[swift-api-extract] BumpPtrAllocate API structs

Fix a bug that pointers to ObjCInterfaces can be invalid when std vector
expended. Now allocate all the interfaces in BumpPtrAllocator so the
pointers to APIs are always valid as long as API class is alive.

rdar://76155450

Author: cachemeifyoucan

lib/TBDGen/APIGen.cpp

@@ -27,24 +27,27 @@ namespace apigen {
 void API::addSymbol(llvm::StringRef symbol, APILoc loc, APILinkage linkage,
                     APIFlags flags, APIAccess access,
                     APIAvailability availability) {
-  globals.emplace_back(symbol, loc, linkage, flags, access, GVKind::Function,
-                       availability);
+  auto *global = new (allocator) GlobalRecord(
+      symbol, loc, linkage, flags, access, GVKind::Function, availability);
+  globals.push_back(global);
 }
 
 ObjCInterfaceRecord *API::addObjCClass(llvm::StringRef name, APILinkage linkage,
                                        APILoc loc, APIAccess access,
                                        APIAvailability availability,
                                        llvm::StringRef superClassName) {
-  interfaces.emplace_back(name, linkage, loc, access, availability,
-                          superClassName);
-  return &interfaces.back();
+  auto *interface = new (allocator) ObjCInterfaceRecord(
+      name, linkage, loc, access, availability, superClassName);
+  interfaces.push_back(interface);
+  return interface;
 }
 
 void API::addObjCMethod(ObjCInterfaceRecord *cls, llvm::StringRef name,
                         APILoc loc, APIAccess access, bool isInstanceMethod,
                         bool isOptional, APIAvailability availability) {
-  cls->methods.emplace_back(name, loc, access, isInstanceMethod, isOptional,
-                            availability);
+  auto method = new (allocator) ObjCMethodRecord(
+      name, loc, access, isInstanceMethod, isOptional, availability);
+  cls->methods.push_back(method);
 }
 
 static void serialize(llvm::json::OStream &OS, APIAccess access) {
@@ -120,6 +123,10 @@ static void serialize(llvm::json::OStream &OS, const ObjCMethodRecord &record) {
   });
 }
 
+static bool sortAPIRecords(const APIRecord *base, const APIRecord *compare) {
+  return base->name < compare->name;
+}
+
 static void serialize(llvm::json::OStream &OS,
                       const ObjCInterfaceRecord &record) {
   OS.object([&]() {
@@ -131,39 +138,34 @@ static void serialize(llvm::json::OStream &OS,
     OS.attribute("super", record.superClassName);
     OS.attributeArray("instanceMethods", [&]() {
       for (auto &method : record.methods) {
-        if (method.isInstanceMethod)
-          serialize(OS, method);
+        if (method->isInstanceMethod)
+          serialize(OS, *method);
       }
     });
     OS.attributeArray("classMethods", [&]() {
       for (auto &method : record.methods) {
-        if (!method.isInstanceMethod)
-          serialize(OS, method);
+        if (!method->isInstanceMethod)
+          serialize(OS, *method);
       }
     });
   });
 }
 
-static bool sortAPIRecords(const APIRecord &base, const APIRecord &compare) {
-  return base.name < compare.name;
-}
-
 void API::writeAPIJSONFile(llvm::raw_ostream &os, bool PrettyPrint) {
   unsigned indentSize = PrettyPrint ? 2 : 0;
   llvm::json::OStream JSON(os, indentSize);
 
-  // FIXME: only write PublicSDKContentRoot now.
   JSON.object([&]() {
     JSON.attribute("target", target.str());
     JSON.attributeArray("globals", [&]() {
       llvm::sort(globals, sortAPIRecords);
-      for (const auto &g : globals)
-        serialize(JSON, g);
+      for (const auto *g : globals)
+        serialize(JSON, *g);
     });
     JSON.attributeArray("interfaces", [&]() {
       llvm::sort(interfaces, sortAPIRecords);
-      for (const auto &i : interfaces)
-        serialize(JSON, i);
+      for (const auto *i : interfaces)
+        serialize(JSON, *i);
     });
     JSON.attribute("version", "1.0");
   });

lib/TBDGen/APIGen.h

@@ -18,6 +18,7 @@
 #include "llvm/ADT/StringRef.h"
 #include "llvm/ADT/StringSet.h"
 #include "llvm/ADT/Triple.h"
+#include "llvm/Support/Allocator.h"
 #include "llvm/Support/Error.h"
 
 namespace llvm {
@@ -86,7 +87,7 @@ struct APIRecord {
 
   APIRecord(llvm::StringRef name, APILoc loc, APILinkage linkage,
             APIFlags flags, APIAccess access, APIAvailability availability)
-      : name(name.str()), loc(loc), linkage(linkage), flags(flags),
+      : name(name.data(), name.size()), loc(loc), linkage(linkage), flags(flags),
         access(access), availability(availability) {}
 
   bool isWeakDefined() const {
@@ -135,7 +136,7 @@ struct ObjCMethodRecord : APIRecord {
 };
 
 struct ObjCContainerRecord : APIRecord {
-  std::vector<ObjCMethodRecord> methods;
+  std::vector<ObjCMethodRecord*> methods;
 
   ObjCContainerRecord(llvm::StringRef name, APILinkage linkage, APILoc loc,
                       APIAccess access, const APIAvailability &availability)
@@ -148,12 +149,13 @@ struct ObjCInterfaceRecord : ObjCContainerRecord {
                       APIAccess access, APIAvailability availability,
                       llvm::StringRef superClassName)
       : ObjCContainerRecord(name, linkage, loc, access, availability),
-        superClassName(superClassName.str()) {}
+        superClassName(superClassName.data(), superClassName.size()) {}
 };
 
 class API {
 public:
   API(const llvm::Triple &triple) : target(triple) {}
+
   const llvm::Triple &getTarget() const { return target; }
 
   void addSymbol(llvm::StringRef symbol, APILoc loc, APILinkage linkage,
@@ -174,8 +176,9 @@ class API {
 private:
   const llvm::Triple target;
 
-  std::vector<GlobalRecord> globals;
-  std::vector<ObjCInterfaceRecord> interfaces;
+  llvm::BumpPtrAllocator allocator;
+  std::vector<GlobalRecord*> globals;
+  std::vector<ObjCInterfaceRecord*> interfaces;
 };
 
 } // end namespace apigen