[libclang] Fix cursors for arguments of Subscript and Call operators

The DeclRefExpr of CXXOperatorCallExpr refering to the custom operator
is visited before the arguments to the operator call. For the Call and
Subscript operator the range of this DeclRefExpr includes the whole call
expression, so that all tokens in that range were mapped to the operator
function, even the tokens of the arguments.

Fix this by ensuring that this particular DeclRefExpr is visited last.

Fixes PR25775.

Fix by Nikolai Kosjar.

Differential Revision: https://reviews.llvm.org/D40481

llvm-svn: 340521

2 files changed, 215 insertions, 1 deletions

diff --git a/clang/test/Index/annotate-operator-call-expr.cpp b/clang/test/Index/annotate-operator-call-expr.cpp
new file mode 100644
index 00000000000..facb841ce33
--- /dev/null
+++ b/clang/test/Index/annotate-operator-call-expr.cpp
@@ -0,0 +1,84 @@
+struct Foo {
+  int operator[](int key);
+  int operator()(int key = 2);
+};
+
+void testFoo(Foo foo, int index) {
+  foo();
+  foo(index);
+
+  foo[index];
+  foo[index + index];
+
+  foo[foo[index]];
+  foo[foo() + foo[index]];
+  foo[foo(index) + foo[index]];
+}
+
+// RUN: c-index-test -test-annotate-tokens=%s:7:1:7:100 %s -std=c++11 -Wno-unused-value | FileCheck %s -check-prefix=CHECK1
+// CHECK1: Identifier: "foo" [7:3 - 7:6] DeclRefExpr=foo:6:18
+// CHECK1: Punctuation: "(" [7:6 - 7:7] DeclRefExpr=operator():3:7 RefName=[7:6 - 7:7] RefName=[7:7 - 7:8]
+// CHECK1: Punctuation: ")" [7:7 - 7:8] DeclRefExpr=operator():3:7 RefName=[7:6 - 7:7] RefName=[7:7 - 7:8]
+// CHECK1: Punctuation: ";" [7:8 - 7:9] CompoundStmt=
+
+// RUN: c-index-test -test-annotate-tokens=%s:8:1:8:100 %s -std=c++11 -Wno-unused-value | FileCheck %s -check-prefix=CHECK2
+// CHECK2: Punctuation: "(" [8:6 - 8:7] DeclRefExpr=operator():3:7 RefName=[8:6 - 8:7] RefName=[8:12 - 8:13]
+// CHECK2: Identifier: "index" [8:7 - 8:12] DeclRefExpr=index:6:27
+// CHECK2: Punctuation: ")" [8:12 - 8:13] DeclRefExpr=operator():3:7 RefName=[8:6 - 8:7] RefName=[8:12 - 8:13]
+// CHECK2: Punctuation: ";" [8:13 - 8:14] CompoundStmt=
+
+// RUN: c-index-test -test-annotate-tokens=%s:10:1:10:100 %s -std=c++11 -Wno-unused-value | FileCheck %s -check-prefix=CHECK3
+// CHECK3: Identifier: "foo" [10:3 - 10:6] DeclRefExpr=foo:6:18
+// CHECK3: Punctuation: "[" [10:6 - 10:7] DeclRefExpr=operator[]:2:7 RefName=[10:6 - 10:7] RefName=[10:12 - 10:13]
+// CHECK3: Identifier: "index" [10:7 - 10:12] DeclRefExpr=index:6:27
+// CHECK3: Punctuation: "]" [10:12 - 10:13] DeclRefExpr=operator[]:2:7 RefName=[10:6 - 10:7] RefName=[10:12 - 10:13]
+// CHECK3: Punctuation: ";" [10:13 - 10:14] CompoundStmt=
+
+// RUN: c-index-test -test-annotate-tokens=%s:11:1:11:100 %s -std=c++11 -Wno-unused-value | FileCheck %s -check-prefix=CHECK4
+// CHECK4: Identifier: "foo" [11:3 - 11:6] DeclRefExpr=foo:6:18
+// CHECK4: Punctuation: "[" [11:6 - 11:7] DeclRefExpr=operator[]:2:7 RefName=[11:6 - 11:7] RefName=[11:20 - 11:21]
+// CHECK4: Identifier: "index" [11:7 - 11:12] DeclRefExpr=index:6:27
+// CHECK4: Punctuation: "+" [11:13 - 11:14] BinaryOperator=
+// CHECK4: Identifier: "index" [11:15 - 11:20] DeclRefExpr=index:6:27
+// CHECK4: Punctuation: "]" [11:20 - 11:21] DeclRefExpr=operator[]:2:7 RefName=[11:6 - 11:7] RefName=[11:20 - 11:21]
+// CHECK4: Punctuation: ";" [11:21 - 11:22] CompoundStmt=
+
+// RUN: c-index-test -test-annotate-tokens=%s:13:1:13:100 %s -std=c++11 -Wno-unused-value | FileCheck %s -check-prefix=CHECK5
+// CHECK5: Identifier: "foo" [13:3 - 13:6] DeclRefExpr=foo:6:18
