[libc][NFC] Refactor FPBits and remove LongDoubleBits specialization (#78192)

This patch removes the `FPBits` specialization for x86 Extended Precision by moving it up to `FPRep`.
It also introduces enums (`Exponent`, `BiasedExponent` and `Significand`) to represent the exponent and significant parts of the floating point numbers. These enums are used to construct and observe floating point representations.

Additionally, we remove `LongDoubleBits.h` that is now unnecessary.

Author: gchatelet

libc/src/__support/FPUtil/FPBits.h

@@ -131,7 +131,7 @@ LIBC_INLINE long double sqrt(long double x) {
     out.set_implicit_bit(1);
     out.set_mantissa((y & (ONE - 1)));
 
-    return out;
+    return out.get_val();
   }
 }
 #endif // LIBC_LONG_DOUBLE_IS_X86_FLOAT80

libc/src/__support/FPUtil/generic/sqrt_80_bit_long_double.h

@@ -61,7 +61,7 @@ LIBC_INLINE long double nextafter(long double from, long double to) {
         from_bits.set_biased_exponent(from_bits.get_biased_exponent() + 1);
         if (from_bits.is_inf())
           raise_except_if_required(FE_OVERFLOW | FE_INEXACT);
-        return from_bits;
+        return from_bits.get_val();
       } else {
         ++int_val;
       }
@@ -75,7 +75,7 @@ LIBC_INLINE long double nextafter(long double from, long double to) {
         // from == 0 is handled separately so decrementing the exponent will not
         // lead to underflow.
         from_bits.set_biased_exponent(from_bits.get_biased_exponent() - 1);
-        return from_bits;
+        return from_bits.get_val();
       } else {
         --int_val;
       }
@@ -94,7 +94,7 @@ LIBC_INLINE long double nextafter(long double from, long double to) {
         // from == 0 is handled separately so decrementing the exponent will not
         // lead to underflow.
         from_bits.set_biased_exponent(from_bits.get_biased_exponent() - 1);
-        return from_bits;
+        return from_bits.get_val();
       } else {
         --int_val;
       }
@@ -109,7 +109,7 @@ LIBC_INLINE long double nextafter(long double from, long double to) {
         from_bits.set_biased_exponent(from_bits.get_biased_exponent() + 1);
         if (from_bits.is_inf())
           raise_except_if_required(FE_OVERFLOW | FE_INEXACT);
-        return from_bits;
+        return from_bits.get_val();
       } else {
         ++int_val;
       }