[Modules TS] Module ownership semantics for redeclarations.

When declaring an entity in the "purview" of a module, it's never a
redeclaration of an entity in the purview of a default module or in no module
("in the global module"). Don't consider those other declarations as possible
redeclaration targets if they're not visible, and reject any cases where we
pick a prior visible declaration that violates this rule.

This reinstates r315251 and r315256, reverted in r315309 and r315308
respectively, tweaked to avoid triggering a linkage calculation when declaring
implicit special members (this exposed our pre-existing issue with typedef
names for linkage changing the linkage of types whose linkage has already been
computed and cached in more cases). A testcase for that regression has been
added in r315366.


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@315379 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 29 insertions, 1 deletions

diff --git a/lib/Sema/SemaLookup.cpp b/lib/Sema/SemaLookup.cpp
index a89d3ace24..6fb0260075 100644
--- a/lib/Sema/SemaLookup.cpp
+++ b/lib/Sema/SemaLookup.cpp
@@ -1609,11 +1609,39 @@ bool Sema::isVisibleSlow(const NamedDecl *D) {
 }
 
 bool Sema::shouldLinkPossiblyHiddenDecl(LookupResult &R, const NamedDecl *New) {
+  // FIXME: If there are both visible and hidden declarations, we need to take
+  // into account whether redeclaration is possible. Example:
+  // 
+  // Non-imported module:
+  //   int f(T);        // #1
+  // Some TU:
+  //   static int f(U); // #2, not a redeclaration of #1
+  //   int f(T);        // #3, finds both, should link with #1 if T != U, but
+  //                    // with #2 if T == U; neither should be ambiguous.
   for (auto *D : R) {
     if (isVisible(D))
       return true;
+    assert(D->isExternallyDeclarable() &&
+           "should not have hidden, non-externally-declarable result here");
   }
-  return New->isExternallyVisible();
+
+  // This function is called once "New" is essentially complete, but before a
+  // previous declaration is attached. We can't query the linkage of "New" in
+  // general, because attaching the previous declaration can change the
+  // linkage of New to match the previous declaration.
+  //
+  // However, because we've just determined that there is no *visible* prior
+  // declaration, we can compute the linkage here. There are two possibilities:
+  //
+  //  * This is not a redeclaration; it's safe to compute the linkage now.
+  //
+  //  * This is a redeclaration of a prior declaration that is externally
+  //    redeclarable. In that case, the linkage of the declaration is not
+  //    changed by attaching the prior declaration, because both are externally
+  //    declarable (and thus ExternalLinkage or VisibleNoLinkage).
+  //
+  // FIXME: This is subtle and fragile.
+  return New->isExternallyDeclarable();
 }
 
 /// \brief Retrieve the visible declaration corresponding to D, if any.