Initial pass for generalized encoding support for parser

src/common/parser/parser.cpp

@@ -22,8 +22,8 @@
 
 namespace jessilib {
 
-object parser::deserialize(std::istream& in_stream) {
-	std::vector<char8_t> data;
+object parser::deserialize_bytes(std::istream& in_stream, encoding in_read_encoding) {
+	std::vector<byte_type> data;
 
 	// Read entire stream into data
 	char buffer[1024];
@@ -33,12 +33,13 @@ object parser::deserialize(std::istream& in_stream) {
 	}
 
 	// Pass data to deserialize
-	return deserialize(std::u8string_view{ &data.front(), data.size() });
+	return deserialize_bytes(bytes_view_type{ &data.front(), data.size() }, in_read_encoding);
 }
 
-void parser::serialize(std::ostream& in_stream, const object& in_object) {
+void parser::serialize_bytes(std::ostream& in_stream, const object& in_object, encoding in_write_encoding) {
 	// TODO: replace this method
-	in_stream << jessilib::string_view_cast<char>(serialize(in_object));
+	auto bytes = serialize_bytes(in_object, in_write_encoding);
+	in_stream << bytes;
 }
 
 } // namespace jessilib

src/common/parsers/json.cpp

@@ -17,141 +17,53 @@
  */
 
 #include "parsers/json.hpp"
-#include <charconv>
-
-using namespace std::literals;
 
 namespace jessilib {
 
-template<typename CharT>
-std::basic_string<CharT> make_json_string(std::u8string_view in_string) {
-	std::basic_string<CharT> result;
-	result.reserve(in_string.size() + 2);
-	result = '\"';
-
-	decode_result decode;
-	while ((decode = decode_codepoint(in_string)).units != 0) {
-		if (decode.codepoint == U'\\') { // backslash
-			result += '\\';
-			result += '\\';
-		}
-		else if (decode.codepoint == U'\"') { // quotation
-			result += '\\';
-			result += '\"';
-		}
-		else if (decode.codepoint < 0x20) { // control characters
-			result += '\\';
-			result += 'u';
-			result += '0';
-			result += '0';
-
-			// overwrite last 2 zeroes with correct hexadecimal sequence
-			char data[2]; // Will only ever use 2 chars
-			char* data_end = data + sizeof(data);
-			auto to_chars_result = std::to_chars(data, data_end, static_cast<uint32_t>(decode.codepoint), 16);
-			if (to_chars_result.ptr == data) {
-				// No bytes written
-				result += '0';
-				result += '0';
-			}
-			else if (to_chars_result.ptr != data_end) {
-				// 1 byte written
-				result += '0';
-				result += data[0];
-			}
-			else {
-				// 2 bytes written
-				result += data[0];
-				result += data[1];
-			}
-		}
-		else {
-			// Valid UTF-8 sequence; copy it over
-			result.append(in_string.data(), decode.units);
-		}
+object json_parser::deserialize_bytes(bytes_view_type in_data, encoding in_write_encoding) {
+	object result;
 
-		in_string.remove_prefix(decode.units);
+	if (in_write_encoding == encoding::utf_8) {
+		std::u8string_view data_view = jessilib::string_view_cast<char8_t>(in_data);
+		deserialize_json<char8_t, true>(result, data_view);
+	}
+	else if (in_write_encoding == encoding::utf_16) {
+		std::u16string_view data_view = jessilib::string_view_cast<char16_t>(in_data);
+		deserialize_json<char16_t, true>(result, data_view);
+	}
+	else if (in_write_encoding == encoding::utf_32) {
+		std::u32string_view data_view = jessilib::string_view_cast<char32_t>(in_data);
+		deserialize_json<char32_t, true>(result, data_view);
+	}
+	else if (in_write_encoding == encoding::wchar) {
+		std::wstring_view data_view = jessilib::string_view_cast<wchar_t>(in_data);
+		deserialize_json<wchar_t, true>(result, data_view);
+	}
+	else if (in_write_encoding == encoding::multibyte) {
+		// TODO: support without copying... somehow
+		auto u8_data = mbstring_to_ustring<char8_t>(jessilib::string_view_cast<char>(in_data));
+		std::u8string_view data_view = u8_data.second;
+		deserialize_json<char8_t, true>(result, data_view);
 	}
 
