[Clang][Sema] fix deducing auto& from const int in template parameters is impossible in partial specializations

[jcsxky]

Fix issue
AutoType is possible cv qualified.

Also fixes #75843.

[llvmbot]

@llvm/pr-subscribers-clang

Author: Qizhi Hu (jcsxky)

Changes

Fix issue
AutoType is possible cv qualified.

---

Full diff: https://github.com/llvm/llvm-project/pull/79733.diff

2 Files Affected:

• (modified) clang/lib/Sema/SemaTemplateDeduction.cpp (+1)
• (added) clang/test/SemaTemplate/PR77189.cpp (+22)

diff --git a/clang/lib/Sema/SemaTemplateDeduction.cpp b/clang/lib/Sema/SemaTemplateDeduction.cpp
index 25e58f7bdd953d1..f08577febcd3e80 100644
--- a/clang/lib/Sema/SemaTemplateDeduction.cpp
+++ b/clang/lib/Sema/SemaTemplateDeduction.cpp
@@ -649,6 +649,7 @@ static bool IsPossiblyOpaquelyQualifiedTypeInternal(const Type *T) {
   case Type::PackIndexing:
   case Type::UnresolvedUsing:
   case Type::TemplateTypeParm:
+  case Type::Auto:
     return true;
 
   case Type::ConstantArray:
diff --git a/clang/test/SemaTemplate/PR77189.cpp b/clang/test/SemaTemplate/PR77189.cpp
new file mode 100644
index 000000000000000..dd2e4511a02cf70
--- /dev/null
+++ b/clang/test/SemaTemplate/PR77189.cpp
@@ -0,0 +1,22 @@
+// RUN: %clang_cc1 -fsyntax-only -std=c++17 -verify %s
+// RUN: %clang_cc1 -fsyntax-only -std=c++20 -verify %s
+// expected-no-diagnostics
+
+struct false_type {
+	static constexpr bool value = false;
+};
+
+struct true_type {
+	static constexpr bool value = true;
+};
+
+template <auto& Value, int>
+struct test : false_type {};
+
+template <auto& Value>
+struct test<Value, 0> : true_type {};
+
+int main() {
+    static constexpr int v = 42;
+    static_assert(test<v, 0>::value);
+}
\ No newline at end of file

[cor3ntin]

Oh wow, thanks for this fix!
It will need a Changelog entry.

Does it fixes any of these issues? #68885

[jcsxky]

Oh wow, thanks for this fix! It will need a Changelog entry.

Does it fixes any of these issues? #68885

This patch can't fix #68885. I have looked into the code and the root cause maybe different (probably relate to ExprConst, I'm not sure). Need more work on that issue.

[cor3ntin]

LGTM, thanks!