[Clang] Don't allow implicit this access when checking function constraints

[zyn0217]

We allowed implicit this access when checking associated constraints after CWG2369. As a result, some of the invalid function call expressions were not properly SFINAE'ed out and ended up as hard errors at evaluation time.

We tried fixing that by mucking around the CurContext, but that spawned additional breakages and I think it's probably safe to revert to the previous behavior to avoid churns.

Though there is CWG2589, which justifies the previous change, it's not what we're pursuing now.

Fixes #151271
Fixes #145505

Reopens #115838

[llvmbot]

@llvm/pr-subscribers-clang

Author: Younan Zhang (zyn0217)

Changes

We allowed implicit this access when checking associated constraints after CWG2369. As a result, some of the invalid function call expressions were not properly SFINAE'ed out and ended up as hard errors at evaluation time.

We tried fixing that by mucking around the CurContext, but that spawned additional breakages and I think it's probably safe to revert to the previous behavior to avoid churns.

Though there is CWG2589, which justifies the previous change, it's not what we're pursuing now.

Fixes #151271

Reopens #115838

---

Full diff: https://github.com/llvm/llvm-project/pull/151276.diff

4 Files Affected:

• (modified) clang/docs/ReleaseNotes.rst (-1)
• (modified) clang/lib/Sema/SemaConcept.cpp (-4)
• (modified) clang/lib/Sema/SemaTemplate.cpp (-2)
• (modified) clang/test/SemaTemplate/concepts.cpp (+17-17)

diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index fcd3887ec7a09..8cf8589005654 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -957,7 +957,6 @@ Bug Fixes to C++ Support
 - Fixed an access checking bug when initializing non-aggregates in default arguments (#GH62444), (#GH83608)
 - Fixed a pack substitution bug in deducing class template partial specializations. (#GH53609)
 - Fixed a crash when constant evaluating some explicit object member assignment operators. (#GH142835)
-- Fixed an access checking bug when substituting into concepts (#GH115838)
 - Fix a bug where private access specifier of overloaded function not respected. (#GH107629)
 - Correctly handles calling an explicit object member function template overload set
   through its address (``(&Foo::bar<baz>)()``).
diff --git a/clang/lib/Sema/SemaConcept.cpp b/clang/lib/Sema/SemaConcept.cpp
index 834417f8e15ac..20567a6d9d1a8 100644
--- a/clang/lib/Sema/SemaConcept.cpp
+++ b/clang/lib/Sema/SemaConcept.cpp
@@ -1097,10 +1097,6 @@ static bool CheckFunctionConstraintsWithoutInstantiation(
   }
 
   Sema::ContextRAII SavedContext(SemaRef, FD);
-  std::optional<Sema::CXXThisScopeRAII> ThisScope;
-  if (auto *Method = dyn_cast<CXXMethodDecl>(FD))
-    ThisScope.emplace(SemaRef, /*Record=*/Method->getParent(),
-                      /*ThisQuals=*/Method->getMethodQualifiers());
   return SemaRef.CheckConstraintSatisfaction(
       Template, TemplateAC, MLTAL, PointOfInstantiation, Satisfaction);
 }
diff --git a/clang/lib/Sema/SemaTemplate.cpp b/clang/lib/Sema/SemaTemplate.cpp
index b76619fc50268..f067aedf1cc9f 100644
--- a/clang/lib/Sema/SemaTemplate.cpp
+++ b/clang/lib/Sema/SemaTemplate.cpp
@@ -4749,8 +4749,6 @@ Sema::CheckConceptTemplateId(const CXXScopeSpec &SS,
   EnterExpressionEvaluationContext EECtx{
       *this, ExpressionEvaluationContext::Unevaluated, CSD};
 
-  ContextRAII CurContext(*this, CSD->getDeclContext(),
-                         /*NewThisContext=*/false);
   if (!AreArgsDependent &&
       CheckConstraintSatisfaction(
           NamedConcept, AssociatedConstraint(NamedConcept->getConstraintExpr()),
diff --git a/clang/test/SemaTemplate/concepts.cpp b/clang/test/SemaTemplate/concepts.cpp
index 62a4f95d79c74..6d6052791507d 100644
--- a/clang/test/SemaTemplate/concepts.cpp
+++ b/clang/test/SemaTemplate/concepts.cpp
@@ -1228,25 +1228,25 @@ template <KnownKind T> struct KnownType {
 
 }
 
-namespace GH115838 {
+namespace CWG2369_Regression_2 {
 
-template<typename T> concept has_x = requires(T t) {{ t.x };};
-
-class Publ { public:    int x = 0; };
-class Priv { private:   int x = 0; };
-class Prot { protected: int x = 0; };
-class Same { protected: int x = 0; };
-
-template<typename T> class D;
-template<typename T> requires ( has_x<T>) class D<T>: public T { public: static constexpr bool has = 1; };
-template<typename T> requires (!has_x<T>) class D<T>: public T { public: static constexpr bool has = 0; };
+template <typename T>
+concept HasFastPropertyForAttribute =
+    requires(T element, int name) { element.propertyForAttribute(name); };
+
+template <typename OwnerType>
+struct SVGPropertyOwnerRegistry {
+  static int fastAnimatedPropertyLookup() {
+    static_assert (HasFastPropertyForAttribute<OwnerType>);
+    return 1;
+  }
+};
 
-// "Same" is identical to "Prot" but queried before used.
-static_assert(!has_x<Same>,  "Protected should be invisible.");
-static_assert(!D<Same>::has, "Protected should be invisible.");
+class SVGCircleElement {
+  friend SVGPropertyOwnerRegistry<SVGCircleElement>;
+  void propertyForAttribute(int);
+};
 
-static_assert( D<Publ>::has, "Public should be visible.");
-static_assert(!D<Priv>::has, "Private should be invisible.");
-static_assert(!D<Prot>::has, "Protected should be invisible.");
+int i = SVGPropertyOwnerRegistry<SVGCircleElement>::fastAnimatedPropertyLookup();
 
 }

[zyn0217]

@tstellar Can you help us check if clang compiles webkitgtk with this patch? thanks!

[zyn0217]

(I'll issue a backport in a few days)

[zyn0217]

#152073