-	result += '\"';
-	return result;
-}
-
-
-
-object json_parser::deserialize(std::u8string_view in_data) {
-	object result;
-	std::u8string_view data_view = jessilib::string_view_cast<char8_t>(in_data);
-	deserialize_json<char8_t, true>(result, data_view);
 	return result;
 }
 
-std::u8string json_parser::serialize(const object& in_object) {
-	static const object::array_type s_null_array;
-	static const object::map_type s_null_map;
-
-	switch (in_object.type()) {
-		case object::type::null:
-			return u8"null"s;
-
-		case object::type::boolean:
-			if (in_object.get<bool>()) {
-				return u8"true"s;
-			}
-			return u8"false"s;
-
-		case object::type::integer:
-			return static_cast<std::u8string>(jessilib::string_view_cast<char8_t>(std::to_string(in_object.get<intmax_t>())));
-
-		case object::type::decimal:
-			return static_cast<std::u8string>(jessilib::string_view_cast<char8_t>(std::to_string(in_object.get<long double>())));
-
-		case object::type::text:
-			return make_json_string<char8_t>(in_object.get<std::u8string>());
-
-		case object::type::array: {
-			if (in_object.size() == 0) {
-				return u8"[]"s;
-			}
-
-			std::u8string result;
-			result = '[';
-
-			// Serialize all objects in array
-			for (auto& obj : in_object.get<object::array_type>(s_null_array)) {
-				result += json_parser::serialize(obj);
-				result += ',';
-			}
-
-			// Return result
-			result.back() = ']';
-			return result;
-		}
-
-		case object::type::map: {
-			if (in_object.size() == 0) {
-				return u8"{}"s;
-			}
-
-			std::u8string result;
-			result = '{';
-
-			// Serialize all objects in map
-			for (auto& item : in_object.get<object::map_type>(s_null_map)) {
-				result += make_json_string<char8_t>(item.first);
-				result += ':';
-				result += json_parser::serialize(item.second);
-				result += ',';
-			}
-
-			// Return result
-			result.back() = '}';
-			return result;
-		}
-
+std::string json_parser::serialize_bytes(const object& in_object, encoding in_write_encoding) {
+	switch (in_write_encoding) {
+		case encoding::utf_8:
+			return serialize_impl<char8_t, char>(in_object);
+		case encoding::utf_16:
+			return serialize_impl<char16_t, char>(in_object);
+		case encoding::utf_32:
+			return serialize_impl<char16_t, char>(in_object);
+		case encoding::wchar:
+			return serialize_impl<char16_t, char>(in_object);
 		default:
-			throw std::invalid_argument{ "Invalid data type: " + std::to_string(static_cast<size_t>(in_object.type())) };
+			break;
 	}
+
+	return {};
 }
 
 } // namespace jessilib

src/common/serialize.cpp

@@ -54,16 +54,16 @@ object deserialize_object(std::u8string_view in_data, const std::string& in_form
 }
 
 object deserialize_object(std::istream& in_stream, const std::string& in_format) {
-	return get_parser(in_format)->deserialize(in_stream);
+	return get_parser(in_format)->deserialize_bytes(in_stream, encoding::utf_8);
 }
 
 /** Serialization */
 std::u8string serialize_object(const object& in_object, const std::string& in_format) {
-	return get_parser(in_format)->serialize(in_object);
+	return get_parser(in_format)->serialize<char8_t>(in_object);
 }
 
 void serialize_object(std::ostream& in_stream, const object& in_object, const std::string& in_format) {
-	get_parser(in_format)->serialize(in_stream, in_object);
+	get_parser(in_format)->serialize_bytes(in_stream, in_object, encoding::utf_8);
 }
 
 } // namespace jessilib

src/include/jessilib/parser.hpp

@@ -20,6 +20,7 @@
 
 #include <memory>
 #include "object.hpp"
+#include "unicode_base.hpp"
 #include "impl/parser_manager.hpp"
 
 namespace jessilib {
@@ -27,20 +28,50 @@ namespace jessilib {
 class parser {
 public:
 	virtual ~parser() = default;
+	using byte_type = uint8_t;
+	using bytes_view_type = std::basic_string_view<byte_type>;
 
