Add Microsoft mangling for _Float16, similar to technique used for _Complex

[mibintc]

Summary

Microsoft doesn't support _Float16, so this patch invents a mangling using the technique that was used for _Complex, putting _Float16 into __clang namespace. What do you think?

[erichkeane]

FWIW, this demangles on both demangler.com and undname as "void cdecl Float16::f(struct clang::_Float16)", which seems like the right thing, since the "Complex" above demangles to: "void cdecl Complex::f(struct clang::_Complex<int>)"

I also note (since upon first reading) that the "Complex::" and "Float16::" in the demanglings are the namespaces, which are confusingly very similar to the actual names :)

This patch should be all that is required to get _Float16 to work on Windows.

[majnemer]

LGTM

cfe/trunk/lib/AST/MicrosoftMangle.cpp

case BuiltinType::OCLReserveID:
     mangleArtificalTagType(TTK_Struct, "ocl_reserveid");
     break;
 
   case BuiltinType::NullPtr:
     Out << "$$T";
     break;
 
   case BuiltinType::Float16:
+    mangleArtificalTagType(TTK_Struct, "_Float16", {"__clang"});
+    break;
+
   case BuiltinType::Float128:
   case BuiltinType::Half: {
     DiagnosticsEngine &Diags = Context.getDiags();
     unsigned DiagID = Diags.getCustomDiagID(
         DiagnosticsEngine::Error, "cannot mangle this built-in %0 type yet");
     Diags.Report(Range.getBegin(), DiagID)
         << T->getName(Context.getASTContext().getPrintingPolicy()) << Range;
     break;

cfe/trunk/test/CodeGenCXX/mangle-ms.cpp

 namespace Atomic {
 // CHECK-DAG: define dso_local void @"?f@Atomic@@YAXU?$_Atomic@H@__clang@@@Z"(
 void f(_Atomic(int)) {}
 }
 namespace Complex {
 // CHECK-DAG: define dso_local void @"?f@Complex@@YAXU?$_Complex@H@__clang@@@Z"(
 void f(_Complex int) {}
 }
+namespace Float16 {
+// CHECK-DAG: define dso_local void @"?f@Float16@@YAXU_Float16@__clang@@@Z"(
+void f(_Float16) {}
+}
 
 namespace PR26029 {
 template <class>
 struct L {
   L() {}
 };
 template <class>
 class H;