[InstallAPI] Collect frontend attributes & ObjCInterface decls (#83378)

* This patch introduces a container class, for holding records and
attributes only collectible from the clang frontend, which is a subclass
of `llvm::MachO::RecordsSlice`
* This also prunes out collecting declarations from headers that aren't
considered input to installapi.
* Uses these constructs for collecting global objective-c interfaces.

clang/include/clang/InstallAPI/Context.h

@@ -12,12 +12,12 @@
 #include "clang/Basic/Diagnostic.h"
 #include "clang/Basic/FileManager.h"
 #include "clang/InstallAPI/HeaderFile.h"
+#include "llvm/ADT/DenseMap.h"
 #include "llvm/TextAPI/InterfaceFile.h"
-#include "llvm/TextAPI/RecordVisitor.h"
-#include "llvm/TextAPI/RecordsSlice.h"
 
 namespace clang {
 namespace installapi {
+class FrontendRecordsSlice;
 
 /// Struct used for generating validating InstallAPI.
 /// The attributes captured represent all necessary information
@@ -37,7 +37,7 @@ struct InstallAPIContext {
   HeaderType Type = HeaderType::Unknown;
 
   /// Active TargetSlice for symbol record collection.
-  std::shared_ptr<llvm::MachO::RecordsSlice> Slice;
+  std::shared_ptr<FrontendRecordsSlice> Slice;
 
   /// FileManager for all I/O operations.
   FileManager *FM = nullptr;
@@ -50,6 +50,30 @@ struct InstallAPIContext {
 
   /// What encoding to write output as.
   llvm::MachO::FileType FT = llvm::MachO::FileType::TBD_V5;
+
+  /// Populate entries of headers that should be included for TextAPI
+  /// generation.
+  void addKnownHeader(const HeaderFile &H);
+
+  /// Record visited files during frontend actions to determine whether to
+  /// include their declarations for TextAPI generation.
+  ///
+  /// \param FE Header that is being parsed.
+  /// \param PP Preprocesser used for querying how header was imported.
+  /// \return Access level of header if it should be included for TextAPI
+  /// generation.
+  std::optional<HeaderType> findAndRecordFile(const FileEntry *FE,
+                                              const Preprocessor &PP);
+
+private:
+  using HeaderMap = llvm::DenseMap<const FileEntry *, HeaderType>;
+
+  // Collection of parsed header files and their access level. If set to
+  // HeaderType::Unknown, they are not used for TextAPI generation.
+  HeaderMap KnownFiles;
+
+  // Collection of expected header includes and the access level for them.
+  llvm::DenseMap<StringRef, HeaderType> KnownIncludes;
 };
 
 } // namespace installapi

clang/include/clang/InstallAPI/Frontend.h

@@ -14,6 +14,7 @@
 #define LLVM_CLANG_INSTALLAPI_FRONTEND_H
 
 #include "clang/AST/ASTConsumer.h"
+#include "clang/AST/Availability.h"
 #include "clang/Frontend/CompilerInstance.h"
 #include "clang/Frontend/FrontendActions.h"
 #include "clang/InstallAPI/Context.h"
@@ -24,23 +25,78 @@
 namespace clang {
 namespace installapi {
 
