Support formatting mpf/mpc in binary

This already supported by ``to_str()``, but probably could be done more efficiently:
```pycon
>>> f'{gmpy2.mpfr(1.1):.15b}'
'1.000110011001101p+0'
>>> mpmath.libmp.to_str(x._mpf_, 16, base=2, show_zero_exponent=True, binary_exp=True)[2:]
'1.000110011001101p+0'
```

For compatibility with gmpy2/MPFR we could utilize same "b" format type.