Mofidy and add the floating properties of single, double, and quad precision floating point numbers in FloatProperties.h

[hedingarcia]

Summary

Establish constants values and bit masks for the quad precision floating point numbers. Specifically for non x86, non-Windows x86, and Windows platform that are defined in PlatfromDefs.h. Also, added values that express the number of bits the exponent occupies in the single and double precision. This is in anticipation of the values that a struct that captures floating point encoding will use.

[hedingarcia]

[libc] Fix error message signaled by a static_assert in FloatProperties.h

[aeubanks]

For the subject, make it imperative, not past tense, e.g. "Modify and add ...", and don't end with a period

The description shouldn't be a copy of the subject
It should probably explain the changes a bit more in detail and why you did it. e.g. why did you start using PlatformDefs.h?
Which float properties are you adding?

https://chris.beams.io/posts/git-commit/ is a good read

Phabricator/arcanist is a bit weird when changing commit messages, you have to do it locally (git commit --amend) and in Edit Revision on Phabricator

[sivachandra]

Prefix subjects with [libc]. This change of yours, in its current state, is only fixing a copy paste error. So, the subject line can be something like:

[libc] Fix error message in a static_assert in FloatProperties.h

This is only an example, not a suggestion for this change as it seems to me like you will add much more shortly?

libc/utils/FPUtil/FloatProperties.h

 template <> struct FloatType<uint64_t> {
   static_assert(sizeof(uint64_t) == sizeof(double),
                 "Unexpected size of 'double' type.");
   typedef double Type;
 };
 
 template <> struct FloatType<__uint128_t> {
   static_assert(sizeof(__uint128_t) == sizeof(long double),
-                "Unexpected size of 'double' type.");
+                "Unexpected size of 'long double' type.");
   typedef long double Type;
 };
 
 template <typename BitsType>
 using FloatTypeT = typename FloatType<BitsType>::Type;
 
 } // namespace fputil
 } // namespace __llvm_libc
 
 #endif // LLVM_LIBC_UTILS_FPUTIL_FLOAT_PROPERTIES_H