 	/** Interface methods */
 
 	/**
-	 * Deserializes an object directly from a stream
+	 * Deserializes an object directly from a stream of bytes
 	 * May throw: invalid_argument
 	 *
 	 * @param in_stream Stream to deserialize object from
 	 * @return A valid (possibly null) object
 	 */
-	virtual object deserialize(std::istream& in_stream);
-	virtual object deserialize(std::u8string_view in_data) = 0; // TODO: serialize from arbitrary unicode strings
-	virtual void serialize(std::ostream& in_stream, const object& in_object);
-	virtual std::u8string serialize(const object& in_object) = 0; // TODO: serialize to arbitrary unicode strings
+	virtual object deserialize_bytes(std::istream& in_stream, encoding in_read_encoding);
+	virtual object deserialize_bytes(bytes_view_type in_data, encoding in_read_encoding) = 0;
+	virtual void serialize_bytes(std::ostream& in_stream, const object& in_object, encoding in_write_encoding);
+	virtual std::string serialize_bytes(const object& in_object, encoding in_write_encoding) = 0;
+
+	template<typename CharT>
+	object deserialize(std::basic_string_view<CharT> in_text) {
+		bytes_view_type byte_view{ reinterpret_cast<const byte_type*>(in_text.data()), in_text.size() * sizeof(CharT) };
+		return deserialize_bytes(byte_view, default_encoding_info<CharT>::text_encoding);
+	}
+
+	// Perhaps this could be condensed down to a simple method such that: serialize(out_variant, in_object, encoding)?
+	virtual std::u8string serialize_u8(const object& in_object) = 0;
+	virtual std::u16string serialize_u16(const object& in_object) = 0;
+	virtual std::u32string serialize_u32(const object& in_object) = 0;
+	virtual std::wstring serialize_w(const object& in_object) = 0;
+
+	template<typename CharT>
+	std::basic_string<CharT> serialize(const object& in_object) {
+		if constexpr (std::is_same_v<CharT, char8_t>) {
+			return serialize_u8(in_object);
+		}
+		else if constexpr (std::is_same_v<CharT, char16_t>) {
+			return serialize_u16(in_object);
+		}
+		else if constexpr (std::is_same_v<CharT, char32_t>) {
+			return serialize_u32(in_object);
+		}
+		else if constexpr (std::is_same_v<CharT, wchar_t>) {
+			return serialize_w(in_object);
+		}
+	}
 }; // parser
 
 template<typename T>
@@ -57,4 +88,47 @@ public:
 	impl::parser_manager::id m_id;
 }; // parser_registration
 
+template<typename OutCharT, typename ResultCharT = OutCharT>
+std::basic_string<ResultCharT> simple_copy(std::u8string_view in_string) {
+	if constexpr (sizeof(OutCharT) == sizeof(ResultCharT)) {
+		return { in_string.begin(), in_string.end() };
+	}
+	else if constexpr (std::is_same_v<ResultCharT, char>) {
+		// Copy in_string into result _as if_ result were of OutCharT
+		for (OutCharT codepoint : in_string) {
+			// TODO: Assuming native for now, but we need to account for endianness later
+			return { reinterpret_cast<const char*>(&codepoint), sizeof(codepoint) };
+		}
+	}
+	// else // Invalid use of simple_copy
+}
+
+template<typename OutCharT, typename ResultCharT>
+void simple_append(std::basic_string<ResultCharT>& out_string, std::u8string_view in_string) {
+	if constexpr (sizeof(OutCharT) == sizeof(ResultCharT)) {
+		out_string.append(in_string.begin(), in_string.end());
+	}
+	else if constexpr (std::is_same_v<ResultCharT, char>) {
+		// Copy in_string into result _as if_ result were of OutCharT
+		for (OutCharT codepoint : in_string) {
+			// TODO: Assuming native for now, but we need to account for endianness later
+			out_string.append(reinterpret_cast<const char*>(&codepoint), sizeof(codepoint));
+		}
+	}
+	// else // Invalid use of simple_append
+}
+
+template<typename OutCharT, typename ResultCharT>
+void simple_append(std::basic_string<ResultCharT>& out_string, char8_t in_character) {
+	if constexpr (sizeof(OutCharT) == sizeof(ResultCharT)) {
+		out_string += in_character;
+	}
+	else if constexpr (std::is_same_v<ResultCharT, char>) {
+		// Copy in_character into result _as if_ result were of OutCharT
+		OutCharT codepoint = in_character;
+		// TODO: Assuming native for now, but we need to account for endianness later
+		out_string.append(reinterpret_cast<const char*>(&codepoint), sizeof(codepoint));
+	}
+}
+
 } // namespace jessilib

