proto_zero_plugin [NOT FOR REVIEW].

components/tracing/proto_zero_plugin/proto_zero_generator.cc

Index: components/tracing/proto_zero_plugin/proto_zero_generator.cc
diff --git a/components/tracing/proto_zero_plugin/proto_zero_generator.cc b/components/tracing/proto_zero_plugin/proto_zero_generator.cc
new file mode 100644
index 0000000000000000000000000000000000000000..c4a504519951571a72358a8426747ec624f5539a
--- /dev/null
+++ b/components/tracing/proto_zero_plugin/proto_zero_generator.cc
@@ -0,0 +1,400 @@
+// Copyright (c) 2016 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "proto_zero_generator.h"
+
+#include <algorithm>
+#include <memory>
+#include <string>
+
+#include "third_party/protobuf/src/google/protobuf/descriptor.h"
+#include "third_party/protobuf/src/google/protobuf/io/printer.h"
+#include "third_party/protobuf/src/google/protobuf/io/zero_copy_stream.h"
+#include "third_party/protobuf/src/google/protobuf/stubs/strutil.h"
+
+namespace tracing {
+namespace proto {
+
+// MessageDescriptor is a longer name but has more obvious meaning.

[Primiano Tucci (use gerrit), 2016/07/11 16:52:07]

It also makes the class harder to codesearch. I'd stick with MessageDescriptor
(so just using google::protobuf::Descriptor)

+using MessageDescriptor = google::protobuf::Descriptor;
+using google::protobuf::EnumDescriptor;
+using google::protobuf::EnumValueDescriptor;
+using google::protobuf::FieldDescriptor;
+using google::protobuf::FileDescriptor;
+using google::protobuf::io::Printer;
+using google::protobuf::io::ZeroCopyOutputStream;
+
+using google::protobuf::Split;
+using google::protobuf::StripPrefixString;
+using google::protobuf::StripString;
+using google::protobuf::StripSuffixString;
+using google::protobuf::UpperString;
+
+namespace {
+
+class GeneratorJob {
+ public:
+  GeneratorJob(const FileDescriptor* file, Printer* printer)
+    : file_(file), printer_(printer) {}
+
+  bool GenerateStubs() {
+    Prepare();
+    PrintPrologue();
+    for (const EnumDescriptor* enumeration: enums_) {

[Primiano Tucci (use gerrit), 2016/07/11 16:52:07]

you can drop the braces for single line for loops (and if-s)

+      PrintEnum(enumeration);
+    }
+    for (const MessageDescriptor* message : messages_) {
+      PrintMessage(message);
+    }
+    PrintEpilogue();
+    return error_.empty();
+  }
+
+  std::string GetFirstError() {

[Primiano Tucci (use gerrit), 2016/07/11 16:52:07]

not sure you want to return this by copy. I think this should be a:
const std::string& GetFirstError() const 

+    return error_;
+  }
+
+ private:
+  // Get C++ class name corresponding to message/enum descriptor.
+  // Nested names are splitted by underscores.
+  template <class T>
+  std::string GetCppClassName(const T* descriptor, bool absolute = false) {
+    std::string local_name = descriptor->name();
+    if (local_name.find('_') != std::string::npos) {
+      Abort("Underscores are not permitted in proto's type name.");
+      return " /* Invalid name */ ";
+    }
+    std::string name =
+        StripPrefixString(descriptor->full_name(), package_ + ".");
+    StripString(&name, ".", '_');
+    if (absolute) {
+      name = full_namespace_prefix_ + name;
+    }
+    return name;
+  }
+
+  // Small enums can be written faster without involving VarInt encoder.
+  inline bool IsEnumOptimal(const EnumDescriptor* enumeration) {
+    for (int i = 0; i < enumeration->value_count(); ++i) {
+      int32_t number = enumeration->value(i)->number();
+      if (number < 0 || number > 0x7F)
+        return false;
+    }
+    return true;
+  }
+
+  // If something went wrong, plugin returns the first error occured.
+  void Abort(const std::string& reason) {
+    if (error_.empty()) {
+      error_ = reason;
+    }
+  }
+
+  // Scanning pass.

[Primiano Tucci (use gerrit), 2016/07/11 16:52:07]

This comment doesn't say anything more than "Prepare". Either drop it or say
what it does. I'd just drop it as you have comments below.
Also I'd probably call it "Preprocess"

