[clang] [C++20] [Modules] merge the type for anony enum from import and #include (PR #214121)

[Vassil Vassilev via cfe-commits]

================
@@ -2667,6 +2667,33 @@ void Sema::MergeTypedefNameDecl(Scope *S, TypedefNameDecl *New,
       else
         New->setTypeSourceInfo(OldTD->getTypeSourceInfo());
 
+      // See https://github.com/llvm/llvm-project/issues/213299 for the case.
+      //
+      // Ideally, we shall merge the new enum with the old enum when we
+      // creating the new enum. But the enum is anonymous and the typedef's name
+      // come after the enum body, it is too late to merge them. This is the
+      // choice 10 years ago: a523022b5384d7a0901beea7a5f36ee9c09ba339. Actually
+      // what we're merging here is the typedef decls.
+      //
+      // Then https://github.com/llvm/llvm-project/pull/114240 removes the logic
+      // to remove the new ED. This the direct trigger for the above issue of
+      // ambiguous look ups.
+      //
+      // We choose to fix the problem by setting the type of new enum to the
+      // type of old enums. This is consistent with the above call to
+      // setTypeSourceInfo.
+      //
+      // The check `M && M->isGlobalModule()` is not necessary but we hopt to
+      // limit the impact of this change. We can relax the check when we find
+      // similar issue later in other cases.
----------------
vgvassilev wrote:

I don't think a person reading the code needs this and if we apply this rule elsewhere the codebase will look more like prose than code. I think the causality of these changes should be in the commit log.

https://github.com/llvm/llvm-project/pull/214121