Add support for C++ namespace-aware typo correction, e.g., correcting

  vector<int>

to

  std::vector<int>

Patch by Kaelyn Uhrain, with minor tweaks + PCH support from me. Fixes
PR5776/<rdar://problem/8652971>.

Thanks Kaelyn!

llvm-svn: 134007

Author: DougGregor

clang/include/clang/Sema/ExternalSemaSource.h

@@ -49,6 +49,11 @@ class ExternalSemaSource : public ExternalASTSource {
   /// instance and factory methods, respectively, with this selector.
   virtual std::pair<ObjCMethodList,ObjCMethodList> ReadMethodPool(Selector Sel);
 
+  /// \brief Load the set of namespaces that are known to the external source,
+  /// which will be used during typo correction.
+  virtual void ReadKnownNamespaces(
+                           llvm::SmallVectorImpl<NamespaceDecl *> &Namespaces);
+  
   /// \brief Do last resort, unqualified lookup on a LookupResult that
   /// Sema cannot find.
   ///

clang/include/clang/Sema/Sema.h

@@ -21,6 +21,7 @@
 #include "clang/Sema/ObjCMethodList.h"
 #include "clang/Sema/DeclSpec.h"
 #include "clang/Sema/LocInfoType.h"
+#include "clang/Sema/TypoCorrection.h"
 #include "clang/AST/Expr.h"
 #include "clang/AST/DeclarationName.h"
 #include "clang/AST/ExternalASTSource.h"
@@ -46,6 +47,8 @@ namespace clang {
   class ASTConsumer;
   class ASTContext;
   class ASTMutationListener;
+  class ASTReader;
+  class ASTWriter;
   class ArrayType;
   class AttributeList;
   class BlockDecl;
@@ -1623,6 +1626,16 @@ class Sema {
                                                    bool ConstThis,
                                                    bool VolatileThis);
 
+  // \brief The set of known/encountered (unique, canonicalized) NamespaceDecls.
+  //
+  // The boolean value will be true to indicate that the namespace was loaded
+  // from an AST/PCH file, or false otherwise.
+  llvm::DenseMap<NamespaceDecl*, bool> KnownNamespaces;
+
+  /// \brief Whether we have already loaded known namespaces from an extenal
+  /// source.
+  bool LoadedExternalKnownNamespaces;
+  
 public:
   /// \brief Look up a name, looking for a single declaration.  Return
   /// null if the results were absent, ambiguous, or overloaded.
@@ -1699,11 +1712,13 @@ class Sema {
     CTC_ObjCMessageReceiver
   };
 
-  DeclarationName CorrectTypo(LookupResult &R, Scope *S, CXXScopeSpec *SS,
-                              DeclContext *MemberContext = 0,
-                              bool EnteringContext = false,
-                              CorrectTypoContext CTC = CTC_Unknown,
-                              const ObjCObjectPointerType *OPT = 0);
+  TypoCorrection CorrectTypo(const DeclarationNameInfo &Typo,
+                             Sema::LookupNameKind LookupKind,
+                             Scope *S, CXXScopeSpec *SS,
+                             DeclContext *MemberContext = NULL,
+                             bool EnteringContext = false,
+                             CorrectTypoContext CTC = CTC_Unknown,
+                             const ObjCObjectPointerType *OPT = NULL);
 
   void FindAssociatedClassesAndNamespaces(Expr **Args, unsigned NumArgs,
                                    AssociatedNamespaceSet &AssociatedNamespaces,
@@ -4721,7 +4736,7 @@ class Sema {
   /// \brief The number of typos corrected by CorrectTypo.
   unsigned TyposCorrected;
 
-  typedef llvm::DenseMap<IdentifierInfo *, std::pair<llvm::StringRef, bool> >
+  typedef llvm::DenseMap<IdentifierInfo *, TypoCorrection>
     UnqualifiedTyposCorrectedMap;
 
   /// \brief A cache containing the results of typo correction for unqualified
@@ -5918,6 +5933,8 @@ class Sema {
 protected:
   friend class Parser;
   friend class InitializationSequence;  
+  friend class ASTReader;
+  friend class ASTWriter;
 
 public:
   /// \brief Retrieve the parser's current scope.

