[modules] Add ability to specify module name to module file mapping (…

…reapply)

Extend the -fmodule-file option to support the [<name>=]<file> value format.
If the name is omitted, then the old semantics is preserved (the module file
is loaded whether needed or not). If the name is specified, then the mapping
is treated as just another prebuilt module search mechanism, similar to
-fprebuilt-module-path, and the module file is only loaded if actually used
(e.g., via import). With one exception: this mapping also overrides module
file references embedded in other modules (which can be useful if module files
are moved/renamed as often happens during remote compilation).

This override semantics requires some extra work: we now store the module name
in addition to the file name in the serialized AST representation.

Reviewed By: rsmith

Differential Revision: https://reviews.llvm.org/D35020

llvm-svn: 312220

clang/docs/Modules.rst

@@ -213,8 +213,14 @@ Command-line parameters
 ``-fno-implicit-modules``
   All modules used by the build must be specified with ``-fmodule-file``.
 
-``-fmodule-file=<file>``
-  Load the given precompiled module file.
+``-fmodule-file=[<name>=]<file>``
+  Specify the mapping of module names to precompiled module files. If the
+  name is omitted, then the module file is loaded whether actually required
+  or not. If the name is specified, then the mapping is treated as another
+  prebuilt module search mechanism (in addition to ``-fprebuilt-module-path``)
+  and the module is only loaded if required. Note that in this case the
+  specified file also overrides this module's paths that might be embedded
+  in other precompiled module files.
 
 ``-fprebuilt-module-path=<directory>``
   Specify the path to the prebuilt modules. If specified, we will look for modules in this directory for a given top-level module name. We don't need a module map for loading prebuilt modules in this directory and the compiler will not try to rebuild these modules. This can be specified multiple times.
@@ -945,4 +951,3 @@ PCHInternals_
 .. [#] The preprocessing context in which the modules are parsed is actually dependent on the command-line options provided to the compiler, including the language dialect and any ``-D`` options. However, the compiled modules for different command-line options are kept distinct, and any preprocessor directives that occur within the translation unit are ignored. See the section on the `Configuration macros declaration`_ for more information.
 
 .. _PCHInternals: PCHInternals.html
-

clang/include/clang/Driver/Options.td

@@ -1125,8 +1125,8 @@ def fmodule_map_file : Joined<["-"], "fmodule-map-file=">,
   Group<f_Group>, Flags<[DriverOption,CC1Option]>, MetaVarName<"<file>">,
   HelpText<"Load this module map file">;
 def fmodule_file : Joined<["-"], "fmodule-file=">,
-  Group<f_Group>, Flags<[DriverOption,CC1Option]>,
-  HelpText<"Load this precompiled module file">, MetaVarName<"<file>">;
+  Group<i_Group>, Flags<[DriverOption,CC1Option]>, MetaVarName<"[<name>=]<file>">,
+  HelpText<"Specify the mapping of module name to precompiled module file, or load a module file if name is omitted.">;
 def fmodules_ignore_macro : Joined<["-"], "fmodules-ignore-macro=">, Group<f_Group>, Flags<[CC1Option]>,
   HelpText<"Ignore the definition of the given macro when building and loading modules">;
 def fmodules_decluse : Flag <["-"], "fmodules-decluse">, Group<f_Group>,

clang/include/clang/Lex/HeaderSearch.h

@@ -471,29 +471,41 @@ class HeaderSearch {
   /// \brief Get filenames for all registered header maps.
   void getHeaderMapFileNames(SmallVectorImpl<std::string> &Names) const;
 
-  /// \brief Retrieve the name of the module file that should be used to 
-  /// load the given module.
+  /// \brief Retrieve the name of the cached module file that should be used
+  /// to load the given module.
   ///
   /// \param Module The module whose module file name will be returned.
   ///
   /// \returns The name of the module file that corresponds to this module,
   /// or an empty string if this module does not correspond to any module file.
-  std::string getModuleFileName(Module *Module);
+  std::string getCachedModuleFileName(Module *Module);
 
