make sure eval_convert_to() do not terminate with super large number

this change is a follow-up of d51f2e9. it intends to
address the exception thrown in a noexcept functon.

a minimal reproducer looks like

```c++

int main() {
    std::string s = "32767456456456456456545678943512357658768763546575675";
    boost::multiprecision::cpp_int num(s);
    std::cout << num.convert_to<float>() << std::endl;
}
```

since boost 1.79, the code above terminates like
```
Program returned: 139
Program stderr
terminate called after throwing an instance of 'boost::wrapexcept<std::domain_error>'
  what():  Error in function float_next<float>(float): Argument must be finite, but got inf
Program terminated with signal: SIGSEGV
```

because `float_next_imp()` throws 'boost::wrapexcept<std::domain_error>'
if the number is NAN of INF. and `eval_convert_to()` is marked as
`noexcept(boost::multiprecision::detail::is_arithmetic<R>::value &&
          std::numeric_limits<R>::has_infinity)`,
but only `overflow_error` is ignored in the policy passed to
`float_next()`.

so, in this change, `std::domain_error` is ignored as well, so that
``num.convert_to<float>()` returns a NaN in this case.

Refs boostorg#553

Signed-off-by: Kefu Chai <tchaikov@gmail.com>

include/boost/multiprecision/cpp_int/misc.hpp

@@ -184,7 +184,9 @@ eval_convert_to(R* result, const cpp_int_backend<MinBits1, MaxBits1, SignType1,
 
 template <class R, std::size_t MinBits1, std::size_t MaxBits1, cpp_integer_type SignType1, cpp_int_check_type Checked1, class Allocator1>
 inline BOOST_MP_CXX14_CONSTEXPR typename std::enable_if<std::is_floating_point<R>::value && !is_trivial_cpp_int<cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value, void>::type
-eval_convert_to(R* result, const cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1>& backend) noexcept(boost::multiprecision::detail::is_arithmetic<R>::value && std::numeric_limits<R>::has_infinity)
+eval_convert_to(R* result, const cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1>& backend) noexcept(boost::multiprecision::detail::is_arithmetic<R>::value &&
+                                                                                                                         (std::numeric_limits<R>::has_infinity ||
+                                                                                                                          std::numeric_limits<R>::has_quiet_NaN))
 {
    BOOST_MP_FLOAT128_USING using std::ldexp;
    if (eval_is_zero(backend))
@@ -244,10 +246,11 @@ eval_convert_to(R* result, const cpp_int_backend<MinBits1, MaxBits1, SignType1,
          if ((eval_lsb_imp(backend) < static_cast<std::size_t>(bits)) || eval_bit_test(backend, static_cast<std::size_t>(bits + 1)))
          {
             #ifdef BOOST_MP_MATH_AVAILABLE
-            BOOST_IF_CONSTEXPR(std::numeric_limits<R>::has_infinity)
+            BOOST_IF_CONSTEXPR(std::numeric_limits<R>::has_infinity || std::numeric_limits<R>::has_quiet_NaN)
             {
                // Must NOT throw:
-               *result = boost::math::float_next(*result, boost::math::policies::make_policy(boost::math::policies::overflow_error<boost::math::policies::ignore_error>()));
+               *result = boost::math::float_next(*result, boost::math::policies::make_policy(boost::math::policies::overflow_error<boost::math::policies::ignore_error>(),
+                                                                                             boost::math::policies::domain_error<boost::math::policies::ignore_error>()));
             }
             else
             {