+  void Prepare() {
+    // Package name evaluates to a series of namespaces.
+    package_ = file_->package();
+    namespaces_ = Split(package_, ".");
+    full_namespace_prefix_ = "::";
+    for (const std::string& ns : namespaces_) {
+      full_namespace_prefix_ += ns + "::";
+    }
+    // Collect message descriptors in DFS order.
+    std::vector<const MessageDescriptor*> stack;
+    for (int i = 0; i < file_->message_type_count(); ++i) {
+      stack.push_back(file_->message_type(i));
+    }
+    while (!stack.empty()) {
+      const MessageDescriptor* message = stack.back();
+      stack.pop_back();
+      messages_.push_back(message);
+      for (int i = 0; i < message->nested_type_count(); ++i) {
+        stack.push_back(message->nested_type(i));
+      }
+    }
+    // Collect enums.
+    for (int i = 0; i < file_->enum_type_count(); ++i) {
+      enums_.push_back(file_->enum_type(i));
+    }
+    for (const MessageDescriptor* message : messages_) {
+      for (int i = 0; i < message->enum_type_count(); ++i) {
+        enums_.push_back(message->enum_type(i));
+      }
+    }
+  }
+
+  // Print top header, namespaces and forward declarations.
+  void PrintPrologue() {
+    std::string guard = package_ + "_" + file_->name() + "_H_";
+    UpperString(&guard);
+    StripString(&guard, ".-", '_');
+    printer_->Print(
+        "// Autogenerated. DO NOT EDIT.\n"
+        "// Protoc has generated these stubs "
+        "with //components/tracing/proto_zero_plugin.\n\n"
+        "#ifndef $guard$\n"
+        "#define $guard$\n\n"
+        "#include <inttypes.h>\n\n"

[Primiano Tucci (use gerrit), 2016/07/11 16:52:07]

s/inttypes/stdint/

+        "// Runtime is optional for testing purposes.\n"
+        "#ifndef PROTO_ZERO_NO_RUNTIME\n"
+        "#include \"components/tracing/core/proto_zero_message.h\"\n"
+        "#endif\n\n",
+        "guard", guard);
+    // Print namespaces.

[Primiano Tucci (use gerrit), 2016/07/11 16:52:07]

add a blank newline before these comments. It makes code easier to read as it
clarifies which line the code belongs to.

