[clang][NFC] Use BuiltinType::isFloatingPoint() in Type::isFloatingType()

[tbaederr]

They both used to do the same comparison of getKind() to Half and Ibm128.

[llvmbot]

@llvm/pr-subscribers-clang

Author: Timm Baeder (tbaederr)

Changes

They both used to do the same comparison of getKind() to Half and Ibm128.

---

Full diff: https://github.com/llvm/llvm-project/pull/139035.diff

1 Files Affected:

• (modified) clang/lib/AST/Type.cpp (+1-2)

diff --git a/clang/lib/AST/Type.cpp b/clang/lib/AST/Type.cpp
index 392a95d042353..82ca2c62eba30 100644
--- a/clang/lib/AST/Type.cpp
+++ b/clang/lib/AST/Type.cpp
@@ -2301,8 +2301,7 @@ bool Type::hasUnsignedIntegerRepresentation() const {
 
 bool Type::isFloatingType() const {
   if (const auto *BT = dyn_cast<BuiltinType>(CanonicalType))
-    return BT->getKind() >= BuiltinType::Half &&
-           BT->getKind() <= BuiltinType::Ibm128;
+    return BT->isFloatingPoint();
   if (const auto *CT = dyn_cast<ComplexType>(CanonicalType))
     return CT->getElementType()->isFloatingType();
   return false;

[tbaederr]

I considered that the comparison is inlined in isFloatingType for performance reasons, but the git log doesn't suggest anything like that.