Reland "[TextAPI] Implement TBDv5 Reader"

Introduce initial reader for TBDv5 which is in JSON. This captures all
the currently understood fields within the internal structure
`InterfaceFile`.
New fields will be followed up in future PRs.

Reviewed By: pete

Differential Revision: https://reviews.llvm.org/D144156

Author: cyndyishida

lld/test/MachO/Inputs/libStubLink.tbd

@@ -22,3 +22,4 @@ current-version: 1.0.0
 exports:
   - targets:         [ arm64-ios-simulator ]
     symbols:         [ _arm64_sim_only ]
+...

lld/test/MachO/invalid/invalid-stub.s

@@ -7,9 +7,8 @@
 # RUN: not %lld -L%t -linvalidYAML %t/test.o -o %t/test 2>&1 | FileCheck %s -DDIR=%t
 # RUN: not %lld -F%t -framework invalidYAML %t/test.o -o %t/test 2>&1 | FileCheck %s -DDIR=%t --check-prefix=CHECK-FRAMEWORK
 
-# CHECK: could not load TAPI file at [[DIR]]{{[\\/]}}libinvalidYAML.tbd: malformed file
-# CHECK-FRAMEWORK: could not load TAPI file at [[DIR]]{{[\\/]}}invalidYAML.framework{{[\\/]}}invalidYAML.tbd: malformed file
-
+# CHECK: could not load TAPI file at [[DIR]]{{[\\/]}}libinvalidYAML.tbd: unsupported file type  
+# CHECK-FRAMEWORK: could not load TAPI file at [[DIR]]{{[\\/]}}invalidYAML.framework{{[\\/]}}invalidYAML.tbd: unsupported file type  
 .globl _main
 _main:
   ret

lld/test/MachO/tapi-link.s

@@ -121,7 +121,6 @@ exports:
     re-exports: [ 'libNested.dylib' ]
 ...
 
-## This tests that weak and thread-local symbols are imported as such.
 #--- libTlvWeak.tbd
 --- !tapi-tbd
 tbd-version:      4
@@ -131,8 +130,8 @@ uuids:
     value:        00000000-0000-0000-0000-000000000000
 install-name:     '/usr/lib/libTlvWeak.dylib'
 current-version:  0001.001.1
-exports:
+exports:                            # Validate weak & thread-local symbols 
   - targets:      [ x86_64-macos ]
     weak-symbols: [ _weak ]
     thread-local-symbols: [ _tlv ]
----
+...

llvm/include/llvm/TextAPI/InterfaceFile.h

@@ -66,6 +66,9 @@ enum FileType : unsigned {
   /// Text-based stub file (.tbd) version 4.0
   TBD_V4  = 1U <<  3,
 
+  /// Text-based stub file (.tbd) version 5.0
+  TBD_V5  = 1U <<  4,
+
   All     = ~0U,
 
   LLVM_MARK_AS_BITMASK_ENUM(/*LargestValue=*/All),

llvm/lib/TextAPI/CMakeLists.txt

@@ -2,6 +2,7 @@ add_llvm_component_library(LLVMTextAPI
   Architecture.cpp
   ArchitectureSet.cpp
   InterfaceFile.cpp
+  TextStubV5.cpp
   PackedVersion.cpp
   Platform.cpp
   Symbol.cpp

llvm/lib/TextAPI/TextStub.cpp

@@ -258,16 +258,6 @@ struct UUIDv4 {
   UUIDv4(const Target &TargetID, const std::string &Value)
       : TargetID(TargetID), Value(Value) {}
 };
-
-// clang-format off
-enum TBDFlags : unsigned {
-  None                         = 0U,
-  FlatNamespace                = 1U << 0,
-  NotApplicationExtensionSafe  = 1U << 1,
-  InstallAPI                   = 1U << 2,
-  LLVM_MARK_AS_BITMASK_ENUM(/*LargestValue=*/InstallAPI),
-};
-// clang-format on
 } // end anonymous namespace.
 
 LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(Architecture)
@@ -1105,10 +1095,49 @@ static void DiagHandler(const SMDiagnostic &Diag, void *Context) {
   File->ErrorMessage = ("malformed file\n" + Message).str();
 }
 
+namespace {
+
+Expected<FileType> canReadFileType(MemoryBufferRef InputBuffer) {
+  auto TAPIFile = InputBuffer.getBuffer().trim();
+  if (TAPIFile.startswith("{") && TAPIFile.endswith("}"))
+    return FileType::TBD_V5;
+
+  if (!TAPIFile.endswith("..."))
+    return createStringError(std::errc::not_supported, "unsupported file type");
+
+  if (TAPIFile.startswith("--- !tapi-tbd\n"))
+    return FileType::TBD_V4;
+
+  if (TAPIFile.startswith("--- !tapi-tbd-v3\n"))
+    return FileType::TBD_V3;
+
+  if (TAPIFile.startswith("--- !tapi-tbd-v2\n"))
+    return FileType::TBD_V2;
+
+  if (TAPIFile.startswith("--- !tapi-tbd-v1\n") ||
+      TAPIFile.startswith("---\narchs:"))
+    return FileType::TBD_V1;
+
+  return createStringError(std::errc::not_supported, "unsupported file type");
+}
+} // namespace
+
 Expected<std::unique_ptr<InterfaceFile>>
 TextAPIReader::get(MemoryBufferRef InputBuffer) {
   TextAPIContext Ctx;
   Ctx.Path = std::string(InputBuffer.getBufferIdentifier());
+  if (auto FTOrErr = canReadFileType(InputBuffer))
+    Ctx.FileKind = *FTOrErr;
+  else
+    return FTOrErr.takeError();
+
+  // Handle JSON Format.
+  if (Ctx.FileKind >= FileType::TBD_V5) {
+    auto FileOrErr = getInterfaceFileFromJSON(InputBuffer.getBuffer());
+    if (!FileOrErr)
+      return FileOrErr.takeError();
+    return std::move(*FileOrErr);
+  }
   yaml::Input YAMLIn(InputBuffer.getBuffer(), &Ctx, DiagHandler, &Ctx);
 
   // Fill vector with interface file objects created by parsing the YAML file.

llvm/lib/TextAPI/TextStubCommon.cpp

@@ -6,7 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// Implememts common Text Stub YAML mappings.
+// Implements common Text Stub YAML mappings.
 //
 //===----------------------------------------------------------------------===//
 

llvm/lib/TextAPI/TextStubCommon.h

@@ -22,6 +22,16 @@
 
 using UUID = std::pair<llvm::MachO::Target, std::string>;
 
+// clang-format off
+enum TBDFlags : unsigned {
+  None                         = 0U,
+  FlatNamespace                = 1U << 0,
+  NotApplicationExtensionSafe  = 1U << 1,
+  InstallAPI                   = 1U << 2,
+  LLVM_MARK_AS_BITMASK_ENUM(/*LargestValue=*/InstallAPI),
+};
+// clang-format on
+
 LLVM_YAML_STRONG_TYPEDEF(llvm::StringRef, FlowStringRef)
 LLVM_YAML_STRONG_TYPEDEF(uint8_t, SwiftVersion)
 LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(UUID)
@@ -30,9 +40,13 @@ LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(FlowStringRef)
 namespace llvm {
 
 namespace MachO {
-    class ArchitectureSet;
-    class PackedVersion;
-}
+class ArchitectureSet;
+class PackedVersion;
+
+Expected<std::unique_ptr<InterfaceFile>>
+getInterfaceFileFromJSON(StringRef JSON);
+} // namespace MachO
+
 namespace yaml {
 
 template <> struct ScalarTraits<FlowStringRef> {