[clang][InstallAPI] Add input file support to library (#81701)

This patch adds support for expected InstallAPI inputs. InstallAPI
accepts a well defined filelist of headers and how those headers
represent a single library.

InstallAPI captures header files to determine linkable symbols to then
compare against what was compiled in a binary dylib and generate TBD
files.

Author: cyndyishida

clang/include/clang/InstallAPI/FileList.h

@@ -0,0 +1,42 @@
+//===- InstallAPI/FileList.h ------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// The JSON file list parser is used to communicate input to InstallAPI.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_INSTALLAPI_FILELIST_H
+#define LLVM_CLANG_INSTALLAPI_FILELIST_H
+
+#include "clang/Basic/Diagnostic.h"
+#include "clang/Basic/FileManager.h"
+#include "clang/InstallAPI/HeaderFile.h"
+#include "llvm/Support/Error.h"
+#include "llvm/Support/MemoryBuffer.h"
+
+namespace clang {
+namespace installapi {
+
+class FileListReader {
+public:
+  /// Decode JSON input and append header input into destination container.
+  /// Headers are loaded in the order they appear in the JSON input.
+  ///
+  /// \param InputBuffer JSON input data.
+  /// \param Destination Container to load headers into.
+  static llvm::Error
+  loadHeaders(std::unique_ptr<llvm::MemoryBuffer> InputBuffer,
+              HeaderSeq &Destination);
+
+  FileListReader() = delete;
+};
+
+} // namespace installapi
+} // namespace clang
+
+#endif // LLVM_CLANG_INSTALLAPI_FILELIST_H

clang/include/clang/InstallAPI/HeaderFile.h

@@ -0,0 +1,72 @@
+//===- InstallAPI/HeaderFile.h ----------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// Representations of a library's headers for InstallAPI.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_INSTALLAPI_HEADERFILE_H
+#define LLVM_CLANG_INSTALLAPI_HEADERFILE_H
+
+#include "clang/Basic/LangStandard.h"
+#include "llvm/ADT/StringRef.h"
+#include "llvm/Support/Regex.h"
+#include <optional>
+#include <string>
+
+namespace clang::installapi {
+enum class HeaderType {
+  /// Represents declarations accessible to all clients.
+  Public,
+  /// Represents declarations accessible to a disclosed set of clients.
+  Private,
+  /// Represents declarations only accessible as implementation details to the
+  /// input library.
+  Project,
+};
+
+class HeaderFile {
+  /// Full input path to header.
+  std::string FullPath;
+  /// Access level of header.
+  HeaderType Type;
+  /// Expected way header will be included by clients.
+  std::string IncludeName;
+  /// Supported language mode for header.
+  std::optional<clang::Language> Language;
+
+public:
+  HeaderFile(StringRef FullPath, HeaderType Type,
+             StringRef IncludeName = StringRef(),
+             std::optional<clang::Language> Language = std::nullopt)
+      : FullPath(FullPath), Type(Type), IncludeName(IncludeName),
+        Language(Language) {}
+
+  static llvm::Regex getFrameworkIncludeRule();
+
+  bool operator==(const HeaderFile &Other) const {
+    return std::tie(Type, FullPath, IncludeName, Language) ==
+           std::tie(Other.Type, Other.FullPath, Other.IncludeName,
+                    Other.Language);
+  }
+};
+
+/// Assemble expected way header will be included by clients.
+/// As in what maps inside the brackets of `#include <IncludeName.h>`
+/// For example,
+/// "/System/Library/Frameworks/Foo.framework/Headers/Foo.h" returns
+/// "Foo/Foo.h"
+///
+/// \param FullPath Path to the header file which includes the library
+/// structure.
+std::optional<std::string> createIncludeHeaderName(const StringRef FullPath);
+using HeaderSeq = std::vector<HeaderFile>;
+
+} // namespace clang::installapi
+
+#endif // LLVM_CLANG_INSTALLAPI_HEADERFILE_H

clang/lib/ExtractAPI/CMakeLists.txt

@@ -16,5 +16,6 @@ add_clang_library(clangExtractAPI
   clangBasic
   clangFrontend
   clangIndex
+  clangInstallAPI
   clangLex
   )

clang/lib/ExtractAPI/ExtractAPIConsumer.cpp

@@ -30,6 +30,7 @@
 #include "clang/Frontend/CompilerInstance.h"
 #include "clang/Frontend/FrontendOptions.h"
 #include "clang/Frontend/MultiplexConsumer.h"
+#include "clang/InstallAPI/HeaderFile.h"
 #include "clang/Lex/MacroInfo.h"
 #include "clang/Lex/PPCallbacks.h"
 #include "clang/Lex/Preprocessor.h"
@@ -61,9 +62,6 @@ std::optional<std::string> getRelativeIncludeName(const CompilerInstance &CI,
          "CompilerInstance does not have a FileNamager!");
 
   using namespace llvm::sys;
-  // Matches framework include patterns
-  const llvm::Regex Rule("/(.+)\\.framework/(.+)?Headers/(.+)");
-
   const auto &FS = CI.getVirtualFileSystem();
 