clang/include/clang/Sema/TypoCorrection.h

@@ -0,0 +1,105 @@
+//===--- TypoCorrection.h - Class for typo correction results ---*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file defines the TypoCorrection class, which stores the results of
+// Sema's typo correction (Sema::CorrectTypo).
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_SEMA_TYPOCORRECTION_H
+#define LLVM_CLANG_SEMA_TYPOCORRECTION_H
+
+#include "clang/AST/DeclCXX.h"
+
+namespace clang {
+
+/// @brief Simple class containing the result of Sema::CorrectTypo
+class TypoCorrection {
+public:
+  TypoCorrection(const DeclarationName &Name, NamedDecl *NameDecl,
+                 NestedNameSpecifier *NNS=NULL, unsigned distance=0)
+      : CorrectionName(Name),
+        CorrectionNameSpec(NNS),
+        CorrectionDecl(NameDecl),
+        EditDistance(distance) {}
+
+  TypoCorrection(NamedDecl *Name, NestedNameSpecifier *NNS=NULL,
+                 unsigned distance=0)
+      : CorrectionName(Name->getDeclName()),
+        CorrectionNameSpec(NNS),
+        CorrectionDecl(Name),
+        EditDistance(distance)  {}
+
+  TypoCorrection(DeclarationName Name, NestedNameSpecifier *NNS=NULL,
+                 unsigned distance=0)
+      : CorrectionName(Name),
+        CorrectionNameSpec(NNS),
+        CorrectionDecl(NULL),
+        EditDistance(distance)  {}
+
+  TypoCorrection()
+      : CorrectionName(), CorrectionNameSpec(NULL), CorrectionDecl(NULL),
+        EditDistance(0) {}
+
+  /// \brief Gets the DeclarationName of the typo correction
+  DeclarationName getCorrection() const { return CorrectionName; }
+  IdentifierInfo* getCorrectionAsIdentifierInfo() const {
+    return CorrectionName.getAsIdentifierInfo();
+  }
+
+  /// \brief Gets the NestedNameSpecifier needed to use the typo correction
+  NestedNameSpecifier* getCorrectionSpecifier() const {
+    return CorrectionNameSpec;
+  }
+  void setCorrectionSpecifier(NestedNameSpecifier* NNS) {
+    CorrectionNameSpec = NNS;
+  }
+
+  /// \brief Gets the "edit distance" of the typo correction from the typo
+  unsigned getEditDistance() const { return EditDistance; }
+
+  /// \brief Gets the pointer to the declaration of the typo correction
+  NamedDecl* getCorrectionDecl() const {
+    return isKeyword() ? NULL : CorrectionDecl;
+  }
+  template <class DeclClass>
+  DeclClass *getCorrectionDeclAs() const {
+    return dyn_cast_or_null<DeclClass>(getCorrectionDecl());
+  }
+  
+  void setCorrectionDecl(NamedDecl *CDecl) {
+    CorrectionDecl = CDecl;
+    if (!CorrectionName)
+      CorrectionName = CDecl->getDeclName();
+  }
+
+  std::string getAsString(const LangOptions &LO) const;
+  std::string getQuoted(const LangOptions &LO) const {
+    return "'" + getAsString(LO) + "'";
+  }
+
+  operator bool() const { return bool(CorrectionName); }
+
+  static inline NamedDecl *KeywordDecl() { return (NamedDecl*)-1; }
+  bool isKeyword() const { return CorrectionDecl == KeywordDecl(); }
+
+  // Returns true if the correction either is a keyword or has a known decl.
+  bool isResolved() const { return CorrectionDecl != NULL; }
+
+private:
+  // Results.
+  DeclarationName CorrectionName;
+  NestedNameSpecifier *CorrectionNameSpec;
+  NamedDecl *CorrectionDecl;
+  unsigned EditDistance;
+};
+
+}
+
+#endif

clang/include/clang/Serialization/ASTBitCodes.h

@@ -371,7 +371,11 @@ namespace clang {
 
       /// \brief Record code for the table of offsets into the block
       /// of file source-location information.
-      FILE_SOURCE_LOCATION_OFFSETS = 45
+      FILE_SOURCE_LOCATION_OFFSETS = 45,
+      
+      /// \brief Record code for the set of known namespaces, which are used
+      /// for typo correction.
+      KNOWN_NAMESPACES = 46
 
     };
 

clang/include/clang/Serialization/ASTReader.h

@@ -625,6 +625,9 @@ class ASTReader
   /// \brief The OpenCL extension settings.
   llvm::SmallVector<uint64_t, 1> OpenCLExtensions;
 