+using SymbolFlags = llvm::MachO::SymbolFlags;
+using RecordLinkage = llvm::MachO::RecordLinkage;
+using GlobalRecord = llvm::MachO::GlobalRecord;
+using ObjCInterfaceRecord = llvm::MachO::ObjCInterfaceRecord;
+
+// Represents a collection of frontend records for a library that are tied to a
+// darwin target triple.
+class FrontendRecordsSlice : public llvm::MachO::RecordsSlice {
+public:
+  FrontendRecordsSlice(const llvm::Triple &T)
+      : llvm::MachO::RecordsSlice({T}) {}
+
+  /// Add non-ObjC global record with attributes from AST.
+  ///
+  /// \param Name The name of symbol.
+  /// \param Linkage The linkage of symbol.
+  /// \param GV The kind of global.
+  /// \param Avail The availability information tied to the active target
+  /// triple.
+  /// \param D The pointer to the declaration from traversing AST.
+  /// \param Access The intended access level of symbol.
+  /// \param Flags The flags that describe attributes of the symbol.
+  /// \return The non-owning pointer to added record in slice.
+  GlobalRecord *addGlobal(StringRef Name, RecordLinkage Linkage,
+                          GlobalRecord::Kind GV,
+                          const clang::AvailabilityInfo Avail, const Decl *D,
+                          const HeaderType Access,
+                          SymbolFlags Flags = SymbolFlags::None);
+
+  /// Add ObjC Class record with attributes from AST.
+  ///
+  /// \param Name The name of class, not symbol.
+  /// \param Linkage The linkage of symbol.
+  /// \param Avail The availability information tied to the active target
+  /// triple.
+  /// \param D The pointer to the declaration from traversing AST.
+  /// \param Access The intended access level of symbol.
+  /// \param IsEHType Whether declaration has an exception attribute.
+  /// \return The non-owning pointer to added record in slice.
+  ObjCInterfaceRecord *addObjCInterface(StringRef Name, RecordLinkage Linkage,
+                                        const clang::AvailabilityInfo Avail,
+                                        const Decl *D, HeaderType Access,
+                                        bool IsEHType);
+
+private:
+  /// Frontend information captured about records.
+  struct FrontendAttrs {
+    const AvailabilityInfo Avail;
+    const Decl *D;
+    const HeaderType Access;
+  };
+
+  /// Mapping of records stored in slice to their frontend attributes.
+  llvm::DenseMap<llvm::MachO::Record *, FrontendAttrs> FrontendRecords;
+};
+
 /// Create a buffer that contains all headers to scan
 /// for global symbols with.
-std::unique_ptr<llvm::MemoryBuffer>
-createInputBuffer(const InstallAPIContext &Ctx);
+std::unique_ptr<llvm::MemoryBuffer> createInputBuffer(InstallAPIContext &Ctx);
 
 class InstallAPIAction : public ASTFrontendAction {
 public:
-  explicit InstallAPIAction(llvm::MachO::RecordsSlice &Records)
-      : Records(Records) {}
+  explicit InstallAPIAction(InstallAPIContext &Ctx) : Ctx(Ctx) {}
 
   std::unique_ptr<ASTConsumer> CreateASTConsumer(CompilerInstance &CI,
                                                  StringRef InFile) override {
-    return std::make_unique<InstallAPIVisitor>(CI.getASTContext(), Records);
+    return std::make_unique<InstallAPIVisitor>(
+        CI.getASTContext(), Ctx, CI.getSourceManager(), CI.getPreprocessor());
   }
 
 private:
-  llvm::MachO::RecordsSlice &Records;
+  InstallAPIContext &Ctx;
 };
 } // namespace installapi
 } // namespace clang

clang/include/clang/InstallAPI/Visitor.h

@@ -17,8 +17,8 @@
 #include "clang/AST/RecursiveASTVisitor.h"
 #include "clang/Basic/TargetInfo.h"
 #include "clang/Frontend/FrontendActions.h"
+#include "clang/InstallAPI/Context.h"
 #include "llvm/ADT/Twine.h"
-#include "llvm/TextAPI/RecordsSlice.h"
 
 namespace clang {
 namespace installapi {
@@ -27,20 +27,29 @@ namespace installapi {
 class InstallAPIVisitor final : public ASTConsumer,
                                 public RecursiveASTVisitor<InstallAPIVisitor> {
 public:
-  InstallAPIVisitor(ASTContext &ASTCtx, llvm::MachO::RecordsSlice &Slice)
-      : Slice(Slice),
+  InstallAPIVisitor(ASTContext &ASTCtx, InstallAPIContext &Ctx,
+                    SourceManager &SrcMgr, Preprocessor &PP)
+      : Ctx(Ctx), SrcMgr(SrcMgr), PP(PP),
         MC(ItaniumMangleContext::create(ASTCtx, ASTCtx.getDiagnostics())),
         Layout(ASTCtx.getTargetInfo().getDataLayoutString()) {}
   void HandleTranslationUnit(ASTContext &ASTCtx) override;
 
   /// Collect global variables.
   bool VisitVarDecl(const VarDecl *D);
 
+  /// Collect Objective-C Interface declarations.
+  /// Every Objective-C class has an interface declaration that lists all the
+  /// ivars, properties, and methods of the class.
+  bool VisitObjCInterfaceDecl(const ObjCInterfaceDecl *D);
+
 private:
   std::string getMangledName(const NamedDecl *D) const;
   std::string getBackendMangledName(llvm::Twine Name) const;
+  std::optional<HeaderType> getAccessForDecl(const NamedDecl *D) const;
 
