increasing precision of mpc produces nan

[ofloveandhate]

making a complex number, then increasing its precision, produces (nan,nan). test below.

// doesn't fail
BOOST_AUTO_TEST_CASE(real_precision_increase)
{
	DefaultPrecision(50);
	mpfr_float a(2);
	a.precision(100);
	std::cout << a << '\n';  // just 2
	BOOST_CHECK(abs(mpfr_float(2)-a) < 1e-50);
}

//fails
BOOST_AUTO_TEST_CASE(complex_precision_increase)
{
	DefaultPrecision(50);
	mpfr_complex a(2,3);
	a.precision(100);
	std::cout << a << '\n';  // (nan,nan) for me...
	BOOST_CHECK(abs(mpfr_complex(2,3)-a) < 1e-50);
}

[jzmaddock]

Confirmed, should be fixed in develop.

[jzmaddock]

CI passing, closing.