These pexpect-based GAP classed in Sage are a mess.
The only reason they are still there is that few things are actually
implemented in Cython, avoiding GAP,
and they are somehow faster (while many things are much slower)...
(one needs to update https://trac.sagemath.org/ticket/26902 here)
Anyhow,
The following lines do what is to be expected:
sage: g = SymmetricGroup(3)
: u = g.subgroup([])
: x1 = u.trivial_character().induct(g)
: dummy = g.conjugacy_classes_subgroups()
: x2 = u.trivial_character().induct(g)
: print(x1 == x2)
: print(x1.scalar_product(x1))
: pri
