[Clang] emit -Wignored-qualifiers diagnostic for cv-qualified base classes

[a-tarasyuk]

Fixes #55474

[llvmbot]

@llvm/pr-subscribers-libcxx

@llvm/pr-subscribers-clang

Author: Oleksandr T. (a-tarasyuk)

Changes

Fixes #55474

---

Full diff: https://github.com/llvm/llvm-project/pull/121419.diff

8 Files Affected:

• (modified) clang/docs/ReleaseNotes.rst (+1)
• (modified) clang/include/clang/Basic/DiagnosticSemaKinds.td (+4)
• (modified) clang/lib/Sema/SemaDeclCXX.cpp (+9)
• (modified) clang/test/CXX/drs/cwg4xx.cpp (+2-1)
• (modified) clang/test/CXX/special/class.inhctor/elsewhere.cpp (+3-2)
• (modified) clang/test/Sema/GH70594.cpp (+7-6)
• (modified) clang/test/SemaCXX/class-base-member-init.cpp (+2-1)
• (added) clang/test/SemaCXX/warn-base-type-qualifiers.cpp (+11)

diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index b7da12bcf65818..659f0ebd97fc46 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -733,6 +733,7 @@ Improvements to Clang's diagnostics
       scope.Unlock();
       require(scope); // Warning!  Requires mu1.
     }
+- Clang now emits a ``-Wignored-qualifiers`` diagnostic when a base class includes cv-qualifiers (#GH55474).
 
 Improvements to Clang's time-trace
 ----------------------------------
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index 330ae045616aba..ce8e9fc99dcbff 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -487,6 +487,10 @@ def err_noreturn_non_function : Error<
 def warn_qual_return_type : Warning<
   "'%0' type qualifier%s1 on return type %plural{1:has|:have}1 no effect">,
   InGroup<IgnoredQualifiers>, DefaultIgnore;
+def warn_qual_base_type : Warning<
+  "'%0' qualifier%s1 on base class type %2 have no effect">,
+  InGroup<IgnoredQualifiers>;
+
 def warn_deprecated_redundant_constexpr_static_def : Warning<
   "out-of-line definition of constexpr static data member is redundant "
   "in C++17 and is deprecated">,
diff --git a/clang/lib/Sema/SemaDeclCXX.cpp b/clang/lib/Sema/SemaDeclCXX.cpp
index c5a72cf812ebc9..df0a15fc44e3e3 100644
--- a/clang/lib/Sema/SemaDeclCXX.cpp
+++ b/clang/lib/Sema/SemaDeclCXX.cpp
@@ -2655,6 +2655,15 @@ CXXBaseSpecifier *Sema::CheckBaseSpecifier(CXXRecordDecl *Class,
       return nullptr;
     }
 
+    if (BaseType.hasQualifiers()) {
+      auto Quals =
+          BaseType.getQualifiers().getAsString(Context.getPrintingPolicy());
+      Diag(BaseLoc, diag::warn_qual_base_type)
+          << Quals << std::count(Quals.begin(), Quals.end(), ' ') + 1
+          << BaseType;
+      Diag(BaseLoc, diag::note_base_class_specified_here) << BaseType;
+    }
+
     // For the MS ABI, propagate DLL attributes to base class templates.
     if (Context.getTargetInfo().getCXXABI().isMicrosoft() ||
         Context.getTargetInfo().getTriple().isPS()) {
diff --git a/clang/test/CXX/drs/cwg4xx.cpp b/clang/test/CXX/drs/cwg4xx.cpp
index 825065840f1181..0c3ed2f4ebf534 100644
--- a/clang/test/CXX/drs/cwg4xx.cpp
+++ b/clang/test/CXX/drs/cwg4xx.cpp
@@ -1309,7 +1309,8 @@ namespace cwg484 { // cwg484: yes
   }
   CA::A() {}
 