   SmallString<128> FilePath(File.begin(), File.end());
@@ -147,7 +145,8 @@ std::optional<std::string> getRelativeIncludeName(const CompilerInstance &CI,
       // include name `<Framework/Header.h>`
       if (Entry.IsFramework) {
         SmallVector<StringRef, 4> Matches;
-        Rule.match(File, &Matches);
+        clang::installapi::HeaderFile::getFrameworkIncludeRule().match(
+            File, &Matches);
         // Returned matches are always in stable order.
         if (Matches.size() != 4)
           return std::nullopt;

clang/lib/InstallAPI/CMakeLists.txt

@@ -5,6 +5,8 @@ set(LLVM_LINK_COMPONENTS
 
 add_clang_library(clangInstallAPI
   Context.cpp
+  FileList.cpp
+  HeaderFile.cpp
 
   LINK_LIBS
   clangAST

clang/lib/InstallAPI/FileList.cpp

@@ -0,0 +1,184 @@
+//===- FileList.cpp ---------------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "clang/InstallAPI/FileList.h"
+#include "clang/Basic/DiagnosticFrontend.h"
+#include "clang/InstallAPI/FileList.h"
+#include "llvm/ADT/StringSwitch.h"
+#include "llvm/Support/Error.h"
+#include "llvm/Support/JSON.h"
+#include "llvm/TextAPI/TextAPIError.h"
+#include <optional>
+
+// clang-format off
+/*
+InstallAPI JSON Input Format specification.
+
+{
+  "headers" : [                              # Required: Key must exist.
+    {                                        # Optional: May contain 0 or more header inputs.
+      "path" : "/usr/include/mach-o/dlfn.h", # Required: Path should point to destination 
+                                             #           location where applicable.
+      "type" : "public",                     # Required: Maps to HeaderType for header.
+      "language": "c++"                      # Optional: Language mode for header.
+    }
+  ],
+  "version" : "3"                            # Required: Version 3 supports language mode 
+                                                         & project header input.
+}
+*/
+// clang-format on
+
+using namespace llvm;
+using namespace llvm::json;
+using namespace llvm::MachO;
+using namespace clang::installapi;
+
+namespace {
+class Implementation {
+private:
+  Expected<StringRef> parseString(const Object *Obj, StringRef Key,
+                                  StringRef Error);
+  Expected<StringRef> parsePath(const Object *Obj);
+  Expected<HeaderType> parseType(const Object *Obj);
+  std::optional<clang::Language> parseLanguage(const Object *Obj);
+  Error parseHeaders(Array &Headers);
+
+public:
+  std::unique_ptr<MemoryBuffer> InputBuffer;
+  unsigned Version;
+  HeaderSeq HeaderList;
+
+  Error parse(StringRef Input);
+};
+
+Expected<StringRef>
+Implementation::parseString(const Object *Obj, StringRef Key, StringRef Error) {
+  auto Str = Obj->getString(Key);
+  if (!Str)
+    return make_error<StringError>(Error, inconvertibleErrorCode());
+  return *Str;
+}
+
+Expected<HeaderType> Implementation::parseType(const Object *Obj) {
+  auto TypeStr =
+      parseString(Obj, "type", "required field 'type' not specified");
+  if (!TypeStr)
+    return TypeStr.takeError();
+
+  if (*TypeStr == "public")
+    return HeaderType::Public;
+  else if (*TypeStr == "private")
+    return HeaderType::Private;
+  else if (*TypeStr == "project" && Version >= 2)
+    return HeaderType::Project;
+
+  return make_error<TextAPIError>(TextAPIErrorCode::InvalidInputFormat,
+                                  "unsupported header type");
+}
+
+Expected<StringRef> Implementation::parsePath(const Object *Obj) {
+  auto Path = parseString(Obj, "path", "required field 'path' not specified");
+  if (!Path)
+    return Path.takeError();
+
+  return *Path;
+}
+
+std::optional<clang::Language>
+Implementation::parseLanguage(const Object *Obj) {
+  auto Language = Obj->getString("language");
+  if (!Language)
+    return std::nullopt;
+
+  return StringSwitch<clang::Language>(*Language)
+      .Case("c", clang::Language::C)
+      .Case("c++", clang::Language::CXX)
+      .Case("objective-c", clang::Language::ObjC)
+      .Case("objective-c++", clang::Language::ObjCXX)
+      .Default(clang::Language::Unknown);
+}
+
+Error Implementation::parseHeaders(Array &Headers) {
+  for (const auto &H : Headers) {
+    auto *Obj = H.getAsObject();
+    if (!Obj)
+      return make_error<StringError>("expect a JSON object",
+                                     inconvertibleErrorCode());
+    auto Type = parseType(Obj);
+    if (!Type)
+      return Type.takeError();
+    auto Path = parsePath(Obj);
+    if (!Path)
+      return Path.takeError();
+    auto Language = parseLanguage(Obj);
+
+    StringRef PathStr = *Path;
+    if (*Type == HeaderType::Project) {
+      HeaderList.emplace_back(
+          HeaderFile{PathStr, *Type, /*IncludeName=*/"", Language});
+      continue;
+    }
+    auto IncludeName = createIncludeHeaderName(PathStr);
+    HeaderList.emplace_back(PathStr, *Type,
+                            IncludeName.has_value() ? IncludeName.value() : "",
+                            Language);
+  }
+
+  return Error::success();
+}
+
+Error Implementation::parse(StringRef Input) {
+  auto Val = json::parse(Input);
+  if (!Val)
+    return Val.takeError();
+
+  auto *Root = Val->getAsObject();
+  if (!Root)
+    return make_error<StringError>("not a JSON object",
+                                   inconvertibleErrorCode());
+
+  auto VersionStr = Root->getString("version");
+  if (!VersionStr)
+    return make_error<TextAPIError>(TextAPIErrorCode::InvalidInputFormat,
+                                    "required field 'version' not specified");
+  if (VersionStr->getAsInteger(10, Version))
+    return make_error<TextAPIError>(TextAPIErrorCode::InvalidInputFormat,
+                                    "invalid version number");
+
+  if (Version < 1 || Version > 3)
+    return make_error<TextAPIError>(TextAPIErrorCode::InvalidInputFormat,
+                                    "unsupported version");
+
+  // Not specifying any header files should be atypical, but valid.
+  auto Headers = Root->getArray("headers");
+  if (!Headers)
+    return Error::success();
+
+  Error Err = parseHeaders(*Headers);
+  if (Err)
+    return Err;
+
+  return Error::success();
+}
+} // namespace
+
+llvm::Error
+FileListReader::loadHeaders(std::unique_ptr<MemoryBuffer> InputBuffer,
+                            HeaderSeq &Destination) {
+  Implementation Impl;
+  Impl.InputBuffer = std::move(InputBuffer);
+
+  if (llvm::Error Err = Impl.parse(Impl.InputBuffer->getBuffer()))
+    return Err;
+
+  Destination.reserve(Destination.size() + Impl.HeaderList.size());
+  llvm::move(Impl.HeaderList, std::back_inserter(Destination));
+
+  return Error::success();
+}