-  /// \brief Retrieve the name of the module file that should be used to 
-  /// load a module with the given name.
+  /// \brief Retrieve the name of the prebuilt module file that should be used
+  /// to load a module with the given name.
+  ///
+  /// \param ModuleName The module whose module file name will be returned.
+  ///
+  /// \param FileMapOnly If true, then only look in the explicit module name
+  //  to file name map and skip the directory search.
+  ///
+  /// \returns The name of the module file that corresponds to this module,
+  /// or an empty string if this module does not correspond to any module file.
+  std::string getPrebuiltModuleFileName(StringRef ModuleName,
+                                        bool FileMapOnly = false);
+
+
+  /// \brief Retrieve the name of the (to-be-)cached module file that should
+  /// be used to load a module with the given name.
   ///
   /// \param ModuleName The module whose module file name will be returned.
   ///
   /// \param ModuleMapPath A path that when combined with \c ModuleName
   /// uniquely identifies this module. See Module::ModuleMap.
   ///
-  /// \param UsePrebuiltPath Whether we should use the prebuilt module path.
-  ///
   /// \returns The name of the module file that corresponds to this module,
   /// or an empty string if this module does not correspond to any module file.
-  std::string getModuleFileName(StringRef ModuleName, StringRef ModuleMapPath,
-                                bool UsePrebuiltPath);
+  std::string getCachedModuleFileName(StringRef ModuleName,
+                                      StringRef ModuleMapPath);
 
   /// \brief Lookup a module Search for a module with the given name.
   ///

clang/include/clang/Lex/HeaderSearchOptions.h

@@ -17,6 +17,7 @@
 #include "llvm/ADT/StringRef.h"
 #include <string>
 #include <vector>
+#include <map>
 
 namespace clang {
 
@@ -94,6 +95,9 @@ class HeaderSearchOptions {
   /// \brief The directory used for a user build.
   std::string ModuleUserBuildPath;
 
+  /// \brief The mapping of module names to prebuilt module files.
+  std::map<std::string, std::string> PrebuiltModuleFiles;
+
   /// \brief The directories used to load prebuilt module files.
   std::vector<std::string> PrebuiltModulePaths;
 

clang/include/clang/Serialization/ASTReader.h

@@ -2142,9 +2142,19 @@ class ASTReader
   // \brief Read a string
   static std::string ReadString(const RecordData &Record, unsigned &Idx);
 
+  // \brief Skip a string
+  static void SkipString(const RecordData &Record, unsigned &Idx) {
+    Idx += Record[Idx] + 1;
+  }
+
   // \brief Read a path
   std::string ReadPath(ModuleFile &F, const RecordData &Record, unsigned &Idx);
 
+  // \brief Skip a path
+  static void SkipPath(const RecordData &Record, unsigned &Idx) {
+    SkipString(Record, Idx);
+  }
+
   /// \brief Read a version tuple.
   static VersionTuple ReadVersionTuple(const RecordData &Record, unsigned &Idx);
 

clang/include/clang/Serialization/ModuleManager.h

@@ -27,6 +27,7 @@ class GlobalModuleIndex;
 class MemoryBufferCache;
 class ModuleMap;
 class PCHContainerReader;
+class HeaderSearch;
 
 namespace serialization {
 
@@ -58,6 +59,9 @@ class ModuleManager {
   /// \brief Knows how to unwrap module containers.
   const PCHContainerReader &PCHContainerRdr;
 
+  /// \brief Preprocessor's HeaderSearchInfo containing the module map.
+  const HeaderSearch &HeaderSearchInfo;
+
   /// \brief A lookup of in-memory (virtual file) buffers
   llvm::DenseMap<const FileEntry *, std::unique_ptr<llvm::MemoryBuffer>>
       InMemoryBuffers;
@@ -128,7 +132,8 @@ class ModuleManager {
   typedef std::pair<uint32_t, StringRef> ModuleOffset;
 
   explicit ModuleManager(FileManager &FileMgr, MemoryBufferCache &PCMCache,
-                         const PCHContainerReader &PCHContainerRdr);
+                         const PCHContainerReader &PCHContainerRdr,
+                         const HeaderSearch &HeaderSearchInfo);
   ~ModuleManager();
 
   /// \brief Forward iterator to traverse all loaded modules.
@@ -163,8 +168,11 @@ class ModuleManager {
   /// \brief Returns the module associated with the given index
   ModuleFile &operator[](unsigned Index) const { return *Chain[Index]; }
 
-  /// \brief Returns the module associated with the given name
-  ModuleFile *lookup(StringRef Name) const;
+  /// \brief Returns the module associated with the given file name.
+  ModuleFile *lookupByFileName(StringRef FileName) const;
+
+  /// \brief Returns the module associated with the given module name.
+  ModuleFile *lookupByModuleName(StringRef ModName) const;
 