-  struct B : CA {
+  struct B : CA { // expected-warning {{'const' qualifier on base class type 'CA' (aka 'const cwg484::A') have no effect}} \
+                  // expected-note {{base class 'CA' (aka 'const cwg484::A') specified here}}
     B() : CA() {}
     void f() { return CA::f(); }
   };
diff --git a/clang/test/CXX/special/class.inhctor/elsewhere.cpp b/clang/test/CXX/special/class.inhctor/elsewhere.cpp
index f86f4b86faa7d7..079422fd99d98a 100644
--- a/clang/test/CXX/special/class.inhctor/elsewhere.cpp
+++ b/clang/test/CXX/special/class.inhctor/elsewhere.cpp
@@ -57,9 +57,10 @@ template<typename T> struct F : D<bool> {
 F<bool> fb; // expected-note {{here}}
 
 template<typename T>
-struct G : T {
+struct G : T {        // expected-warning {{'const' qualifier on base class type 'const B1' have no effect}} \
+                      // expected-note {{base class 'const B1' specified here}}
   using T::T;
   G(int &) : G(0) {}
 };
 G<B1> g(123);
-G<const B1> g2(123);
+G<const B1> g2(123); // expected-note {{in instantiation of template class 'G<const B1>' requested here}}
diff --git a/clang/test/Sema/GH70594.cpp b/clang/test/Sema/GH70594.cpp
index ce98e9b12b5cba..91ec31f6b053cb 100644
--- a/clang/test/Sema/GH70594.cpp
+++ b/clang/test/Sema/GH70594.cpp
@@ -1,12 +1,11 @@
 // RUN: %clang_cc1 -fsyntax-only -std=c++11 %s -verify
 // RUN: %clang_cc1 -fsyntax-only -std=c++23 %s -verify
 
-// expected-no-diagnostics
-
 struct A {};
 using CA = const A;
 
-struct S1 : CA {
+struct S1 : CA {           // expected-warning {{'const' qualifier on base class type 'CA' (aka 'const A') have no effect}} \
+                           // expected-note {{base class 'CA' (aka 'const A') specified here}}
   constexpr S1() : CA() {}
 };
 
@@ -14,15 +13,17 @@ struct S2 : A {
   constexpr S2() : CA() {}
 };
 
-struct S3 : CA {
+struct S3 : CA {          // expected-warning {{'const' qualifier on base class type 'CA' (aka 'const A') have no effect}} \
+                          // expected-note {{base class 'CA' (aka 'const A') specified here}}
   constexpr S3() : A() {}
 };
 
 struct Int {};
 
 template <class _Hp>
-struct __tuple_leaf : _Hp {
+struct __tuple_leaf : _Hp {           // expected-warning {{'const' qualifier on base class type 'const Int' have no effect}} \
+                                      // expected-note {{base class 'const Int' specified here}}
   constexpr __tuple_leaf() : _Hp() {}
 };
 
-constexpr __tuple_leaf<const Int> t;
+constexpr __tuple_leaf<const Int> t;  // expected-note {{in instantiation of template class '__tuple_leaf<const Int>' requested here}}
diff --git a/clang/test/SemaCXX/class-base-member-init.cpp b/clang/test/SemaCXX/class-base-member-init.cpp
index a6bb4410a81c84..e46c79a2b68b9c 100644
--- a/clang/test/SemaCXX/class-base-member-init.cpp
+++ b/clang/test/SemaCXX/class-base-member-init.cpp
@@ -103,7 +103,8 @@ namespace PR16596 {
   class A { public: virtual ~A(); };
   typedef const A Foo;
   void Apply(Foo processor);
-  struct Bar : public Foo {};
+  struct Bar : public Foo {}; // expected-warning {{'const' qualifier on base class type 'Foo' (aka 'const PR16596::A') have no effect}}\
+                              // expected-note {{base class 'Foo' (aka 'const PR16596::A') specified here}}
   void Fetch() {
     Apply(Bar());
   }
diff --git a/clang/test/SemaCXX/warn-base-type-qualifiers.cpp b/clang/test/SemaCXX/warn-base-type-qualifiers.cpp
new file mode 100644
index 00000000000000..5074140700cc36
--- /dev/null
+++ b/clang/test/SemaCXX/warn-base-type-qualifiers.cpp
@@ -0,0 +1,11 @@
+// RUN: %clang_cc1 %s -std=c++11 -Wignored-qualifiers -verify
+
+class A { };
+
+typedef const A A_Const;
+class B : public A_Const { }; // expected-warning {{'const' qualifier on base class type 'A_Const' (aka 'const A') have no effect}} \
+                              // expected-note {{base class 'A_Const' (aka 'const A') specified here}}
+
+typedef const volatile A A_Const_Volatile;
+class C : public A_Const_Volatile { }; // expected-warning {{'const volatile' qualifiers on base class type 'A_Const_Volatile' (aka 'const volatile A') have no effect}} \
+                                       // expected-note {{base class 'A_Const_Volatile' (aka 'const volatile A') specified here}}