src/include/jessilib/parsers/json.hpp

@@ -18,6 +18,7 @@
 
 #pragma once
 
+#include "fmt/xchar.h" // fmt::format
 #include "jessilib/parser.hpp"
 #include "jessilib/unicode.hpp" // join
 #include "jessilib/unicode_syntax.hpp" // syntax trees
@@ -29,8 +30,23 @@ namespace jessilib {
 class json_parser : public parser {
 public:
 	/** deserialize/serialize overrides */
-	virtual object deserialize(std::u8string_view in_data) override;
-	virtual std::u8string serialize(const object& in_object) override;
+	object deserialize_bytes(bytes_view_type in_data, encoding in_write_encoding) override;
+	std::string serialize_bytes(const object& in_object, encoding in_write_encoding) override;
+
+	std::u8string serialize_u8(const object& in_object) override { return serialize_impl<char8_t>(in_object); }
+	std::u16string serialize_u16(const object& in_object) override { return serialize_impl<char16_t>(in_object); }
+	std::u32string serialize_u32(const object& in_object) override { return serialize_impl<char32_t>(in_object); }
+	std::wstring serialize_w(const object& in_object) override { return serialize_impl<wchar_t>(in_object); }
+
+	template<typename CharT, typename ResultCharT = CharT>
+	std::basic_string<ResultCharT> serialize_impl(const object& in_object) {
+		std::basic_string<ResultCharT> result;
+		serialize_impl<CharT, ResultCharT>(result, in_object);
+		return result;
+	}
+
+	template<typename CharT, typename ResultCharT = CharT>
+	void serialize_impl(std::basic_string<ResultCharT>& out_string, const object& in_object);
 };
 
 /**
@@ -451,4 +467,165 @@ bool deserialize_json(object& out_object, std::basic_string_view<CharT>& inout_r
 		(context, inout_read_view);
 }
 
