Bug 1741705
<http://sourceforge.net/tracker/index.php?func=detail&aid=1741705&group_id=4933&atid=104933>
says the integral from 0 to 8 is wrong. I have a fix for this that
depends on floor(x/%pi) working correctly when x is a number, or
expression possibly including %pi. I believe floor is working
correctly, and the changes I've made are also working correctly.
However, what should maxima do for the integral from 0 to z, say?
Currently, it returns
sqrt(2)/2*atan(sqrt(2)*sin(z)/cos(z)) + %pi/sqrt(2)
This isn't quite right, unless you carefully select the correct branch
for atan. This is, howveer, the correct answer if z < 2*%pi.
What should maxima do here? What do Macsyma, Maple, and/or
Mathematica return for this integral?
Ray