[clang] [clang][Dependency Scanning] Refactor Scanning Compiler Instance Initialization (PR #161300)

[Qiongsi Wu via cfe-commits]

================
@@ -480,26 +624,63 @@ bool DependencyScanningAction::runInvocation(
   case ScanningOutputFormat::P1689:
   case ScanningOutputFormat::Full:
     MDC = std::make_shared<ModuleDepCollector>(
-        Service, std::move(Opts), ScanInstance, Consumer, Controller,
-        OriginalInvocation, std::move(PrebuiltModulesASTMap), StableDirs);
+        Service, std::move(Opts), ScanInstance, Consumer, Controller, Inv,
+        std::move(PrebuiltModulesASTMap), StableDirs);
     ScanInstance.addDependencyCollector(MDC);
     break;
   }
 
-  // Consider different header search and diagnostic options to create
-  // different modules. This avoids the unsound aliasing of module PCMs.
-  //
-  // TODO: Implement diagnostic bucketing to reduce the impact of strict
-  // context hashing.
-  ScanInstance.getHeaderSearchOpts().ModulesStrictContextHash = true;
-  ScanInstance.getHeaderSearchOpts().ModulesSerializeOnlyPreprocessor = true;
-  ScanInstance.getHeaderSearchOpts().ModulesSkipDiagnosticOptions = true;
-  ScanInstance.getHeaderSearchOpts().ModulesSkipHeaderSearchPaths = true;
-  ScanInstance.getHeaderSearchOpts().ModulesSkipPragmaDiagnosticMappings = true;
-  ScanInstance.getHeaderSearchOpts().ModulesForceValidateUserHeaders = false;
+  return MDC;
+}
+} // namespace clang::tooling::dependencies
 
-  // Avoid some checks and module map parsing when loading PCM files.
-  ScanInstance.getPreprocessorOpts().ModulesCheckRelocated = false;
+bool DependencyScanningAction::runInvocation(
+    std::unique_ptr<CompilerInvocation> Invocation,
+    IntrusiveRefCntPtr<llvm::vfs::FileSystem> FS,
+    std::shared_ptr<PCHContainerOperations> PCHContainerOps,
+    DiagnosticConsumer *DiagConsumer) {
+  // Making sure that we canonicalize the defines before we create the deep
+  // copy to avoid unnecessary variants in the scanner and in the resulting
+  // explicit command lines.
+  if (any(Service.getOptimizeArgs() & ScanningOptimizations::Macros))
+    canonicalizeDefines(Invocation->getPreprocessorOpts());
+
+  // Make a deep copy of the original Clang invocation.
+  CompilerInvocation OriginalInvocation(*Invocation);
+
+  if (Scanned) {
+    // Scanning runs once for the first -cc1 invocation in a chain of driver
+    // jobs. For any dependent jobs, reuse the scanning result and just
+    // update the LastCC1Arguments to correspond to the new invocation.
+    // FIXME: to support multi-arch builds, each arch requires a separate scan
+    setLastCC1Arguments(std::move(OriginalInvocation));
+    return true;
+  }
+
+  Scanned = true;
+
+  // Create a compiler instance to handle the actual work.
+  auto ModCache = makeInProcessModuleCache(Service.getModuleCacheEntries());
+  ScanInstanceStorage.emplace(std::move(Invocation), std::move(PCHContainerOps),
+                              ModCache.get());
+  CompilerInstance &ScanInstance = *ScanInstanceStorage;
+
+  assert(!DiagConsumerFinished && "attempt to reuse finished consumer");
+  if (!initializeScanCompilerInstance(ScanInstance, FS, DiagConsumer, Service,
+                                      DepFS))
+    return false;
+
+  llvm::SmallVector<StringRef> StableDirs = getInitialStableDirs(ScanInstance);
+  auto MaybePrebuiltModulesASTMap =
+      computePrebuiltModulesASTMap(ScanInstance, StableDirs);
+  if (!MaybePrebuiltModulesASTMap)
+    return false;
+
+  auto DepOutputOpts = getDependencyOutputOptions(ScanInstance);
+
+  MDC = initializeScanInstanceDependencyCollector(
+      ScanInstance, *DepOutputOpts, WorkingDirectory, Consumer, Service,
----------------
qiongsiwu wrote:

I was thinking that when we reuse the compiler instance, we want to preserve a copy of `DepOutputOpts`, so I made a copy inside the function. But I think you are right that I can actually create the copy when I call the function and pass in the newly created copy so we can preserve `DepOutputOpts` only when necessary. Fixed! 

https://github.com/llvm/llvm-project/pull/161300