Incorrect answer on QF_FP formula with fp.roundToIntegral and RoundingMode

Hi,
For this formula:
```
(declare-fun a () (_ FloatingPoint 2 10))
(declare-fun b () RoundingMode)
(assert (distinct (fp.roundToIntegral b (fp.roundToIntegral b a)) (fp.roundToIntegral b a)))
(check-sat)
```
Z3  incorrectly gives sat, while this formula should be unsat.

OS: Ubuntu 18.04
Commit: 7d976e4
