Use correct 64bit full multiplication for MinGW on ARM64

[vejbomar]

Update fast_float code to use correct 64bit full multiplication for MinGW on ARM64. This fixes the following error when building with MinGW for ARM64 from https://github.com/Windows-on-ARM-Experiments/mingw-woarm64-build:

In file included from ./boost/charconv/detail/fast_float/fast_float.hpp:11,
                 from libs/charconv/build/../src/from_chars.cpp:16:
./boost/charconv/detail/fast_float/float_common.hpp: In function ‘boost::charconv::detail::fast_float::value128 boost::charconv::detail::fast_float::full_multiplication(uint64_t, uint64_t)’:
./boost/charconv/detail/fast_float/float_common.hpp:254:16: error: ‘_umul128’ was not declared in this scope; did you mean ‘umul128’?
  254 |   answer.low = _umul128(a, b, &high); // _umul128 not available on ARM64
      |                ^~~~~~~~
      |                umul128

Corresponding change in fast_float repository: fastfloat/fast_float#269

Please let me know if there is some other way how the code from fast_float is updated here. If this is fine, I'll create a similar PR in https://github.com/boostorg/json

[mborland]

LGTM. Thanks. @grisumbras is there any need to update in JSON? I thought you were moving to charconv this cycle with dropping GCC 4.X series.

[grisumbras]

I'll be switching to Charconv after the next release (that is, after 1.87).

[mborland]

I'll be switching to Charconv after the next release (that is, after 1.87).

Ok. @vejbomar if you could transpose this into the same place in Boost.JSON for the meantime that would be great. I'm going to merge it and then fix the CI. Need to update the Drone images to LTS.

[vejbomar]

Thank you for the brisk merging! I've created PR for Boost.JSON as suggested: boostorg/json#1057