Plotting singular functions



> Realpart seems to be working, but plot2d(asin(x),[x,-2,2]) looks 
> strange. Left part should be -%PI/2. Why it positive?

Realpart is broken when numer=true or float=true.  See bug 940835.  A
related bug is 816166.  I will write a fix for 2pistrip (the responsible
function) next week.  (I already have a draft.)

I am assuming that the correct value for carg(exp(%i*10)) (in the
symbolic, not numeric, case) is 10-4*%pi.  But how far can we take that?
Should carg(exp(%i*(sqrt(5)+3))) be sqrt(5)+3-2*%pi?

After all, for carg(exp(%i*x)) I will give x, and not fff(x), where fff
is some range-reduction function (which does not currently exist).

This is similar to the range reduction problem for trigonometric
functions, with the MAJOR difference that sin(10) = sin(10-4*%pi).

        -s