[modules] Separately track whether an identifier's preprocessor information and

name lookup information have changed since deserialization. For a C++ modules
build, we do not need to re-emit the identifier into the serialized identifier
table if only the name lookup information has changed (and in all cases, we
don't need to re-emit the macro information if only the name lookup information
has changed).

llvm-svn: 259901

clang/include/clang/Basic/IdentifierTable.h

@@ -62,6 +62,9 @@ class IdentifierInfo {
                                    // partially) from an AST file.
   bool ChangedAfterLoad       : 1; // True if identifier has changed from the
                                    // definition loaded from an AST file.
+  bool FEChangedAfterLoad     : 1; // True if identifier's frontend information
+                                   // has changed from the definition loaded
+                                   // from an AST file.
   bool RevertedTokenID        : 1; // True if revertTokenIDToIdentifier was
                                    // called.
   bool OutOfDate              : 1; // True if there may be additional
@@ -69,7 +72,7 @@ class IdentifierInfo {
                                    // stored externally.
   bool IsModulesImport        : 1; // True if this is the 'import' contextual
                                    // keyword.
-  // 30 bit left in 64-bit word.
+  // 29 bit left in 64-bit word.
 
   void *FETokenInfo;               // Managed by the language front-end.
   llvm::StringMapEntry<IdentifierInfo*> *Entry;
@@ -303,6 +306,18 @@ public:
     ChangedAfterLoad = true;
   }
 
+  /// \brief Determine whether the frontend token information for this
+  /// identifier has changed since it was loaded from an AST file.
+  bool hasFETokenInfoChangedSinceDeserialization() const {
+    return FEChangedAfterLoad;
+  }
+  
+  /// \brief Note that the frontend token information for this identifier has
+  /// changed since it was loaded from an AST file.
+  void setFETokenInfoChangedSinceDeserialization() {
+    FEChangedAfterLoad = true;
+  }
+
   /// \brief Determine whether the information for this identifier is out of
   /// date with respect to the external source.
   bool isOutOfDate() const { return OutOfDate; }

clang/lib/Basic/IdentifierTable.cpp

@@ -42,6 +42,7 @@ IdentifierInfo::IdentifierInfo() {
   NeedsHandleIdentifier = false;
   IsFromAST = false;
   ChangedAfterLoad = false;
+  FEChangedAfterLoad = false;
   RevertedTokenID = false;
   OutOfDate = false;
   IsModulesImport = false;

clang/lib/Sema/IdentifierResolver.cpp

@@ -381,7 +381,7 @@ void IdentifierResolver::updatingIdentifier(IdentifierInfo &II) {
     PP.getExternalSource()->updateOutOfDateIdentifier(II);
   
   if (II.isFromAST())
-    II.setChangedSinceDeserialization();
+    II.setFETokenInfoChangedSinceDeserialization();
 }
 
 //===----------------------------------------------------------------------===//

clang/lib/Serialization/ASTWriter.cpp

@@ -3162,6 +3162,8 @@ public:
         NeedDecls(!IsModule || !Writer.getLangOpts().CPlusPlus),
         InterestingIdentifierOffsets(InterestingIdentifierOffsets) {}
 
+  bool needDecls() const { return NeedDecls; }
+
   static hash_value_type ComputeHash(const IdentifierInfo* II) {
     return llvm::HashString(II->getName());
   }
@@ -3307,7 +3309,9 @@ void ASTWriter::WriteIdentifierTable(Preprocessor &PP,
       auto *II = const_cast<IdentifierInfo *>(IdentIDPair.first);
       IdentID ID = IdentIDPair.second;
       assert(II && "NULL identifier in identifier table");
-      if (!Chain || !II->isFromAST() || II->hasChangedSinceDeserialization())
+      if (!Chain || !II->isFromAST() || II->hasChangedSinceDeserialization() ||
+          (Trait.needDecls() &&
+           II->hasFETokenInfoChangedSinceDeserialization()))
         Generator.insert(II, ID, Trait);
     }
 

clang/test/Modules/minimal-identifier-tables.cpp

@@ -0,0 +1,10 @@
+// RUN: rm -rf %t
+// RUN: mkdir %t
+// RUN: echo 'extern int some_long_variable_name;' > %t/x.h
+// RUN: echo 'extern int some_long_variable_name;' > %t/y.h
+// RUN: echo 'module X { header "x.h" } module Y { header "y.h" }' > %t/map
+// RUN: %clang_cc1 -fmodules -x c++ -fmodule-name=X %t/map -emit-module -o %t/x.pcm
+// RUN: %clang_cc1 -fmodules -x c++ -fmodule-name=Y %t/map -fmodule-file=%t/x.pcm -emit-module -o %t/y.pcm
+// RUN: cat %t/y.pcm | FileCheck %s
+//
+// CHECK-NOT: some_long_variable_name