-  llvm::MachO::RecordsSlice &Slice;
+  InstallAPIContext &Ctx;
+  SourceManager &SrcMgr;
+  Preprocessor &PP;
   std::unique_ptr<clang::ItaniumMangleContext> MC;
   StringRef Layout;
 };

clang/lib/InstallAPI/Frontend.cpp

@@ -16,6 +16,73 @@ using namespace llvm::MachO;
 
 namespace clang::installapi {
 
+GlobalRecord *FrontendRecordsSlice::addGlobal(
+    StringRef Name, RecordLinkage Linkage, GlobalRecord::Kind GV,
+    const clang::AvailabilityInfo Avail, const Decl *D, const HeaderType Access,
+    SymbolFlags Flags) {
+
+  auto *GR = llvm::MachO::RecordsSlice::addGlobal(Name, Linkage, GV, Flags);
+  FrontendRecords.insert({GR, FrontendAttrs{Avail, D, Access}});
+  return GR;
+}
+
+ObjCInterfaceRecord *FrontendRecordsSlice::addObjCInterface(
+    StringRef Name, RecordLinkage Linkage, const clang::AvailabilityInfo Avail,
+    const Decl *D, HeaderType Access, bool IsEHType) {
+  ObjCIFSymbolKind SymType =
+      ObjCIFSymbolKind::Class | ObjCIFSymbolKind::MetaClass;
+  if (IsEHType)
+    SymType |= ObjCIFSymbolKind::EHType;
+  auto *ObjCR =
+      llvm::MachO::RecordsSlice::addObjCInterface(Name, Linkage, SymType);
+  FrontendRecords.insert({ObjCR, FrontendAttrs{Avail, D, Access}});
+  return ObjCR;
+}
+
+std::optional<HeaderType>
+InstallAPIContext::findAndRecordFile(const FileEntry *FE,
+                                     const Preprocessor &PP) {
+  if (!FE)
+    return std::nullopt;
+
+  // Check if header has been looked up already and whether it is something
+  // installapi should use.
+  auto It = KnownFiles.find(FE);
+  if (It != KnownFiles.end()) {
+    if (It->second != HeaderType::Unknown)
+      return It->second;
+    else
+      return std::nullopt;
+  }
+
+  // If file was not found, search by how the header was
+  // included. This is primarily to resolve headers found
+  // in a different location than what passed directly as input.
+  StringRef IncludeName = PP.getHeaderSearchInfo().getIncludeNameForHeader(FE);
+  auto BackupIt = KnownIncludes.find(IncludeName.str());
+  if (BackupIt != KnownIncludes.end()) {
+    KnownFiles[FE] = BackupIt->second;
+    return BackupIt->second;
+  }
+
+  // Record that the file was found to avoid future string searches for the
+  // same file.
+  KnownFiles.insert({FE, HeaderType::Unknown});
+  return std::nullopt;
+}
+
+void InstallAPIContext::addKnownHeader(const HeaderFile &H) {
+  auto FE = FM->getFile(H.getPath());
+  if (!FE)
+    return; // File does not exist.
+  KnownFiles[*FE] = H.getType();
+
+  if (!H.useIncludeName())
+    return;
+
+  KnownIncludes[H.getIncludeName()] = H.getType();
+}
+
 static StringRef getFileExtension(clang::Language Lang) {
   switch (Lang) {
   default:
@@ -31,7 +98,7 @@ static StringRef getFileExtension(clang::Language Lang) {
   }
 }
 
-std::unique_ptr<MemoryBuffer> createInputBuffer(const InstallAPIContext &Ctx) {
+std::unique_ptr<MemoryBuffer> createInputBuffer(InstallAPIContext &Ctx) {
   assert(Ctx.Type != HeaderType::Unknown &&
          "unexpected access level for parsing");
   SmallString<4096> Contents;
@@ -47,6 +114,8 @@ std::unique_ptr<MemoryBuffer> createInputBuffer(const InstallAPIContext &Ctx) {
       OS << "<" << H.getIncludeName() << ">";
     else
       OS << "\"" << H.getPath() << "\"";
+
+    Ctx.addKnownHeader(H);
   }
   if (Contents.empty())
     return nullptr;

clang/lib/InstallAPI/Visitor.cpp

@@ -7,8 +7,8 @@
 //===----------------------------------------------------------------------===//
 
