Fix for bug [ 1370433 ] trigsimp(sqrt(%i2))!=sqrt(trigsimp(%i2))?


> Doesn't sqrt(x) mean the principal square root?

I actually pursue this question in a chapter of my PhD. Since there are
indeterminates inside the root, there is a need for some better idea here. I
define a "positive real interpretation" of a rational expression. But if the
inside has trig forms, I think I don't say anything useful.