clang/lib/InstallAPI/HeaderFile.cpp

@@ -0,0 +1,37 @@
+//===- HeaderFile.cpp ------------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "clang/InstallAPI/HeaderFile.h"
+
+using namespace llvm;
+namespace clang::installapi {
+
+llvm::Regex HeaderFile::getFrameworkIncludeRule() {
+  return llvm::Regex("/(.+)\\.framework/(.+)?Headers/(.+)");
+}
+
+std::optional<std::string> createIncludeHeaderName(const StringRef FullPath) {
+  // Headers in usr(/local)*/include.
+  std::string Pattern = "/include/";
+  auto PathPrefix = FullPath.find(Pattern);
+  if (PathPrefix != StringRef::npos) {
+    PathPrefix += Pattern.size();
+    return FullPath.drop_front(PathPrefix).str();
+  }
+
+  // Framework Headers.
+  SmallVector<StringRef, 4> Matches;
+  HeaderFile::getFrameworkIncludeRule().match(FullPath, &Matches);
+  // Returned matches are always in stable order.
+  if (Matches.size() != 4)
+    return std::nullopt;
+
+  return Matches[1].drop_front(Matches[1].rfind('/') + 1).str() + "/" +
+         Matches[3].str();
+}
+} // namespace clang::installapi

clang/unittests/CMakeLists.txt

@@ -51,5 +51,6 @@ endif()
 add_subdirectory(DirectoryWatcher)
 add_subdirectory(Rename)
 add_subdirectory(Index)
+add_subdirectory(InstallAPI)
 add_subdirectory(Serialization)
 add_subdirectory(Support)