[Dependency Scanning] Include initial PCM arguments on Clang module dependency details

Doing so will allow clients to know which Swift-specific PCM arguments are already captured from the scan that first discovered this module.
SwiftDriver, in particular, will be able to use this information to avoid re-scanning a given Clang module if the initial scan was sufficient for all possible sets of PCM arguments on Swift modules that depend on said Clang module.

Author: artemcm

include/swift-c/DependencyScan/DependencyScan.h

@@ -203,6 +203,9 @@ swiftscan_clang_detail_get_context_hash(swiftscan_module_details_t details);
 SWIFTSCAN_PUBLIC swiftscan_string_set_t *
 swiftscan_clang_detail_get_command_line(swiftscan_module_details_t details);
 
+SWIFTSCAN_PUBLIC swiftscan_string_set_t *
+swiftscan_clang_detail_get_captured_pcm_args(swiftscan_module_details_t details);
+
 //=== Batch Scan Input Functions ------------------------------------------===//
 
 /// Create an \c swiftscan_batch_scan_input_t instance.

include/swift/AST/ModuleDependencies.h

@@ -247,16 +247,22 @@ class ClangModuleDependenciesStorage : public ModuleDependenciesStorageBase {
   /// The file dependencies
   const std::vector<std::string> fileDependencies;
 
+  /// The swift-specific PCM arguments captured by this dependencies object
+  /// as found by the scanning action that discovered it
+  const std::vector<std::string> capturedPCMArgs;
+
   ClangModuleDependenciesStorage(
       const std::string &moduleMapFile,
       const std::string &contextHash,
       const std::vector<std::string> &nonPathCommandLine,
-      const std::vector<std::string> &fileDependencies
+      const std::vector<std::string> &fileDependencies,
+      const std::vector<std::string> &capturedPCMArgs
   ) : ModuleDependenciesStorageBase(ModuleDependenciesKind::Clang),
       moduleMapFile(moduleMapFile),
       contextHash(contextHash),
       nonPathCommandLine(nonPathCommandLine),
-      fileDependencies(fileDependencies) { }
+      fileDependencies(fileDependencies),
+      capturedPCMArgs(capturedPCMArgs) {}
 
   ModuleDependenciesStorageBase *clone() const override {
     return new ClangModuleDependenciesStorage(*this);
@@ -361,10 +367,12 @@ class ModuleDependencies {
       const std::string &moduleMapFile,
       const std::string &contextHash,
       const std::vector<std::string> &nonPathCommandLine,
-      const std::vector<std::string> &fileDependencies) {
+      const std::vector<std::string> &fileDependencies,
+      const std::vector<std::string> &capturedPCMArgs) {
     return ModuleDependencies(
         std::make_unique<ClangModuleDependenciesStorage>(
-          moduleMapFile, contextHash, nonPathCommandLine, fileDependencies));
+          moduleMapFile, contextHash, nonPathCommandLine,
+          fileDependencies, capturedPCMArgs));
   }
 
   /// Describe a placeholder dependency swift module.

include/swift/DependencyScan/DependencyScanImpl.h

@@ -129,6 +129,9 @@ typedef struct {
 
   /// Options to the compile command required to build this clang modulemap
   swiftscan_string_set_t *command_line;
+
+  /// The swift-specific PCM arguments captured by this dependencies object
+  swiftscan_string_set_t *captured_pcm_args;
 } swiftscan_clang_details_t;
 
 struct swiftscan_module_details_s {

include/swift/DependencyScan/SerializedModuleDependencyCacheFormat.h

@@ -46,7 +46,6 @@ const unsigned MODULE_DEPENDENCY_CACHE_FORMAT_VERSION_MINOR = 0;
 using IdentifierIDField = BCVBR<13>;
 using FileIDField = IdentifierIDField;
 using ModuleIDField = IdentifierIDField;
-using CompilerFlagField = IdentifierIDField;
 using ContextHashField = IdentifierIDField;
 
