Add support for __float128 conversions

[benruijl]

Starting from MPR 4.0, the functions mpfr_set_float128 and mpfr_get_float128 can be used to convert from and to __f128 (see test cases here for usage: https://fossies.org/linux/misc/mpfr-4.1.0.tar.gz/mpfr-4.1.0/tests/tset_float128.c).

It would be great if mpreal supports these conversions as well.

[benruijl]

In my forked version benruijl@9cf67d1 I've added support for __float128 by essentially copying all long double functions. These functions may all need to be wrapped in #if clauses to support platforms that don't have __float128 (I'm also very open to volunteers, as this is not currently a high priority for me). Once this is done, I can make a pull request.

[maxale]

Support for __int128 is also missing.

[advanpix]

Support for __int128 is also missing.

MPFR has no direct support for this GCC-specific extension (__int128) as this is an integer type, and out of scope of MPFR functionality. GMP/MPIR would be a better choice I guess. Or you can add the conversion to/from __int128 if you need this specific functionality.