+  /// \brief A list of the namespaces we've seen.
+  llvm::SmallVector<uint64_t, 4> KnownNamespaces;
+
   //@}
 
   /// \brief Diagnostic IDs and their mappings that the user changed.
@@ -1125,6 +1128,11 @@ class ASTReader
   virtual std::pair<ObjCMethodList, ObjCMethodList>
     ReadMethodPool(Selector Sel);
 
+  /// \brief Load the set of namespaces that are known to the external source,
+  /// which will be used during typo correction.
+  virtual void ReadKnownNamespaces(
+                           llvm::SmallVectorImpl<NamespaceDecl *> &Namespaces);
+
   /// \brief Load a selector from disk, registering its ID if it exists.
   void LoadSelector(Selector Sel);
 

clang/lib/Sema/Sema.cpp

@@ -154,6 +154,8 @@ Sema::Sema(Preprocessor &pp, ASTContext &ctxt, ASTConsumer &consumer,
     AnalysisWarnings(*this)
 {
   TUScope = 0;
+  LoadedExternalKnownNamespaces = false;
+  
   if (getLangOptions().CPlusPlus)
     FieldCollector.reset(new CXXFieldCollector());
 
@@ -778,6 +780,10 @@ ExternalSemaSource::ReadMethodPool(Selector Sel) {
   return std::pair<ObjCMethodList, ObjCMethodList>();
 }
 
+void ExternalSemaSource::ReadKnownNamespaces(
+                           llvm::SmallVectorImpl<NamespaceDecl *> &Namespaces) {  
+}
+
 void PrettyDeclStackTraceEntry::print(llvm::raw_ostream &OS) const {
   SourceLocation Loc = this->Loc;
   if (!Loc.isValid() && TheDecl) Loc = TheDecl->getLocation();

clang/lib/Sema/SemaCXXScopeSpec.cpp

@@ -464,26 +464,29 @@ bool Sema::BuildCXXNestedNameSpecifier(Scope *S,
     // We haven't found anything, and we're not recovering from a
     // different kind of error, so look for typos.
     DeclarationName Name = Found.getLookupName();
-    if (CorrectTypo(Found, S, &SS, LookupCtx, EnteringContext,  
-                    CTC_NoKeywords) &&
-        Found.isSingleResult() &&
-        isAcceptableNestedNameSpecifier(Found.getAsSingle<NamedDecl>())) {
+    TypoCorrection Corrected;
+    Found.clear();
+    if ((Corrected = CorrectTypo(Found.getLookupNameInfo(),
+                                 Found.getLookupKind(), S, &SS, LookupCtx,
+                                 EnteringContext, CTC_NoKeywords)) &&
+        isAcceptableNestedNameSpecifier(Corrected.getCorrectionDecl())) {
+      std::string CorrectedStr(Corrected.getAsString(getLangOptions()));
+      std::string CorrectedQuotedStr(Corrected.getQuoted(getLangOptions()));
       if (LookupCtx)
         Diag(Found.getNameLoc(), diag::err_no_member_suggest)
-          << Name << LookupCtx << Found.getLookupName() << SS.getRange()
-          << FixItHint::CreateReplacement(Found.getNameLoc(),
-                                          Found.getLookupName().getAsString());
+          << Name << LookupCtx << CorrectedQuotedStr << SS.getRange()
+          << FixItHint::CreateReplacement(Found.getNameLoc(), CorrectedStr);
       else
         Diag(Found.getNameLoc(), diag::err_undeclared_var_use_suggest)
-          << Name << Found.getLookupName()
-          << FixItHint::CreateReplacement(Found.getNameLoc(),
-                                          Found.getLookupName().getAsString());
+          << Name << CorrectedQuotedStr
+          << FixItHint::CreateReplacement(Found.getNameLoc(), CorrectedStr);
 
-      if (NamedDecl *ND = Found.getAsSingle<NamedDecl>())
-        Diag(ND->getLocation(), diag::note_previous_decl)
-          << ND->getDeclName();
+      if (NamedDecl *ND = Corrected.getCorrectionDecl()) {
+        Diag(ND->getLocation(), diag::note_previous_decl) << CorrectedQuotedStr;
+        Found.addDecl(ND);
+      }
+      Found.setLookupName(Corrected.getCorrection());
     } else {
-      Found.clear();
       Found.setLookupName(&Identifier);
     }
   }