[PATCH] D148274: [clang] Fix overly aggressive lifetime checks for parenthesized aggregate initialization

[Shafik Yaghmour via Phabricator via cfe-commits]

shafik added inline comments.


================
Comment at: clang/include/clang/Sema/Initialization.h:400
+  static InitializedEntity InitializeMemberFromParenAggInit(FieldDecl *Member) {
+    return InitializedEntity(Member, nullptr, false, false, true);
+  }
----------------



================
Comment at: clang/include/clang/Sema/Initialization.h:509
   bool isImplicitMemberInitializer() const {
-    return getKind() == EK_Member && Variable.IsImplicitFieldInit;
+    return (getKind() == EK_Member || getKind() == EK_ParenAggInitMember) &&
+           Variable.IsImplicitFieldInit;
----------------
Does this really apply? It looks like this function is only used in `canPerformArrayCopy`


================
Comment at: clang/lib/Sema/SemaAccess.cpp:1657
     const FieldDecl *Field = cast<FieldDecl>(Entity.getDecl());
     PD = PDiag(diag::err_access_field_ctor);
     PD << Field->getType() << getSpecialMember(Constructor);
----------------
It looks like don't have a test that triggers this diagnostic, I was looking for `field of type.*has.*` 


================
Comment at: clang/lib/Sema/SemaInit.cpp:583
+              EK == InitializedEntity::EK_Member ||
+              EK == InitializedEntity::EK_ParenAggInitMember)
             SemaRef.Diag(Entity.getDecl()->getLocation(),
----------------
I don't see a test triggering this diagnostic for parens init.


================
Comment at: clang/lib/Sema/SemaInit.cpp:597
+          EK == InitializedEntity::EK_Member ||
+          EK == InitializedEntity::EK_ParenAggInitMember)
         SemaRef.Diag(Entity.getDecl()->getLocation(),
----------------
I don't see a test triggering this diagnostic for parens init.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D148274/new/

https://reviews.llvm.org/D148274