[clang] clang: Remove useFP16ConversionIntrinsics target option (PR #207212)

[Erich Keane via cfe-commits]

================
@@ -969,19 +969,10 @@ class ScalarExprEmitter : public StmtVisitor<ScalarExprEmitter, mlir::Value> {
     if (srcType->isHalfType() &&
         !cgf.getContext().getLangOpts().NativeHalfType) {
       // Cast to FP using the intrinsic if the half type itself isn't supported.
-      if (mlir::isa<cir::FPTypeInterface>(mlirDstType)) {
----------------
erichkeane wrote:

This is an odd case too, I'd love to see a test to make sure we cover the dest-type being a floating point too (actually a test for both of these would be nice).

https://github.com/llvm/llvm-project/pull/207212