Again a bug from the Sage groups. Defint breaks with keepfloat:true.
(%i1) integrate(1.0/((1.0-x)^2+1), x, 4, 5);
Maxima encountered a Lisp error:
Error during processing of --eval option "(cl-user::run)":
The value 1.0 is not of type FIXNUM.
Automatically continuing.
To enable the Lisp debugger set *debugger-hook* to nil.
The problem is that real-roots in defint calls solve, which expects
keepfloat to be set to false in the toplevel $solve. The workaround is
to set keepfloat to false in solve (or any code which calls solve).
Andrej