[Dependency Scanning] Scan embedded header content if file doesn't exist

Fallback to scan embedded header content when the bridging header path
encoded in the swift binary module doesn't exit on the disk.

rdar://144261730

Author: cachemeifyoucan

lib/DependencyScan/ModuleDependencyScanner.cpp

@@ -36,6 +36,7 @@
 #include "llvm/CAS/CachingOnDiskFileSystem.h"
 #include "llvm/Support/Error.h"
 #include "llvm/Support/MemoryBuffer.h"
+#include "llvm/Support/MemoryBufferRef.h"
 #include "llvm/Support/Path.h"
 #include "llvm/Support/Threading.h"
 #include "llvm/Support/VersionTuple.h"
@@ -1063,6 +1064,42 @@ void ModuleDependencyScanner::resolveHeaderDependenciesForModule(
   if (!isTextualModuleWithABridgingHeader && !isBinaryModuleWithHeaderInput)
     return;
 
+
+  std::optional<std::string> headerPath;
+  std::unique_ptr<llvm::MemoryBuffer> sourceBuffer;
+  std::optional<llvm::MemoryBufferRef> sourceBufferRef;
+
+  auto extractHeaderContent =
+      [&](const SwiftBinaryModuleDependencyStorage &binaryMod)
+      -> std::unique_ptr<llvm::MemoryBuffer> {
+    auto header = binaryMod.headerImport;
+    // Check to see if the header input exists on disk.
+    auto FS = ScanASTContext.SourceMgr.getFileSystem();
+    if (FS->exists(header))
+      return nullptr;
+
+    auto moduleBuf = FS->getBufferForFile(binaryMod.compiledModulePath);
+    if (!moduleBuf)
+      return nullptr;
+
+    auto content = extractEmbeddedBridgingHeaderContent(std::move(*moduleBuf),
+                                                        ScanASTContext);
+    if (content.empty())
+      return nullptr;
+
+    return llvm::MemoryBuffer::getMemBufferCopy(content, header);
+  };
+
+  if (isBinaryModuleWithHeaderInput) {
+    auto &binaryMod = *moduleDependencyInfo.getAsSwiftBinaryModule();
+    if (auto embeddedHeader = extractHeaderContent(binaryMod)) {
+      sourceBuffer = std::move(embeddedHeader);
+      sourceBufferRef = sourceBuffer->getMemBufferRef();
+    } else
+      headerPath = binaryMod.headerImport;
+  } else
+    headerPath = *moduleDependencyInfo.getBridgingHeader();
+
   withDependencyScanningWorker(
       [&](ModuleDependencyScanningWorker *ScanningWorker) {
         auto clangImporter = static_cast<ClangImporter *>(
@@ -1072,12 +1109,9 @@ void ModuleDependencyScanner::resolveHeaderDependenciesForModule(
         std::optional<std::string> includeTreeID;
         std::vector<std::string> bridgingHeaderCommandLine;
         auto headerScan = clangImporter->getHeaderDependencies(
-            moduleID,
-            isTextualModuleWithABridgingHeader
-                ? *moduleDependencyInfo.getBridgingHeader()
-                : moduleDependencyInfo.getAsSwiftBinaryModule()->headerImport,
-            /*sourceBuffer=*/std::nullopt, ScanningWorker->clangScanningTool,
-            cache, headerClangModuleDependencies, headerFileInputs,
+            moduleID, headerPath, sourceBufferRef,
+            ScanningWorker->clangScanningTool, cache,
+            headerClangModuleDependencies, headerFileInputs,
             bridgingHeaderCommandLine, includeTreeID);
         if (!headerScan) {
           // Record direct header Clang dependencies

test/ScanDependencies/embedded-header.swift

@@ -0,0 +1,86 @@
+// REQUIRES: objc_interop
+// RUN: %empty-directory(%t)
+// RUN: split-file %s %t
+
+// RUN: %target-swift-frontend -emit-module -module-name Test -module-cache-path %t/clang-module-cache -O \
+// RUN:   -disable-implicit-string-processing-module-import -disable-implicit-concurrency-module-import \
+// RUN:   %t/test.swift -o %t/Test.swiftmodule \
+// RUN:   -Xcc -fmodule-map-file=%t/a.modulemap -Xcc -fmodule-map-file=%t/b.modulemap -import-objc-header %t/Bridging.h
+
+// RUN: %target-swift-frontend -scan-dependencies -module-name User -module-cache-path %t/clang-module-cache -O \
+// RUN:   -disable-implicit-string-processing-module-import -disable-implicit-concurrency-module-import \
+// RUN:   %t/user.swift -o %t/deps.json \
+// RUN:   -Xcc -fmodule-map-file=%t/a.modulemap -Xcc -fmodule-map-file=%t/b.modulemap -I %t
+
+/// Remove bridging header from disk and rescan
+// RUN: rm -rf %t/Bridging.h %t/Foo.h %t/Foo2.h
+// RUN: %target-swift-frontend -scan-dependencies -module-name User -module-cache-path %t/clang-module-cache -O \
+// RUN:   -disable-implicit-string-processing-module-import -disable-implicit-concurrency-module-import \
+// RUN:   %t/user.swift -o %t/deps2.json \
+// RUN:   -Xcc -fmodule-map-file=%t/a.modulemap -Xcc -fmodule-map-file=%t/b.modulemap -I %t
+
+// RUN: %{python} %S/../CAS/Inputs/SwiftDepsExtractor.py %t/deps.json swiftPrebuiltExternal:Test headerModuleDependencies | %FileCheck %s --check-prefix=MODULE
+// RUN: %{python} %S/../CAS/Inputs/SwiftDepsExtractor.py %t/deps.json swiftPrebuiltExternal:Test headerDependenciesSourceFiles | %FileCheck %s --check-prefix=FILE
+// RUN: %{python} %S/../CAS/Inputs/SwiftDepsExtractor.py %t/deps2.json swiftPrebuiltExternal:Test headerModuleDependencies | %FileCheck %s --check-prefix=MODULE
+// RUN: %{python} %S/../CAS/Inputs/SwiftDepsExtractor.py %t/deps2.json swiftPrebuiltExternal:Test headerDependenciesSourceFiles | %FileCheck %s --check-prefix=FILE
+
+// MODULE: "A"
+// FILE: Bridging.h
+
+//--- test.swift
+public func test() {
+    b()
+}
+public class TestB: B {}
+
+//--- user.swift
+import Test
+
+func user() {
+  var b: TestB
+  test()
+}
+
+extension A {
+    public func testA() {}
+}
+
+//--- Bridging.h
+#include "Foo.h"
+#include "Foo2.h"
+
+//--- Foo.h
+#import "a.h"
+#ifndef IMPORT_FOO
+#define IMPORT_FOO
+int Foo = 0;
+#endif
+
+//--- Foo2.h
+#ifndef IMPORT_FOO2
+#define IMPORT_FOO2
+int Foo2 = 0;
+#endif
+
+//--- a.h
+#include "b.h"
+struct A {
+  int a;
+};
+
+//--- b.h
+void b(void);
+@interface B
+@end
+
+//--- a.modulemap
+module A {
+  header "a.h"
+  export *
+}
+
+//--- b.modulemap
+module B {
+  header "b.h"
+  export *
+}