[clang] [clang] Rework `hasBooleanRepresentation`. (PR #136038)

[Michele Scandale via cfe-commits]

================
@@ -2336,16 +2336,9 @@ bool Type::isArithmeticType() const {
 }
 
 bool Type::hasBooleanRepresentation() const {
-  if (isBooleanType())
-    return true;
-
-  if (const EnumType *ET = getAs<EnumType>())
-    return ET->getDecl()->getIntegerType()->isBooleanType();
-
-  if (const AtomicType *AT = getAs<AtomicType>())
-    return AT->getValueType()->hasBooleanRepresentation();
-
-  return false;
+  if (const auto *VT = dyn_cast<VectorType>(CanonicalType))
----------------
michele-scandale wrote:

I don't think referring to the how the type is lowered in IR make sense for a function in `Type`.

The overall question should be about this family of functions:
```
  /// Determine whether this type has an integer representation
  /// of some sort, e.g., it is an integer type or a vector.
  bool hasIntegerRepresentation() const;

  /// Determine whether this type has an signed integer representation
  /// of some sort, e.g., it is an signed integer type or a vector.
  bool hasSignedIntegerRepresentation() const;

  /// Determine whether this type has an unsigned integer representation
  /// of some sort, e.g., it is an unsigned integer type or a vector.
  bool hasUnsignedIntegerRepresentation() const;

  /// Determine whether this type has a floating-point representation
  /// of some sort, e.g., it is a floating-point type or a vector thereof.
  bool hasFloatingRepresentation() const;
```

In here I'm just trying to have `hasBooleanRepresentation` be consistent with those.

https://github.com/llvm/llvm-project/pull/136038