The calculation of the following integral did not converge after several minutes:
integrate( x**10 * cos(x) )
Whereas the equivalent integral written in complex formalism succeed:
x = Symbol("x", real=True)
( integrate( x**10 * (exp(Ix)+exp(-Ix))/2 ) ).expand(complex=True)
The similar issue holds for sin(x).
It's stuck in heurisch (unsurprisingly). Until risch can handle trig functions maybe we should add cases like these to trigintegrate (or manualintegrate).