   /// \brief Returns the module associated with the given module file.
   ModuleFile *lookup(const FileEntry *File) const;

clang/lib/Driver/ToolChains/Clang.cpp

@@ -3651,9 +3651,11 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
 
   if (HaveAnyModules) {
     // -fprebuilt-module-path specifies where to load the prebuilt module files.
-    for (const Arg *A : Args.filtered(options::OPT_fprebuilt_module_path))
+    for (const Arg *A : Args.filtered(options::OPT_fprebuilt_module_path)) {
       CmdArgs.push_back(Args.MakeArgString(
           std::string("-fprebuilt-module-path=") + A->getValue()));
+      A->claim();
+    }
   }
 
   // -fmodule-name specifies the module that is currently being built (or
@@ -3676,7 +3678,10 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
     }
   }
 
-  // -fmodule-file can be used to specify files containing precompiled modules.
+  // The -fmodule-file=<name>=<file> form specifies the mapping of module
+  // names to precompiled module files (the module is loaded only if used).
+  // The -fmodule-file=<file> form can be used to unconditionally load
+  // precompiled module files (whether used or not).
   if (HaveAnyModules)
     Args.AddAllArgs(CmdArgs, options::OPT_fmodule_file);
   else

clang/lib/Frontend/CompilerInstance.cpp

@@ -1641,6 +1641,14 @@ CompilerInstance::loadModule(SourceLocation ImportLoc,
   } else if (ModuleName == getLangOpts().CurrentModule) {
     // This is the module we're building. 
     Module = PP->getHeaderSearchInfo().lookupModule(ModuleName);
+    /// FIXME: perhaps we should (a) look for a module using the module name
+    //  to file map (PrebuiltModuleFiles) and (b) diagnose if still not found?
+    //if (Module == nullptr) {
+    //  getDiagnostics().Report(ModuleNameLoc, diag::err_module_not_found)
+    //    << ModuleName;
+    //  ModuleBuildFailed = true;
+    //  return ModuleLoadResult();
+    //}
     Known = KnownModules.insert(std::make_pair(Path[0].first, Module)).first;
   } else {
     // Search for a module with the given name.
@@ -1662,16 +1670,17 @@ CompilerInstance::loadModule(SourceLocation ImportLoc,
     }
 
     // Try to load the module from the prebuilt module path.
-    if (Source == ModuleNotFound && !HSOpts.PrebuiltModulePaths.empty()) {
-      ModuleFileName = PP->getHeaderSearchInfo().getModuleFileName(
-          ModuleName, "", /*UsePrebuiltPath*/ true);
+    if (Source == ModuleNotFound && (!HSOpts.PrebuiltModuleFiles.empty() ||
+                                     !HSOpts.PrebuiltModulePaths.empty())) {
+      ModuleFileName =
+        PP->getHeaderSearchInfo().getPrebuiltModuleFileName(ModuleName);
       if (!ModuleFileName.empty())
         Source = PrebuiltModulePath;
     }
 
     // Try to load the module from the module cache.
     if (Source == ModuleNotFound && Module) {
-      ModuleFileName = PP->getHeaderSearchInfo().getModuleFileName(Module);
+      ModuleFileName = PP->getHeaderSearchInfo().getCachedModuleFileName(Module);
       Source = ModuleCache;
     }
 

clang/lib/Frontend/CompilerInvocation.cpp

@@ -1009,9 +1009,12 @@ static void ParseDependencyOutputArgs(DependencyOutputOptions &Opts,
   // They won't be discovered by the regular preprocessor, so
   // we let make / ninja to know about this implicit dependency.
   Opts.ExtraDeps = Args.getAllArgValues(OPT_fdepfile_entry);
-  auto ModuleFiles = Args.getAllArgValues(OPT_fmodule_file);
-  Opts.ExtraDeps.insert(Opts.ExtraDeps.end(), ModuleFiles.begin(),
-                        ModuleFiles.end());
+  // Only the -fmodule-file=<file> form.
+  for (const Arg *A : Args.filtered(OPT_fmodule_file)) {
+    StringRef Val = A->getValue();
+    if (Val.find('=') == StringRef::npos)
+      Opts.ExtraDeps.push_back(Val);
+  }
 }
 
