Strange float behavior



C Y  writes:

> I'm seeing some odd behavior of the float command when I try
> using it on output from the current units package.  Can anyone
> reproduce this or is this just a local quirk?  (Using latest cvs
> everything + sbcl 0.8.20 + post_eval_function patches.)

I think this is clearer with explicit calls to `processunits' instead
of the post_eval_function patch:

(c1) load("unit/unit.mac");
[...]
(d1)         /usr/src/maxima-newlines/share/contrib/unit/unit.mac
(c2) setunits([km,minutes]);
(d2)                                 done
(c3) load("physics/physconst.mac");
(d3)         /usr/src/maxima-newlines/share/physics/physconst.mac
(c4) %%c,numer;
                                  299792458 m
(d4)                              -----------
                                       s
(c5) processunits(%);
                                 449688687 km
(d5)                             ------------
                                   25 %min
(c6) float(%);
                               1.798754748e+7 km
(d6)                           -----------------
                                     %min
(c7) processunits(%);
                                  35975095 km
(d7)                              -----------
                                    2 %min
(c8) d5-d7;
                                       km
(d8)                               - -------
                                     50 %min

So it seems just a numerical error introduced by rationalizing d6.

Wolfgang