libstdc++: Import MSVC floating-point std::to_chars testcases

The testcases are imported almost verbatim, with the only change being
to the -double_nan and -float_nan testcases.  We expect these values to
be formatted as "-nan" instead of "-nan(ind)".

libstdc++-v3/ChangeLog:

	* testsuite/20_util/to_chars/double.cc: New test, consisting of
	testcases imported from the MSVC STL testsuite.
	* testsuite/20_util/to_chars/float.cc: Likewise.