 static bool parseShowColorsArgs(const ArgList &Args, bool DefaultColor) {
@@ -1340,7 +1343,12 @@ static InputKind ParseFrontendArgs(FrontendOptions &Opts, ArgList &Args,
   Opts.UseGlobalModuleIndex = !Args.hasArg(OPT_fno_modules_global_index);
   Opts.GenerateGlobalModuleIndex = Opts.UseGlobalModuleIndex;
   Opts.ModuleMapFiles = Args.getAllArgValues(OPT_fmodule_map_file);
-  Opts.ModuleFiles = Args.getAllArgValues(OPT_fmodule_file);
+  // Only the -fmodule-file=<file> form.
+  for (const Arg *A : Args.filtered(OPT_fmodule_file)) {
+    StringRef Val = A->getValue();
+    if (Val.find('=') == StringRef::npos)
+      Opts.ModuleFiles.push_back(Val);
+  }
   Opts.ModulesEmbedFiles = Args.getAllArgValues(OPT_fmodules_embed_file_EQ);
   Opts.ModulesEmbedAllFiles = Args.hasArg(OPT_fmodules_embed_all_files);
   Opts.IncludeTimestamps = !Args.hasArg(OPT_fno_pch_timestamp);
@@ -1544,6 +1552,12 @@ static void ParseHeaderSearchArgs(HeaderSearchOptions &Opts, ArgList &Args,
   Opts.ModuleCachePath = P.str();
 
   Opts.ModuleUserBuildPath = Args.getLastArgValue(OPT_fmodules_user_build_path);
+  // Only the -fmodule-file=<name>=<file> form.
+  for (const Arg *A : Args.filtered(OPT_fmodule_file)) {
+    StringRef Val = A->getValue();
+    if (Val.find('=') != StringRef::npos)
+      Opts.PrebuiltModuleFiles.insert(Val.split('='));
+  }
   for (const Arg *A : Args.filtered(OPT_fprebuilt_module_path))
     Opts.AddPrebuiltModulePath(A->getValue());
   Opts.DisableModuleHash = Args.hasArg(OPT_fdisable_module_hash);

clang/lib/Frontend/FrontendActions.cpp

@@ -185,8 +185,8 @@ GenerateModuleFromModuleMapAction::CreateOutputFile(CompilerInstance &CI,
 
     HeaderSearch &HS = CI.getPreprocessor().getHeaderSearchInfo();
     CI.getFrontendOpts().OutputFile =
-        HS.getModuleFileName(CI.getLangOpts().CurrentModule, ModuleMapFile,
-                             /*UsePrebuiltPath=*/false);
+        HS.getCachedModuleFileName(CI.getLangOpts().CurrentModule,
+                                   ModuleMapFile);
   }
 
   // We use createOutputFile here because this is exposed via libclang, and we

clang/lib/Lex/HeaderSearch.cpp

@@ -128,21 +128,24 @@ void HeaderSearch::getHeaderMapFileNames(
     Names.push_back(HM.first->getName());
 }
 
-std::string HeaderSearch::getModuleFileName(Module *Module) {
+std::string HeaderSearch::getCachedModuleFileName(Module *Module) {
   const FileEntry *ModuleMap =
       getModuleMap().getModuleMapFileForUniquing(Module);
-  return getModuleFileName(Module->Name, ModuleMap->getName(),
-                           /*UsePrebuiltPath*/false);
+  return getCachedModuleFileName(Module->Name, ModuleMap->getName());
 }
 
-std::string HeaderSearch::getModuleFileName(StringRef ModuleName,
-                                            StringRef ModuleMapPath,
-                                            bool UsePrebuiltPath) {
-  if (UsePrebuiltPath) {
-    if (HSOpts->PrebuiltModulePaths.empty())
+std::string HeaderSearch::getPrebuiltModuleFileName(StringRef ModuleName,
+                                                    bool FileMapOnly) {
+  // First check the module name to pcm file map.
+  auto i (HSOpts->PrebuiltModuleFiles.find(ModuleName));
+  if (i != HSOpts->PrebuiltModuleFiles.end())
+    return i->second;
+
+  if (FileMapOnly || HSOpts->PrebuiltModulePaths.empty())
       return std::string();
 