 /// A bit that indicates whether or not a module is a framework
@@ -166,7 +165,8 @@ using ClangModuleDetailsLayout =
                    FileIDField,               // moduleMapPath
                    ContextHashField,          // contextHash
                    FlagIDArrayIDField,        // commandLine
-                   FileIDArrayIDField         // fileDependencies
+                   FileIDArrayIDField,        // fileDependencies
+                   FlagIDArrayIDField         // capturedPCMArgs
                    >;
 } // namespace graph_block
 

lib/ClangImporter/ClangModuleDependencyScanner.cpp

@@ -193,13 +193,16 @@ static ClangModuleDependenciesCacheImpl *getOrCreateClangImpl(
 void ClangImporter::recordModuleDependencies(
     ModuleDependenciesCache &cache,
     const FullDependenciesResult &clangDependencies) {
-  struct ModuleInfo {
-    std::string PCMPath;
-    std::string ModuleMapPath;
-  };
   auto &ctx = Impl.SwiftContext;
   auto currentSwiftSearchPathSet = ctx.getAllModuleSearchPathsSet();
 
+  // This scanner invocation's already-captured APINotes version
+  std::vector<std::string> capturedPCMArgs = {
+    "-Xcc",
+    ("-fapinotes-swift-version=" +
+     ctx.LangOpts.EffectiveLanguageVersion.asAPINotesVersionString())
+  };
+
   for (const auto &clangModuleDep : clangDependencies.DiscoveredModules) {
     // If we've already cached this information, we're done.
     if (cache.hasDependencies(
@@ -293,7 +296,8 @@ void ClangImporter::recordModuleDependencies(
         clangModuleDep.ClangModuleMapFile,
         clangModuleDep.ContextHash,
         swiftArgs,
-        fileDeps);
+        fileDeps,
+        capturedPCMArgs);
     for (const auto &moduleName : clangModuleDep.ClangModuleDeps) {
       dependencies.addModuleDependency(moduleName.ModuleName, &alreadyAddedModules);
     }

lib/DependencyScan/ModuleDependencyCacheSerialization.cpp

@@ -447,10 +447,11 @@ bool Deserializer::readGraph(GlobalModuleDependenciesCache &cache) {
         llvm::report_fatal_error("Unexpected CLANG_MODULE_DETAILS_NODE record");
       cache.configureForTriple(getTriple());
       unsigned moduleMapPathID, contextHashID, commandLineArrayID,
-          fileDependenciesArrayID;
+               fileDependenciesArrayID, capturedPCMArgsArrayID;
       ClangModuleDetailsLayout::readRecord(Scratch, moduleMapPathID,
                                            contextHashID, commandLineArrayID,
-                                           fileDependenciesArrayID);
+                                           fileDependenciesArrayID,
+                                           capturedPCMArgsArrayID);
       auto moduleMapPath = getIdentifier(moduleMapPathID);
       if (!moduleMapPath)
         llvm::report_fatal_error("Bad module map path");
@@ -463,10 +464,15 @@ bool Deserializer::readGraph(GlobalModuleDependenciesCache &cache) {
       auto fileDependencies = getArray(fileDependenciesArrayID);
       if (!fileDependencies)
         llvm::report_fatal_error("Bad file dependencies");
+      auto capturedPCMArgs = getArray(capturedPCMArgsArrayID);
+      if (!capturedPCMArgs)
+        llvm::report_fatal_error("Bad captured PCM Args");
 
       // Form the dependencies storage object
       auto moduleDep = ModuleDependencies::forClangModule(
-          *moduleMapPath, *contextHash, *commandLineArgs, *fileDependencies);
+          *moduleMapPath, *contextHash, *commandLineArgs, *fileDependencies,
+          *capturedPCMArgs);
+
       // Add dependencies of this module
       for (const auto &moduleName : *currentModuleDependencies)
         moduleDep.addModuleDependency(moduleName);
@@ -572,6 +578,7 @@ enum ModuleIdentifierArrayKind : uint8_t {
   BridgingModuleDependencies,
   NonPathCommandLine,
   FileDependencies,
+  CapturedPCMArgs,
   LastArrayKind
 };
 
