[clang][serialization] Enable ASTWriter to work with Preprocessor only (llvm#115237)

This PR builds on top of
llvm#115235 and makes it possible
to call `ASTWriter::WriteAST()` with `Preprocessor` only instead of full
`Sema` object. So far, there are no clients that leverage the new
capability - that will come in a follow-up commit.

(cherry picked from commit a2f9d1d)

Author: jansvoboda11

clang/include/clang/AST/ASTContext.h

@@ -476,14 +476,6 @@ class ASTContext : public RefCountedBase<ASTContext> {
   llvm::StringMap<const Module *> PrimaryModuleNameMap;
   llvm::DenseMap<const Module *, const Module *> SameModuleLookupSet;
 
-  /// The include tree that is being built, if any.
-  /// See \c FrontendOptions::CASIncludeTreeID.
-  std::optional<std::string> CASIncludeTreeID;
-
-  /// The cas-fs tree that is being built, if any.
-  /// See \c FileSystemOptions::CASFileSystemRootID.
-  std::optional<std::string> CASFileSystemRootID;
-
   static constexpr unsigned ConstantArrayTypesLog2InitSize = 8;
   static constexpr unsigned GeneralTypesLog2InitSize = 9;
   static constexpr unsigned FunctionProtoTypesLog2InitSize = 12;
@@ -1123,20 +1115,6 @@ class ASTContext : public RefCountedBase<ASTContext> {
   /// a module fragment or a module unit but not a C++20 module.
   bool isInSameModule(const Module *M1, const Module *M2);
 
-  std::optional<std::string> getCASIncludeTreeID() const {
-    return CASIncludeTreeID;
-  }
-  void setCASIncludeTreeID(std::string ID) {
-    CASIncludeTreeID = std::move(ID);
-  }
-
-  std::optional<std::string> getCASFileSystemRootID() const {
-    return CASFileSystemRootID;
-  }
-  void setCASFileSystemRootID(std::string ID) {
-    CASFileSystemRootID = std::move(ID);
-  }
-
   TranslationUnitDecl *getTranslationUnitDecl() const {
     return TUDecl->getMostRecentDecl();
   }

clang/include/clang/Lex/Preprocessor.h

@@ -2428,6 +2428,28 @@ class Preprocessor {
            !IsAtImport;
   }
 
+private:
+  /// The include tree that is being built, if any.
+  /// See \c FrontendOptions::CASIncludeTreeID.
+  std::optional<std::string> CASIncludeTreeID;
+
+  /// The cas-fs tree that is being built, if any.
+  /// See \c FileSystemOptions::CASFileSystemRootID.
+  std::optional<std::string> CASFileSystemRootID;
+
+public:
+  std::optional<std::string> getCASIncludeTreeID() const {
+    return CASIncludeTreeID;
+  }
+  void setCASIncludeTreeID(std::string ID) { CASIncludeTreeID = std::move(ID); }
+
+  std::optional<std::string> getCASFileSystemRootID() const {
+    return CASFileSystemRootID;
+  }
+  void setCASFileSystemRootID(std::string ID) {
+    CASFileSystemRootID = std::move(ID);
+  }
+
   /// Allocate a new MacroInfo object with the provided SourceLocation.
   MacroInfo *AllocateMacroInfo(SourceLocation L);
 

clang/include/clang/Serialization/ASTWriter.h

@@ -537,8 +537,7 @@ class ASTWriter : public ASTDeserializationListener,
   void WriteSubStmt(ASTContext &Context, Stmt *S);
 
   void WriteBlockInfoBlock();
-  void WriteControlBlock(ASTContext &Context, Preprocessor &PP,
-                         StringRef isysroot);
+  void WriteControlBlock(Preprocessor &PP, StringRef isysroot);
 