-    // Go though each prebuilt module path and try to find the pcm file.
+  // Then go through each prebuilt module directory and try to find the pcm
+  // file.
     for (const std::string &Dir : HSOpts->PrebuiltModulePaths) {
       SmallString<256> Result(Dir);
       llvm::sys::fs::make_absolute(Result);
@@ -154,6 +157,8 @@ std::string HeaderSearch::getModuleFileName(StringRef ModuleName,
     return std::string();
   }
 
+std::string HeaderSearch::getCachedModuleFileName(StringRef ModuleName,
+                                                  StringRef ModuleMapPath) {
   // If we don't have a module cache path or aren't supposed to use one, we
   // can't do anything.
   if (getModuleCachePath().empty())

clang/lib/Serialization/ASTReader.cpp

@@ -2485,7 +2485,23 @@ ASTReader::ReadControlBlock(ModuleFile &F,
             {{(uint32_t)Record[Idx++], (uint32_t)Record[Idx++],
               (uint32_t)Record[Idx++], (uint32_t)Record[Idx++],
               (uint32_t)Record[Idx++]}}};
-        auto ImportedFile = ReadPath(F, Record, Idx);
+
+        std::string ImportedName = ReadString(Record, Idx);
+        std::string ImportedFile;
+
+        // For prebuilt and explicit modules first consult the file map for
+        // an override. Note that here we don't search prebuilt module
+        // directories, only the explicit name to file mappings. Also, we will
+        // still verify the size/signature making sure it is essentially the
+        // same file but perhaps in a different location.
+        if (ImportedKind == MK_PrebuiltModule || ImportedKind == MK_ExplicitModule)
+          ImportedFile = PP.getHeaderSearchInfo().getPrebuiltModuleFileName(
+            ImportedName, /*FileMapOnly*/ true);
+
+        if (ImportedFile.empty())
+          ImportedFile = ReadPath(F, Record, Idx);
+        else
+          SkipPath(Record, Idx);
 
         // If our client can't cope with us being out of date, we can't cope with
         // our dependency being missing.
@@ -3420,12 +3436,18 @@ void ASTReader::ReadModuleOffsetMap(ModuleFile &F) const {
   RemapBuilder TypeRemap(F.TypeRemap);
 
   while (Data < DataEnd) {
-    // FIXME: Looking up dependency modules by filename is horrible.
+    // FIXME: Looking up dependency modules by filename is horrible. Let's
+    // start fixing this with prebuilt and explicit modules and see how it
+    // goes...
     using namespace llvm::support;
+    ModuleKind Kind = static_cast<ModuleKind>(
+      endian::readNext<uint8_t, little, unaligned>(Data));
     uint16_t Len = endian::readNext<uint16_t, little, unaligned>(Data);
     StringRef Name = StringRef((const char*)Data, Len);
     Data += Len;
-    ModuleFile *OM = ModuleMgr.lookup(Name);
+    ModuleFile *OM = (Kind == MK_PrebuiltModule || Kind == MK_ExplicitModule
+                      ? ModuleMgr.lookupByModuleName(Name)
+                      : ModuleMgr.lookupByFileName(Name));
     if (!OM) {
       std::string Msg =
           "SourceLocation remap refers to unknown module, cannot find ";
@@ -4756,6 +4778,7 @@ bool ASTReader::readASTFileControlBlock(
       while (Idx < N) {
         // Read information about the AST file.
         Idx += 5; // ImportLoc, Size, ModTime, Signature
+        SkipString(Record, Idx); // Module name; FIXME: pass to listener?
         std::string Filename = ReadString(Record, Idx);
         ResolveImportedPath(Filename, ModuleDir);
         Listener.visitImport(Filename);
@@ -10346,7 +10369,8 @@ ASTReader::ASTReader(Preprocessor &PP, ASTContext *Context,
       SourceMgr(PP.getSourceManager()), FileMgr(PP.getFileManager()),
       PCHContainerRdr(PCHContainerRdr), Diags(PP.getDiagnostics()), PP(PP),
       ContextObj(Context),
-      ModuleMgr(PP.getFileManager(), PP.getPCMCache(), PCHContainerRdr),
+      ModuleMgr(PP.getFileManager(), PP.getPCMCache(), PCHContainerRdr,
+                PP.getHeaderSearchInfo()),
       PCMCache(PP.getPCMCache()), DummyIdResolver(PP),
       ReadTimer(std::move(ReadTimer)), isysroot(isysroot),
       DisableValidation(DisableValidation),