Add arbitrary string type to the native object API (#4218)

* Custom strings are very common for optimizations around small objects
  or growth style optimizations, i.e.: grow at 1.57 times vs doubling vs..

  A second common strategy is to cooperate w/ the memory allocator
  see FBString[1] and seastar[2] string for examples.

[1] fbstring: https://github.com/facebook/folly/blob/master/folly/docs/FBString.md
[2] sstring: https://github.com/scylladb/seastar/blob/master/core/sstring.hh

docs/source/CppUsage.md

@@ -200,6 +200,15 @@ pointer type (`my_ptr<T>`), or by specifying `naked` as the type to get `T *`
 pointers. Unlike the smart pointers, naked pointers do not manage memory for
 you, so you'll have to manage their lifecycles manually.
 
+
+# Using different string type.
+
+By default the object tree is built out of `std::string`, but you can
+influence this either globally (using the `--cpp-str-type` argument to
+`flatc`) or per field using the `cpp_str_type` attribute.
+
+The type must support T::c_str() and T::length() as member functions.
+
 ## Reflection (& Resizing)
 
 There is experimental support for reflection in FlatBuffers, allowing you to

include/flatbuffers/flatbuffers.h

@@ -980,6 +980,15 @@ FLATBUFFERS_FINAL_CLASS
     return str ? CreateString(str->c_str(), str->Length()) : 0;
   }
 
+  /// @brief Store a string in the buffer, which can contain any binary data.
+  /// @param[in] str A const reference to a std::string like type with support
+  /// of T::c_str() and T::length() to store in the buffer.
+  /// @return Returns the offset in the buffer where the string starts.
+  template<typename T>
+    Offset<String> CreateString(const T &str) {
+    return CreateString(str.c_str(), str.length());
+  }
+
   /// @brief Store a string in the buffer, which can contain any binary data.
   /// If a string with this exact contents has already been serialized before,
   /// instead simply returns the offset of the existing string.

include/flatbuffers/idl.h

@@ -353,6 +353,7 @@ struct IDLOptions {
   bool escape_proto_identifiers;
   bool generate_object_based_api;
   std::string cpp_object_api_pointer_type;
+  std::string cpp_object_api_string_type;
   bool union_value_namespacing;
   bool allow_non_utf8;
   std::string include_prefix;
@@ -479,6 +480,7 @@ class Parser : public ParserState {
     known_attributes_["idempotent"] = true;
     known_attributes_["cpp_type"] = true;
     known_attributes_["cpp_ptr_type"] = true;
+    known_attributes_["cpp_str_type"] = true;
     known_attributes_["native_inline"] = true;
     known_attributes_["native_type"] = true;
     known_attributes_["native_default"] = true;
@@ -743,4 +745,3 @@ bool GenerateGoGRPC(const Parser &parser,
 }  // namespace flatbuffers
 
 #endif  // FLATBUFFERS_IDL_H_
-

src/flatc.cpp

@@ -86,6 +86,8 @@ std::string FlatCompiler::GetUsageString(const char* program_name) const {
       "  --escape-proto-ids Disable appending '_' in namespaces names.\n"
       "  --gen-object-api   Generate an additional object-based API.\n"
       "  --cpp-ptr-type T   Set object API pointer type (default std::unique_ptr)\n"
+      "  --cpp-str-type T   Set object API string type (default std::string)\n"
+      "                     T::c_str() and T::length() must be supported\n"
       "  --no-js-exports    Removes Node.js style export lines in JS.\n"
       "  --goog-js-export   Uses goog.exports* for closure compiler exporting in JS.\n"
       "  --raw-binary       Allow binaries without file_indentifier to be read.\n"
@@ -178,6 +180,9 @@ int FlatCompiler::Compile(int argc, const char** argv) {
       } else if (arg == "--cpp-ptr-type") {
         if (++argi >= argc) Error("missing type following" + arg, true);
         opts.cpp_object_api_pointer_type = argv[argi];
+      } else if (arg == "--cpp-str-type") {
+        if (++argi >= argc) Error("missing type following" + arg, true);
+        opts.cpp_object_api_string_type = argv[argi];
       } else if(arg == "--gen-all") {
         opts.generate_all = true;
         opts.include_dependence_headers = false;

src/idl_gen_cpp.cpp

@@ -362,6 +362,15 @@ class CppGenerator : public BaseGenerator {
     return attr ? attr->constant : parser_.opts.cpp_object_api_pointer_type;
   }
 
+  const std::string NativeString(const FieldDef *field) {
+    auto attr = field ? field->attributes.Lookup("cpp_str_type") : nullptr;
+    auto &ret = attr ? attr->constant : parser_.opts.cpp_object_api_string_type;
+    if (ret.empty()) {
+      return "std::string";
+    }
+    return ret;
+  }
+
   std::string GenTypeNativePtr(const std::string &type, const FieldDef *field,
                                bool is_constructor) {
     auto &ptr_type = PtrType(field);
@@ -383,7 +392,7 @@ class CppGenerator : public BaseGenerator {
                             const FieldDef &field) {
     switch (type.base_type) {
       case BASE_TYPE_STRING: {
-        return "std::string";
+        return NativeString(&field);
       }
       case BASE_TYPE_VECTOR: {
         const auto type_name = GenTypeNative(type.VectorType(), true, field);