[clang][bytecode] Change the way we do init chains (#122871)

See the comment in Compiler<>::VisitCXXThisExpr.
We need to mark the InitList explicitly, so we later know what to refer
to when the init chain is active.

Author: tbaederr

clang/lib/AST/ByteCode/Compiler.cpp

@@ -90,6 +90,8 @@ bool InitLink::emit(Compiler<Emitter> *Ctx, const Expr *E) const {
     if (!Ctx->emitConstUint32(Offset, E))
       return false;
     return Ctx->emitArrayElemPtrPopUint32(E);
+  case K_InitList:
+    return true;
   default:
     llvm_unreachable("Unhandled InitLink kind");
   }
@@ -1717,6 +1719,8 @@ bool Compiler<Emitter>::VisitArraySubscriptExpr(const ArraySubscriptExpr *E) {
 template <class Emitter>
 bool Compiler<Emitter>::visitInitList(ArrayRef<const Expr *> Inits,
                                       const Expr *ArrayFiller, const Expr *E) {
+  InitLinkScope<Emitter> ILS(this, InitLink::InitList());
+
   QualType QT = E->getType();
   if (const auto *AT = QT->getAs<AtomicType>())
     QT = AT->getValueType();
@@ -1754,6 +1758,7 @@ bool Compiler<Emitter>::visitInitList(ArrayRef<const Expr *> Inits,
     auto initPrimitiveField = [=](const Record::Field *FieldToInit,
                                   const Expr *Init, PrimType T) -> bool {
       InitStackScope<Emitter> ISS(this, isa<CXXDefaultInitExpr>(Init));
+      InitLinkScope<Emitter> ILS(this, InitLink::Field(FieldToInit->Offset));
       if (!this->visit(Init))
         return false;
 
@@ -1766,6 +1771,7 @@ bool Compiler<Emitter>::visitInitList(ArrayRef<const Expr *> Inits,
                                   const Expr *Init) -> bool {
       InitStackScope<Emitter> ISS(this, isa<CXXDefaultInitExpr>(Init));
       InitLinkScope<Emitter> ILS(this, InitLink::Field(FieldToInit->Offset));
+
       // Non-primitive case. Get a pointer to the field-to-initialize
       // on the stack and recurse into visitInitializer().
       if (!this->emitGetPtrField(FieldToInit->Offset, Init))
@@ -3812,6 +3818,7 @@ template <class Emitter> bool Compiler<Emitter>::visit(const Expr *E) {
 
     if (!this->emitGetPtrLocal(*LocalIndex, E))
       return false;
+    InitLinkScope<Emitter> ILS(this, InitLink::Temp(*LocalIndex));
     return this->visitInitializer(E);
   }
 
@@ -4848,18 +4855,49 @@ bool Compiler<Emitter>::VisitCXXThisExpr(const CXXThisExpr *E) {
   // instance pointer of the current function frame, but e.g. to the declaration
   // currently being initialized. Here we emit the necessary instruction(s) for
   // this scenario.
-  if (!InitStackActive || !E->isImplicit())
+  if (!InitStackActive)
     return this->emitThis(E);
 
-  if (InitStackActive && !InitStack.empty()) {
+  if (!InitStack.empty()) {
+    // If our init stack is, for example:
+    // 0 Stack: 3 (decl)
+    // 1 Stack: 6 (init list)
+    // 2 Stack: 1 (field)
+    // 3 Stack: 6 (init list)
+    // 4 Stack: 1 (field)
+    //
+    // We want to find the LAST element in it that's an init list,
+    // which is marked with the K_InitList marker. The index right
+    // before that points to an init list. We need to find the
+    // elements before the K_InitList element that point to a base
+    // (e.g. a decl or This), optionally followed by field, elem, etc.
+    // In the example above, we want to emit elements [0..2].
     unsigned StartIndex = 0;
+    unsigned EndIndex = 0;
+    // Find the init list.
     for (StartIndex = InitStack.size() - 1; StartIndex > 0; --StartIndex) {
+      if (InitStack[StartIndex].Kind == InitLink::K_InitList ||
+          InitStack[StartIndex].Kind == InitLink::K_This) {
+        EndIndex = StartIndex;
+        --StartIndex;
+        break;
+      }
+    }
+
+    // Walk backwards to find the base.
+    for (; StartIndex > 0; --StartIndex) {
+      if (InitStack[StartIndex].Kind == InitLink::K_InitList)
+        continue;
+
       if (InitStack[StartIndex].Kind != InitLink::K_Field &&
           InitStack[StartIndex].Kind != InitLink::K_Elem)
         break;
     }
 
-    for (unsigned I = StartIndex, N = InitStack.size(); I != N; ++I) {
+    // Emit the instructions.
+    for (unsigned I = StartIndex; I != EndIndex; ++I) {
+      if (InitStack[I].Kind == InitLink::K_InitList)
+        continue;
       if (!InitStack[I].template emit<Emitter>(this, E))
         return false;
     }

clang/lib/AST/ByteCode/Compiler.h

@@ -51,9 +51,11 @@ struct InitLink {
     K_Temp = 2,
     K_Decl = 3,
     K_Elem = 5,
+    K_InitList = 6
   };
 
   static InitLink This() { return InitLink{K_This}; }
+  static InitLink InitList() { return InitLink{K_InitList}; }
   static InitLink Field(unsigned Offset) {
     InitLink IL{K_Field};
     IL.Offset = Offset;

clang/test/AST/ByteCode/records.cpp

@@ -1678,3 +1678,9 @@ namespace NonConst {
     static_assert(s.getSize() == 10, "");
   }
 }
+
+namespace ExplicitThisInTemporary {
+  struct B { B *p = this; };
+  constexpr bool g(B b) { return &b == b.p; }
+  static_assert(g({}), "");
+}

clang/test/AST/ByteCode/unions.cpp

@@ -401,4 +401,24 @@ namespace UnionInBase {
                                         // both-note {{subobject 'y' is not initialized}}
   static_assert(return_uninit().a.x == 2);
 }
+
+/// FIXME: Our diagnostic here is a little off.
+namespace One {
+  struct A { long x; };
+
+  union U;
+  constexpr A foo(U *up);
+  union U {
+    A a = foo(this); // both-note {{in call to 'foo(&u)'}}
+    int y;
+  };
+
+  constexpr A foo(U *up) {
+    return {up->y}; // both-note {{read of member 'y' of union}}
+  }
+
+  constinit U u = {}; // both-error {{constant init}} \
+                      // both-note {{constinit}}
+}
+
 #endif