf);
cc:taylor(mu*bb^(2/3),kbt,0,order);

Here's what I get:

          kill(all);
(%o0)                                done
(%i1) ratvars(kbt,mu);
(%o1)                              [kbt, mu]
(%i2) f(x):=(kbt*x/mu+1)**(3/2);
                                     kbt x     3/2
(%o2)                       f(x) := (----- + 1)
                                      mu
(%i3) order:4;
(%o3)                                  4
(%i4) aa:ratdisrep(taylor(f(x),x,0,order));
                      4  4      3  3        2  2
                 3 kbt  x    kbt  x    3 kbt  x    3 kbt x
(%o4)            --------- - ------- + --------- + ------- + 1
                        4         3          2      2 mu
                  128 mu     16 mu       8 mu
(%i5) bb:integrate(exp(x)/(exp(x)+1)^2*aa,x,minf,inf);
Is  mu  positive, negative, or zero?

p;
Is  kbt  positive, negative, or zero?

p;
         4         2    2   2        4    4
                   640 mu  + 80 %pi  kbt  mu  + 7 %pi  kbt
(%o5)              ----------------------------------------
                                         4
                                   640 mu
(%i6)                 cc:taylor(mu*bb^(2/3),kbt,0,order);
                      2    2      4    4
                           %pi  kbt    %pi  kbt
(%o6)/T/              mu + --------- + --------- + . . .
                             12 mu            3
                                        180 mu

The exponents are in the wrong place.  When I run from an xterm or
from xmaxima, everything is fine.

David