+// CHECK5: Punctuation: "[" [13:6 - 13:7] DeclRefExpr=operator[]:2:7 RefName=[13:6 - 13:7] RefName=[13:17 - 13:18]
+// CHECK5: Identifier: "foo" [13:7 - 13:10] DeclRefExpr=foo:6:18
+// CHECK5: Punctuation: "[" [13:10 - 13:11] DeclRefExpr=operator[]:2:7 RefName=[13:10 - 13:11] RefName=[13:16 - 13:17]
+// CHECK5: Identifier: "index" [13:11 - 13:16] DeclRefExpr=index:6:27
+// CHECK5: Punctuation: "]" [13:16 - 13:17] DeclRefExpr=operator[]:2:7 RefName=[13:10 - 13:11] RefName=[13:16 - 13:17]
+// CHECK5: Punctuation: "]" [13:17 - 13:18] DeclRefExpr=operator[]:2:7 RefName=[13:6 - 13:7] RefName=[13:17 - 13:18]
+// CHECK5: Punctuation: ";" [13:18 - 13:19] CompoundStmt=
+
+// RUN: c-index-test -test-annotate-tokens=%s:14:1:14:100 %s -std=c++11 -Wno-unused-value | FileCheck %s -check-prefix=CHECK6
+// CHECK6: Identifier: "foo" [14:3 - 14:6] DeclRefExpr=foo:6:18
+// CHECK6: Punctuation: "[" [14:6 - 14:7] DeclRefExpr=operator[]:2:7 RefName=[14:6 - 14:7] RefName=[14:25 - 14:26]
+// CHECK6: Identifier: "foo" [14:7 - 14:10] DeclRefExpr=foo:6:18
+// CHECK6: Punctuation: "(" [14:10 - 14:11] DeclRefExpr=operator():3:7 RefName=[14:10 - 14:11] RefName=[14:11 - 14:12]
+// CHECK6: Punctuation: ")" [14:11 - 14:12] DeclRefExpr=operator():3:7 RefName=[14:10 - 14:11] RefName=[14:11 - 14:12]
+// CHECK6: Punctuation: "+" [14:13 - 14:14] BinaryOperator=
+// CHECK6: Identifier: "foo" [14:15 - 14:18] DeclRefExpr=foo:6:18
+// CHECK6: Punctuation: "[" [14:18 - 14:19] DeclRefExpr=operator[]:2:7 RefName=[14:18 - 14:19] RefName=[14:24 - 14:25]
+// CHECK6: Identifier: "index" [14:19 - 14:24] DeclRefExpr=operator[]:2:7 RefName=[14:6 - 14:7] RefName=[14:25 - 14:26]
+// CHECK6: Punctuation: "]" [14:24 - 14:25] DeclRefExpr=operator[]:2:7 RefName=[14:18 - 14:19] RefName=[14:24 - 14:25]
+// CHECK6: Punctuation: "]" [14:25 - 14:26] DeclRefExpr=operator[]:2:7 RefName=[14:6 - 14:7] RefName=[14:25 - 14:26]
+// CHECK6: Punctuation: ";" [14:26 - 14:27] CompoundStmt=
+
+// RUN: c-index-test -test-annotate-tokens=%s:15:1:15:100 %s -std=c++11 -Wno-unused-value | FileCheck %s -check-prefix=CHECK7
+// CHECK7: Identifier: "foo" [15:3 - 15:6] DeclRefExpr=foo:6:18
+// CHECK7: Punctuation: "[" [15:6 - 15:7] DeclRefExpr=operator[]:2:7 RefName=[15:6 - 15:7] RefName=[15:30 - 15:31]
+// CHECK7: Identifier: "foo" [15:7 - 15:10] DeclRefExpr=foo:6:18
+// CHECK7: Punctuation: "(" [15:10 - 15:11] DeclRefExpr=operator():3:7 RefName=[15:10 - 15:11] RefName=[15:16 - 15:17]
+// CHECK7: Identifier: "index" [15:11 - 15:16] DeclRefExpr=index:6:27
+// CHECK7: Punctuation: ")" [15:16 - 15:17] DeclRefExpr=operator():3:7 RefName=[15:10 - 15:11] RefName=[15:16 - 15:17]
+// CHECK7: Punctuation: "+" [15:18 - 15:19] BinaryOperator=
+// CHECK7: Identifier: "foo" [15:20 - 15:23] DeclRefExpr=foo:6:18
+// CHECK7: Punctuation: "[" [15:23 - 15:24] DeclRefExpr=operator[]:2:7 RefName=[15:23 - 15:24] RefName=[15:29 - 15:30]
+// CHECK7: Identifier: "index" [15:24 - 15:29] DeclRefExpr=index:6:27
+// CHECK7: Punctuation: "]" [15:29 - 15:30] DeclRefExpr=operator[]:2:7 RefName=[15:23 - 15:24] RefName=[15:29 - 15:30]
+// CHECK7: Punctuation: "]" [15:30 - 15:31] DeclRefExpr=operator[]:2:7 RefName=[15:6 - 15:7] RefName=[15:30 - 15:31]
+// CHECK7: Punctuation: ";" [15:31 - 15:32] CompoundStmt=
+
diff --git a/clang/tools/libclang/CIndex.cpp b/clang/tools/libclang/CIndex.cpp
index 92df69da4d4..6d2cf888f38 100644
--- a/clang/tools/libclang/CIndex.cpp
+++ b/clang/tools/libclang/CIndex.cpp
@@ -6794,11 +6794,18 @@ class AnnotateTokensWorker {
   SourceManager &SrcMgr;
   bool HasContextSensitiveKeywords;
 