 #include "clang/InstallAPI/Visitor.h"
-#include "clang/AST/Availability.h"
 #include "clang/Basic/Linkage.h"
+#include "clang/InstallAPI/Frontend.h"
 #include "llvm/ADT/SmallString.h"
 #include "llvm/ADT/StringRef.h"
 #include "llvm/IR/DataLayout.h"
@@ -62,7 +62,66 @@ std::string InstallAPIVisitor::getBackendMangledName(Twine Name) const {
   return std::string(FinalName);
 }
 
-/// Collect all global variables.
+std::optional<HeaderType>
+InstallAPIVisitor::getAccessForDecl(const NamedDecl *D) const {
+  SourceLocation Loc = D->getLocation();
+  if (Loc.isInvalid())
+    return std::nullopt;
+
+  // If the loc refers to a macro expansion, InstallAPI needs to first get the
+  // file location of the expansion.
+  auto FileLoc = SrcMgr.getFileLoc(Loc);
+  FileID ID = SrcMgr.getFileID(FileLoc);
+  if (ID.isInvalid())
+    return std::nullopt;
+
+  const FileEntry *FE = SrcMgr.getFileEntryForID(ID);
+  if (!FE)
+    return std::nullopt;
+
+  auto Header = Ctx.findAndRecordFile(FE, PP);
+  if (!Header.has_value())
+    return std::nullopt;
+
+  HeaderType Access = Header.value();
+  assert(Access != HeaderType::Unknown && "unexpected access level for global");
+  return Access;
+}
+
+/// Check if the interface itself or any of its super classes have an
+/// exception attribute. InstallAPI needs to export an additional symbol
+/// ("OBJC_EHTYPE_$CLASS_NAME") if any of the classes have the exception
+/// attribute.
+static bool hasObjCExceptionAttribute(const ObjCInterfaceDecl *D) {
+  for (; D != nullptr; D = D->getSuperClass())
+    if (D->hasAttr<ObjCExceptionAttr>())
+      return true;
+
+  return false;
+}
+
+bool InstallAPIVisitor::VisitObjCInterfaceDecl(const ObjCInterfaceDecl *D) {
+  // Skip forward declaration for classes (@class)
+  if (!D->isThisDeclarationADefinition())
+    return true;
+
+  // Skip over declarations that access could not be collected for.
+  auto Access = getAccessForDecl(D);
+  if (!Access)
+    return true;
+
+  StringRef Name = D->getObjCRuntimeNameAsString();
+  const RecordLinkage Linkage =
+      isExported(D) ? RecordLinkage::Exported : RecordLinkage::Internal;
+  const AvailabilityInfo Avail = AvailabilityInfo::createFromDecl(D);
+  const bool IsEHType =
+      (!D->getASTContext().getLangOpts().ObjCRuntime.isFragile() &&
+       hasObjCExceptionAttribute(D));
+
+  Ctx.Slice->addObjCInterface(Name, Linkage, Avail, D, *Access, IsEHType);
+  return true;
+}
+
 bool InstallAPIVisitor::VisitVarDecl(const VarDecl *D) {
   // Skip function parameters.
   if (isa<ParmVarDecl>(D))
@@ -81,13 +140,18 @@ bool InstallAPIVisitor::VisitVarDecl(const VarDecl *D) {
       D->getTemplateSpecializationKind() == TSK_Undeclared)
     return true;
 
-  // TODO: Capture SourceLocation & Availability for Decls.
+  // Skip over declarations that access could not collected for.
+  auto Access = getAccessForDecl(D);
+  if (!Access)
+    return true;
+
   const RecordLinkage Linkage =
       isExported(D) ? RecordLinkage::Exported : RecordLinkage::Internal;
   const bool WeakDef = D->hasAttr<WeakAttr>();
   const bool ThreadLocal = D->getTLSKind() != VarDecl::TLS_None;
-  Slice.addGlobal(getMangledName(D), Linkage, GlobalRecord::Kind::Variable,
-                  getFlags(WeakDef, ThreadLocal));
+  const AvailabilityInfo Avail = AvailabilityInfo::createFromDecl(D);
+  Ctx.Slice->addGlobal(getMangledName(D), Linkage, GlobalRecord::Kind::Variable,
+                       Avail, D, *Access, getFlags(WeakDef, ThreadLocal));
   return true;
 }