The test is silly, but it's been with us for a long time. I suppose the reason for the test is that integrate is required to return an expression that is known
to be continuous on the real domain of the integrand, but not necessarily continuous on the complex domain of the integrand. Oh surely
Maxima should be able to detect that this particular integrand is entire.
--Barton
________________________________________
I was surprised to see this (in current git):
(%i1) integrate(x,x,sqrt(a),b);
defint: lower limit of integration must be real; found sqrt(a)
-- an error. To debug this try: debugmode(true);
Isn't this test for real limits a little too strong?
Andrej