+  struct PostChildrenAction {
+    CXCursor cursor;
+    enum Action { Invalid, Ignore, Postpone } action;
+  };
+  using PostChildrenActions = SmallVector<PostChildrenAction, 0>;
+
   struct PostChildrenInfo {
     CXCursor Cursor;
     SourceRange CursorRange;
     unsigned BeforeReachingCursorIdx;
     unsigned BeforeChildrenTokenIdx;
+    PostChildrenActions ChildActions;
   };
   SmallVector<PostChildrenInfo, 8> PostChildrenInfos;
 
@@ -6844,7 +6851,13 @@ public:
 
   void VisitChildren(CXCursor C) { AnnotateVis.VisitChildren(C); }
   enum CXChildVisitResult Visit(CXCursor cursor, CXCursor parent);
+  bool IsIgnoredChildCursor(CXCursor cursor) const;
+  PostChildrenActions DetermineChildActions(CXCursor Cursor) const;
+
   bool postVisitChildren(CXCursor cursor);
+  void HandlePostPonedChildCursors(const PostChildrenInfo &Info);
+  void HandlePostPonedChildCursor(CXCursor Cursor, unsigned StartTokenIndex);
+
   void AnnotateTokens();
   
   /// Determine whether the annotator saw any cursors that have 
@@ -6865,6 +6878,67 @@ void AnnotateTokensWorker::AnnotateTokens() {
   AnnotateVis.visitFileRegion();
 }
 
+bool AnnotateTokensWorker::IsIgnoredChildCursor(CXCursor cursor) const {
+  if (PostChildrenInfos.empty())
+    return false;
+
+  for (const auto &ChildAction : PostChildrenInfos.back().ChildActions) {
+    if (ChildAction.cursor == cursor &&
+        ChildAction.action == PostChildrenAction::Ignore) {
+      return true;
+    }
+  }
+
+  return false;
+}
+
+const CXXOperatorCallExpr *GetSubscriptOrCallOperator(CXCursor Cursor) {
+  if (!clang_isExpression(Cursor.kind))
+    return nullptr;
+
+  const Expr *E = getCursorExpr(Cursor);
+  if (const auto *OCE = dyn_cast<CXXOperatorCallExpr>(E)) {
+    const OverloadedOperatorKind Kind = OCE->getOperator();
+    if (Kind == OO_Call || Kind == OO_Subscript)
+      return OCE;
+  }
+
+  return nullptr;
+}
+
+AnnotateTokensWorker::PostChildrenActions
+AnnotateTokensWorker::DetermineChildActions(CXCursor Cursor) const {
+  PostChildrenActions actions;
+
+  // The DeclRefExpr of CXXOperatorCallExpr refering to the custom operator is
+  // visited before the arguments to the operator call. For the Call and
+  // Subscript operator the range of this DeclRefExpr includes the whole call
+  // expression, so that all tokens in that range would be mapped to the
+  // operator function, including the tokens of the arguments. To avoid that,
+  // ensure to visit this DeclRefExpr as last node.
+  if (const auto *OCE = GetSubscriptOrCallOperator(Cursor)) {
+    const Expr *Callee = OCE->getCallee();
+    if (const ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(Callee)) {
+      const Expr *SubExpr = ICE->getSubExpr();
+      if (const DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(SubExpr)) {
+        const Decl *parentDecl = getCursorParentDecl(Cursor);
+        CXTranslationUnit TU = clang_Cursor_getTranslationUnit(Cursor);
+
+        // Visit the DeclRefExpr as last.
+        CXCursor cxChild = MakeCXCursor(DRE, parentDecl, TU);
+        actions.push_back({cxChild, PostChildrenAction::Postpone});
+
+        // The parent of the DeclRefExpr, an ImplicitCastExpr, has an equally
+        // wide range as the DeclRefExpr. We can skip visiting this entirely.
+        cxChild = MakeCXCursor(ICE, parentDecl, TU);
+        actions.push_back({cxChild, PostChildrenAction::Ignore});
+      }
+    }
+  }
+
+  return actions;
+}
+
 static inline void updateCursorAnnotation(CXCursor &Cursor,
                                           const CXCursor &updateC) {
   if (clang_isInvalid(updateC.kind) || !clang_isInvalid(Cursor.kind))
@@ -6941,7 +7015,10 @@ AnnotateTokensWorker::Visit(CXCursor cursor, CXCursor parent) {
   SourceRange cursorRange = getRawCursorExtent(cursor);
   if (cursorRange.isInvalid())
     return CXChildVisit_Recurse;
-      
+
+  if (IsIgnoredChildCursor(cursor))
+    return CXChildVisit_Continue;
+
   if (!HasContextSensitiveKeywords) {
     // Objective-C properties can have context-sensitive keywords.
     if (cursor.kind == CXCursor_ObjCPropertyDecl) {
@@ -7089,6 +7166,7 @@ AnnotateTokensWorker::Visit(CXCursor cursor, CXCursor parent) {
   Info.CursorRange = cursorRange;
   Info.BeforeReachingCursorIdx = BeforeReachingCursorIdx;
   Info.BeforeChildrenTokenIdx = NextToken();
+  Info.ChildActions = DetermineChildActions(cursor);
   PostChildrenInfos.push_back(Info);
 