[erichkeane]

I have a bit of a concern that the template cases are going to be REALLY noisy, but we can deal with restricting individual use cases to a separate warning-group in the future if need-be.

[a-tarasyuk]

...that the template cases are going to be REALLY noisy

@erichkeane It looks like that's why I have two failed tests from llvm-libc++-shared-cfg-in due to the use of qualifiers

[erichkeane]

...that the template cases are going to be REALLY noisy

@erichkeane It looks like that's why I have two failed tests from llvm-libc++-shared-cfg-in due to the use of qualifiers

Yep, that looks to be exactly the case! I'd be interested to see if @ldionne would be willing to accept a std::remove_cv_t or something on those. This diagnostic is likely going to require/cause a level of const/volatile cleanliness in base classes that no one has ever really thought about before.

[a-tarasyuk]

@erichkeane @ldionne should the libcxx tests (common_reference.compile.pass.cpp, apply_extended_types.pass.cpp) be updated or should the new warnings/errors be treated as the new baseline?

[erichkeane]

@erichkeane @ldionne should the libcxx tests (common_reference.compile.pass.cpp, apply_extended_types.pass.cpp) be updated or should the new warnings/errors be treated as the new baseline?

This is entirely a @ldionne question, though I see in github that he's on vacation at the moment, so we'll have to wait until he returns.

[ldionne]

@erichkeane @ldionne should the libcxx tests (common_reference.compile.pass.cpp, apply_extended_types.pass.cpp) be updated or should the new warnings/errors be treated as the new baseline?

This is entirely a @ldionne question, though I see in github that he's on vacation at the moment, so we'll have to wait until he returns.

Thanks for the ping. I've returned but Github doesn't clear the vacation status automatically - fixed now.

This seems reasonable to me, we should "fix" libc++. I think this should do it:

diff --git a/libcxx/include/tuple b/libcxx/include/tuple
index aca14ba408d3..90257e9bae0a 100644
--- a/libcxx/include/tuple
+++ b/libcxx/include/tuple
@@ -257,6 +257,7 @@ template <class... Types>
 #  include <__type_traits/maybe_const.h>
 #  include <__type_traits/nat.h>
 #  include <__type_traits/negation.h>
+#  include <__type_traits/remove_cv.h>
 #  include <__type_traits/remove_cvref.h>
 #  include <__type_traits/remove_reference.h>
 #  include <__type_traits/unwrap_ref.h>
@@ -535,7 +536,7 @@ __memberwise_forward_assign(_Dest& __dest, _Source&& __source, __tuple_types<_Up
 
 template <class... _Tp>
 class _LIBCPP_TEMPLATE_VIS tuple {
-  typedef __tuple_impl<typename __make_tuple_indices<sizeof...(_Tp)>::type, _Tp...> _BaseT;
+  typedef __tuple_impl<typename __make_tuple_indices<sizeof...(_Tp)>::type, __remove_cv_t<_Tp>...> _BaseT;
 
   _BaseT __base_;
 

[erichkeane]

As far as the CFE is concerned, we're OK with this. Though note, I anticipate quite a bit of push back as libraries discover this warning, though most should change their code instead, so I don't anticipate a good reason to revert.

[a-tarasyuk]

@ldionne could you take a look at the latest changes? thanks