@@ -846,7 +853,8 @@ void Serializer::writeModuleInfo(ModuleDependencyID moduleID,
         getIdentifier(clangDeps->moduleMapFile),
         getIdentifier(clangDeps->contextHash),
         getArray(moduleID, ModuleIdentifierArrayKind::NonPathCommandLine),
-        getArray(moduleID, ModuleIdentifierArrayKind::FileDependencies));
+        getArray(moduleID, ModuleIdentifierArrayKind::FileDependencies),
+        getArray(moduleID, ModuleIdentifierArrayKind::CapturedPCMArgs));
 
     break;
   }
@@ -1015,6 +1023,8 @@ void Serializer::collectStringsAndArrays(
                    clangDeps->nonPathCommandLine);
           addArray(moduleID, ModuleIdentifierArrayKind::FileDependencies,
                    clangDeps->fileDependencies);
+          addArray(moduleID, ModuleIdentifierArrayKind::CapturedPCMArgs,
+                   clangDeps->capturedPCMArgs);
           break;
         }
         default:

lib/DependencyScan/ScanDependencies.cpp

@@ -716,6 +716,10 @@ static void writeJSON(llvm::raw_ostream &out,
 
       // Command line.
       writeJSONSingleField(out, "commandLine", clangDeps->command_line, 5,
+                           /*trailingComma=*/true);
+
+      // Captured PCM arguments.
+      writeJSONSingleField(out, "capturedPCMArgs", clangDeps->captured_pcm_args, 5,
                            /*trailingComma=*/false);
     }
 
@@ -841,7 +845,6 @@ generateFullDependencyGraph(CompilerInstance &instance,
         // TODO: Once the clients are taught about the new dependency kind,
         // switch to using a bespoke kind here.
         details->kind = SWIFTSCAN_DEPENDENCY_INFO_SWIFT_TEXTUAL;
-
         details->swift_textual_details = {
             moduleInterfacePath,
             create_empty_set(),
@@ -870,7 +873,9 @@ generateFullDependencyGraph(CompilerInstance &instance,
         details->clang_details = {
             create_clone(clangDeps->moduleMapFile.c_str()),
             create_clone(clangDeps->contextHash.c_str()),
-            create_set(clangDeps->nonPathCommandLine)};
+            create_set(clangDeps->nonPathCommandLine),
+            create_set(clangDeps->capturedPCMArgs)
+        };
       }
       return details;
     };

test/ScanDependencies/module_deps.swift

@@ -133,6 +133,11 @@ import SubE
 // CHECK-NEXT: "-module-name",
 // CHECK-NEXT: "C"
 
+// CHECK: "capturedPCMArgs": [
+// CHECK-NEXT:   "-Xcc",
+// CHECK-NEXT:   "-fapinotes-swift-version=4"
+// CHECK-NEXT: ]
+
 /// --------Swift module E
 // CHECK: "swift": "E"
 // CHECK-LABEL: modulePath": "E.swiftmodule"

tools/libSwiftScan/libSwiftScan.cpp

@@ -73,6 +73,7 @@ void swiftscan_dependency_info_details_dispose(
     swiftscan_string_dispose(details_impl->clang_details.module_map_path);
     swiftscan_string_dispose(details_impl->clang_details.context_hash);
     swiftscan_string_set_dispose(details_impl->clang_details.command_line);
+    swiftscan_string_set_dispose(details_impl->clang_details.captured_pcm_args);
     break;
   }
   delete details_impl;
@@ -345,6 +346,11 @@ swiftscan_clang_detail_get_command_line(swiftscan_module_details_t details) {
   return details->clang_details.command_line;
 }
 
+swiftscan_string_set_t *
+swiftscan_clang_detail_get_captured_pcm_args(swiftscan_module_details_t details) {
+  return details->clang_details.captured_pcm_args;
+}
+
 //=== Batch Scan Input Functions ------------------------------------------===//
 
 swiftscan_batch_scan_input_t *swiftscan_batch_scan_input_create() {

tools/libSwiftScan/libSwiftScan.exports

@@ -24,6 +24,7 @@ swiftscan_swift_placeholder_detail_get_module_source_info_path
 swiftscan_clang_detail_get_module_map_path
 swiftscan_clang_detail_get_context_hash
 swiftscan_clang_detail_get_command_line
+swiftscan_clang_detail_get_captured_pcm_args
 swiftscan_batch_scan_input_set_modules
 swiftscan_batch_scan_entry_set_module_name
 swiftscan_batch_scan_entry_set_arguments