+    for (const std::string& ns : namespaces_) {
+      printer_->Print("namespace $ns$ {\n", "ns", ns);
+    }
+    printer_->Print("\n");
+    // Print forward declarations.
+    for (const MessageDescriptor* message : messages_) {
+      printer_->Print(
+          "class $class$;\n",
+          "class", GetCppClassName(message));
+    }
+    printer_->Print("\n");
+  }
+
+  void PrintSimpleField(const FieldDescriptor* field) {
+    std::string appender;
+    std::string cpp_type;
+    std::string action = field->is_repeated() ? "add" : "set";
+
+    switch (field->type()) {
+      case FieldDescriptor::TYPE_BOOL: {
+        appender = "AppendBool";
+        cpp_type = "bool";
+        break;
+      }
+      case FieldDescriptor::TYPE_INT32: {
+        appender = "AppendInt32";
+        cpp_type = "int32_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_INT64: {
+        appender = "AppendInt64";
+        cpp_type = "int64_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_UINT32: {
+        appender = "AppendUint32";
+        cpp_type = "uint32_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_UINT64: {
+        appender = "AppendUint64";
+        cpp_type = "uint64_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_SINT32: {
+        appender = "AppendSint32";
+        cpp_type = "int32_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_SINT64: {
+        appender = "AppendSint64";
+        cpp_type = "int64_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_FIXED32: {
+        appender = "AppendFixed32";
+        cpp_type = "uint32_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_FIXED64: {
+        appender = "AppendFixed64";
+        cpp_type = "uint64_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_SFIXED32: {
+        appender = "AppendSfixed32";
+        cpp_type = "int32_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_SFIXED64: {
+        appender = "AppendSfixed64";
+        cpp_type = "int64_t";
+        break;
+      }
+      case FieldDescriptor::TYPE_FLOAT: {
+        appender = "AppendFloat";
+        cpp_type = "float";
+        break;
+      }
+      case FieldDescriptor::TYPE_DOUBLE: {
+        appender = "AppendDouble";
+        cpp_type = "double";
+        break;
+      }
+      case FieldDescriptor::TYPE_ENUM: {
+        if (IsEnumOptimal(field->enum_type())) {
+          appender = "AppendTinyNumber";
+        } else {
+          appender = "AppendInt32";
+        }
+        cpp_type = GetCppClassName(field->enum_type(), true);
+        break;
+      }
+      case FieldDescriptor::TYPE_STRING: {
+        appender = "AppendString";
+        cpp_type = "const char*";
+        break;
+      }
+      case FieldDescriptor::TYPE_BYTES: {
+        printer_->Print(
+            "void $action$_$name$(const uint8_t* data, size_t size) {\n"
+            "  AppendBytes($key$, data, size);\n"
+            "}\n",
+            "action", action,
+            "name", field->name(),
+            "key", std::to_string(field->number()));
+        return;
+      }
+      default: {
+        Abort("Unsupported field type.");
+        return;
+      }
+    }
+    printer_->Print(
+        "void $action$_$name$($cpp_type$ value) {\n"
+        "  $appender$($key$, value);\n"
+        "}\n",
+        "appender", appender,
+        "cpp_type", cpp_type,
+        "action", action,
+        "name", field->name(),
+        "key", std::to_string(field->number()));
+  }
+
+  // Embedded message field.
+  void PrintNestedMessageField(const FieldDescriptor* field) {
+    printer_->Print(
+        "$class$* $action$_$name$() {\n"
+        "  $class$* nested;\n"
+        "  OpenNestedMessage($key$, &nested);\n"
+        "  return nested;\n"
+        "}\n",
+        "action", field->is_repeated() ? "add" : "set",
+        "class", GetCppClassName(field->message_type()),
+        "name", field->name(),
+        "key", std::to_string(field->number()));
+  }
+
+  // Print fields of the message.
+  void PrintFields(const MessageDescriptor* message) {
+    for (int i = 0; i < message->field_count(); ++i) {
+      const FieldDescriptor* field = message->field(i);
+      if (field->is_required()) {
+        Abort("Required fields are not supported.");
+        return;
+      }
+      if (field->is_packed()) {
+        Abort("Packed repeated fields are not supported.");
+        return;
+      }
+      if (field->type() != FieldDescriptor::TYPE_MESSAGE) {
+        PrintSimpleField(field);
+      } else {
+        PrintNestedMessageField(field);
+      }
+    }
+  }
+
+  // Print enum type definition.
+  void PrintEnum(const EnumDescriptor* enumeration) {
+    printer_->Print(
+        "enum $class$ : int32_t {\n",
+        "class", GetCppClassName(enumeration));
+    printer_->Indent();
+    for (int i = 0; i < enumeration->value_count(); ++i) {
+      const EnumValueDescriptor* value = enumeration->value(i);
+      printer_->Print(
+          "$name$ = $number$,\n",
+          "name", value->name(),
+          "number", std::to_string(value->number()));
+    }
+    printer_->Outdent();
+    printer_->Print("};\n\n");
+  }
+
+  // Print message class definition.
+  void PrintMessage(const MessageDescriptor* message) {
+    printer_->Print(
+        "class $class$ : public $base$ {\n"
+        " public:\n",
+
+        "class", GetCppClassName(message),
+        "base", "::tracing::proto::AppendOnlyProtoMessage");
+    printer_->Indent();
+    // Using nested messages.
+    for (int i = 0; i < message->nested_type_count(); ++i) {
+      const MessageDescriptor* nested_message = message->nested_type(i);
+      printer_->Print(
+          "using $local_name$ = $global_name$;\n",
+          "local_name", nested_message->name(),
+          "global_name", GetCppClassName(nested_message, true));
+    }
+    // Using nested enums.
+    for (int i = 0; i < message->enum_type_count(); ++i) {
+      const EnumDescriptor* nested_enum = message->enum_type(i);
+      printer_->Print(
+          "using $local_name$ = $global_name$;\n",
+          "local_name", nested_enum->name(),
+          "global_name", GetCppClassName(nested_enum, true));
+    }
+    PrintFields(message);
+    printer_->Outdent();
+    printer_->Print("};\n\n");
+  }
+
+  // Closing braces.
+  void PrintEpilogue() {
+    for (auto it = namespaces_.rbegin(); it != namespaces_.rend(); ++it) {
+      printer_->Print("} // namespace $ns$\n", "ns", *it);
+    }
+    printer_->Print("#endif  // Include guard.\n");
+  }
+
+  const FileDescriptor* const file_;
+  Printer* const printer_;
+  std::string error_;
+
+  std::string package_;
+  std::vector<std::string> namespaces_;
+  std::string full_namespace_prefix_;
+  std::vector<const MessageDescriptor*> messages_;
+  std::vector<const EnumDescriptor*> enums_;
+};
+
+}  // namespace
+
+Generator::Generator() {
+}
+
+Generator::~Generator() {
+}
+
+bool Generator::Generate(const FileDescriptor* file,
+                         const std::string& options,
+                         google::protobuf::compiler::GeneratorContext* context,
+                         std::string* error) const {
+
+  std::string stub_name = StripSuffixString(file->name(), ".proto") + ".zeropb";
+  std::unique_ptr<ZeroCopyOutputStream> pb_h(context->Open(stub_name + ".h"));
+  std::unique_ptr<ZeroCopyOutputStream> pb_cc(context->Open(stub_name + ".cc"));
+  // $ used as variable delimiter.
+  Printer pb_h_printer(pb_h.get(), '$');
+  Printer pb_cc_printer(pb_cc.get(), '$');
+
+  GeneratorJob job(file, &pb_h_printer);
+  if (!job.GenerateStubs()) {
+    *error = job.GetFirstError();
+    return false;
+  }
+  pb_cc_printer.Print("// This file intentionally left blank.\n");
+
+  return true;
+}
+
+}  // namespace proto
+}  // namespace tracing