   return CXChildVisit_Recurse;
@@ -7101,6 +7179,8 @@ bool AnnotateTokensWorker::postVisitChildren(CXCursor cursor) {
   if (!clang_equalCursors(Info.Cursor, cursor))
     return false;
 
+  HandlePostPonedChildCursors(Info);
+
   const unsigned BeforeChildren = Info.BeforeChildrenTokenIdx;
   const unsigned AfterChildren = NextToken();
   SourceRange cursorRange = Info.CursorRange;
@@ -7127,6 +7207,56 @@ bool AnnotateTokensWorker::postVisitChildren(CXCursor cursor) {
   return false;
 }
 
+void AnnotateTokensWorker::HandlePostPonedChildCursors(
+    const PostChildrenInfo &Info) {
+  for (const auto &ChildAction : Info.ChildActions) {
+    if (ChildAction.action == PostChildrenAction::Postpone) {
+      HandlePostPonedChildCursor(ChildAction.cursor,
+                                 Info.BeforeChildrenTokenIdx);
+    }
+  }
+}
+
+void AnnotateTokensWorker::HandlePostPonedChildCursor(
+    CXCursor Cursor, unsigned StartTokenIndex) {
+  const auto flags = CXNameRange_WantQualifier | CXNameRange_WantQualifier;
+  unsigned I = StartTokenIndex;
+
+  // The bracket tokens of a Call or Subscript operator are mapped to
+  // CallExpr/CXXOperatorCallExpr because we skipped visiting the corresponding
+  // DeclRefExpr. Remap these tokens to the DeclRefExpr cursors.
+  for (unsigned RefNameRangeNr = 0; I < NumTokens; RefNameRangeNr++) {
+    const CXSourceRange CXRefNameRange =
+        clang_getCursorReferenceNameRange(Cursor, flags, RefNameRangeNr);
+    if (clang_Range_isNull(CXRefNameRange))
+      break; // All ranges handled.
+
+    SourceRange RefNameRange = cxloc::translateCXSourceRange(CXRefNameRange);
+    while (I < NumTokens) {
+      const SourceLocation TokenLocation = GetTokenLoc(I);
+      if (!TokenLocation.isValid())
+        break;
+
+      // Adapt the end range, because LocationCompare() reports
+      // RangeOverlap even for the not-inclusive end location.
+      const SourceLocation fixedEnd =
+          RefNameRange.getEnd().getLocWithOffset(-1);
+      RefNameRange = SourceRange(RefNameRange.getBegin(), fixedEnd);
+
+      const RangeComparisonResult ComparisonResult =
+          LocationCompare(SrcMgr, TokenLocation, RefNameRange);
+
+      if (ComparisonResult == RangeOverlap) {
+        Cursors[I++] = Cursor;
+      } else if (ComparisonResult == RangeBefore) {
+        ++I; // Not relevant token, check next one.
+      } else if (ComparisonResult == RangeAfter) {
+        break; // All tokens updated for current range, check next.
+      }
+    }
+  }
+}
+
 static enum CXChildVisitResult AnnotateTokensVisitor(CXCursor cursor,
                                                      CXCursor parent,
                                                      CXClientData client_data) {