   /// Write out the signature and diagnostic options, and return the signature.
   void writeUnhashedControlBlock(Preprocessor &PP);
@@ -554,7 +553,7 @@ class ASTWriter : public ASTDeserializationListener,
   void WriteHeaderSearch(const HeaderSearch &HS);
   void WritePreprocessorDetail(PreprocessingRecord &PPRec,
                                uint64_t MacroOffsetsBase);
-  void WriteSubmodules(Module *WritingModule, ASTContext &Context);
+  void WriteSubmodules(Module *WritingModule, ASTContext *Context);
 
   void WritePragmaDiagnosticMappings(const DiagnosticsEngine &Diag,
                                      bool isModule);
@@ -575,7 +574,7 @@ class ASTWriter : public ASTDeserializationListener,
   void WriteComments(ASTContext &Context);
   void WriteSelectors(Sema &SemaRef);
   void WriteReferencedSelectorsPool(Sema &SemaRef);
-  void WriteIdentifierTable(Preprocessor &PP, IdentifierResolver &IdResolver,
+  void WriteIdentifierTable(Preprocessor &PP, IdentifierResolver *IdResolver,
                             bool IsModule);
   void WriteDeclAndTypes(ASTContext &Context);
   void PrepareWritingSpecialDecls(Sema &SemaRef);
@@ -632,7 +631,7 @@ class ASTWriter : public ASTDeserializationListener,
   void WriteDeclAbbrevs();
   void WriteDecl(ASTContext &Context, Decl *D);
 
-  ASTFileSignature WriteASTCore(Sema &SemaRef, StringRef isysroot,
+  ASTFileSignature WriteASTCore(Sema *SemaPtr, StringRef isysroot,
                                 Module *WritingModule);
 
 public:
@@ -652,10 +651,13 @@ class ASTWriter : public ASTDeserializationListener,
   /// include timestamps in the output file.
   time_t getTimestampForOutput(const FileEntry *E) const;
 
-  /// Write a precompiled header for the given semantic analysis.
+  /// Write a precompiled header or a module with the AST produced by the
+  /// \c Sema object, or a dependency scanner module with the preprocessor state
+  /// produced by the \c Preprocessor object.
   ///
-  /// \param SemaRef a reference to the semantic analysis object that processed
-  /// the AST to be written into the precompiled header.
+  /// \param Subject The \c Sema object that processed the AST to be written, or
+  /// in the case of a dependency scanner module the \c Preprocessor that holds
+  /// the state.
   ///
   /// \param WritingModule The module that we are writing. If null, we are
   /// writing a precompiled header.
@@ -666,8 +668,9 @@ class ASTWriter : public ASTDeserializationListener,
   ///
   /// \return the module signature, which eventually will be a hash of
   /// the module but currently is merely a random 32-bit number.
-  ASTFileSignature WriteAST(Sema &SemaRef, StringRef OutputFile,
-                            Module *WritingModule, StringRef isysroot,
+  ASTFileSignature WriteAST(llvm::PointerUnion<Sema *, Preprocessor *> Subject,
+                            StringRef OutputFile, Module *WritingModule,
+                            StringRef isysroot,
                             bool ShouldCacheASTInMemory = false);
 
   /// Emit a token.

clang/lib/Frontend/ASTUnit.cpp

@@ -2358,7 +2358,7 @@ bool ASTUnit::Save(StringRef File) {
 
 static bool serializeUnit(ASTWriter &Writer, SmallVectorImpl<char> &Buffer,
                           Sema &S, raw_ostream &OS) {
-  Writer.WriteAST(S, std::string(), nullptr, "");
+  Writer.WriteAST(&S, std::string(), nullptr, "");
 
   // Write the generated bitstream to "Out".
   if (!Buffer.empty())

clang/lib/Frontend/FrontendAction.cpp

@@ -1208,7 +1208,7 @@ bool FrontendAction::BeginSourceFile(CompilerInstance &CI,
     }
 
     if (!CI.getFrontendOpts().CASIncludeTreeID.empty())
-      CI.getASTContext().setCASIncludeTreeID(
+      CI.getPreprocessor().setCASIncludeTreeID(
           CI.getFrontendOpts().CASIncludeTreeID);
 
     CI.setASTConsumer(std::move(Consumer));