+template<typename CharT, typename ResultCharT>
+void make_json_string(std::basic_string<ResultCharT>& out_string, std::u8string_view in_string) {
+	using namespace std::literals;
+	out_string.reserve(out_string.size() + in_string.size() + 2);
+	simple_append<CharT, ResultCharT>(out_string, '\"');
+
+	decode_result decode;
+	while ((decode = decode_codepoint(in_string)).units != 0) {
+		if (decode.codepoint == U'\\') { // backslash
+			simple_append<CharT, ResultCharT>(out_string, u8"\\\\"sv);
+		}
+		else if (decode.codepoint == U'\"') { // quotation
+			simple_append<CharT, ResultCharT>(out_string, u8"\\\""sv);
+		}
+		else if (decode.codepoint < 0x20) { // control characters
+			simple_append<CharT, ResultCharT>(out_string, u8"\\u00"sv);
+
+			// overwrite last 2 zeroes with correct hexadecimal sequence
+			char data[2]; // Will only ever use 2 chars
+			char* data_end = data + sizeof(data);
+			auto to_chars_result = std::to_chars(data, data_end, static_cast<uint32_t>(decode.codepoint), 16);
+			if (to_chars_result.ptr == data) {
+				// No bytes written
+				simple_append<CharT, ResultCharT>(out_string, u8"00"sv);
+			}
+			else if (to_chars_result.ptr != data_end) {
+				// 1 byte written
+				simple_append<CharT, ResultCharT>(out_string, '0');
+				simple_append<CharT, ResultCharT>(out_string, data[0]);
+			}
+			else {
+				// 2 bytes written
+				simple_append<CharT, ResultCharT>(out_string, std::u8string_view{ reinterpret_cast<char8_t*>(data), sizeof(data) });
+			}
+		}
+		else {
+			if constexpr (sizeof(CharT) == sizeof(char8_t) && sizeof(CharT) == sizeof(ResultCharT)) {
+				// Valid UTF-8 sequence; copy it over
+				out_string.append(reinterpret_cast<const ResultCharT*>(in_string.data()), decode.units);
+			}
+			else if constexpr (sizeof(CharT) == sizeof(ResultCharT)){
+				// Valid UTF-8 codepoint; append it
+				encode_codepoint(out_string, decode.codepoint);
+			}
+			else {
+				// Valid UTF-8 codepoint; encode & append it
+				encode_buffer_type<CharT> buffer;
+				size_t units_written = encode_codepoint(buffer, decode.codepoint);
+				out_string.append(reinterpret_cast<ResultCharT*>(buffer), units_written * sizeof(CharT));
+			}
+		}
+
+		in_string.remove_prefix(decode.units);
+	}
+
+	simple_append<CharT, ResultCharT>(out_string, '\"');
+}
+
+template<typename CharT>
+static constexpr CharT empty_format_arg[3]{ '{', '}', 0 };
+
+template<typename CharT, typename ResultCharT>
+void json_parser::serialize_impl(std::basic_string<ResultCharT>& out_string, const object& in_object) {
+	using namespace std::literals;
+	static const object::array_type s_null_array;
+	static const object::map_type s_null_map;
+
+	switch (in_object.type()) {
+		case object::type::null:
+			simple_append<CharT, ResultCharT>(out_string, u8"null"sv);
+			return;
+
+		case object::type::boolean:
+			if (in_object.get<bool>()) {
+				simple_append<CharT, ResultCharT>(out_string, u8"true"sv);
+				return;
+			}
+			simple_append<CharT, ResultCharT>(out_string, u8"false"sv);
+			return;
+
+		case object::type::integer:
+			if constexpr (sizeof(CharT) == sizeof(ResultCharT)) {
+				out_string += fmt::format(empty_format_arg<ResultCharT>, in_object.get<intmax_t>());
+			}
+			else if constexpr (std::is_same_v<ResultCharT, char>){
+				auto encoded = fmt::format(empty_format_arg<CharT>, in_object.get<intmax_t>());
+				out_string.append(reinterpret_cast<ResultCharT*>(encoded.data()), encoded.size() * sizeof(CharT));
+			}
+			return;
+
+		case object::type::decimal:
+			if constexpr (sizeof(CharT) == sizeof(ResultCharT)) {
+				out_string += fmt::format(empty_format_arg<ResultCharT>, in_object.get<long double>());
+			}
+			else if constexpr (std::is_same_v<ResultCharT, char>){
+				auto encoded = fmt::format(empty_format_arg<CharT>, in_object.get<long double>());
+				out_string.append(reinterpret_cast<ResultCharT*>(encoded.data()), encoded.size() * sizeof(CharT));
+			}
+			return;
+
+		case object::type::text:
+			make_json_string<CharT, ResultCharT>(out_string, in_object.get<std::u8string>());
+			return;
+
+		case object::type::array: {
+			if (in_object.size() == 0) {
+				simple_append<CharT, ResultCharT>(out_string, u8"[]"sv);
+			}
+
+			simple_append<CharT, ResultCharT>(out_string, '[');
+
+			// Serialize all objects in array
+			for (auto& obj : in_object.get<object::array_type>(s_null_array)) {
+				json_parser::serialize_impl<CharT, ResultCharT>(out_string, obj);
+				simple_append<CharT, ResultCharT>(out_string, ',');
+			}
+
+			// Replace last comma with ']'
+			if constexpr (sizeof(CharT) == sizeof(ResultCharT)) {
+				out_string.back() = ']';
+			}
+			else if constexpr (std::is_same_v<ResultCharT, char>) {
+				out_string.erase(out_string.size() - sizeof(CharT));
+				simple_append<CharT, ResultCharT>(out_string, ']');
+			}
+			// else // not supported
+			return;
+		}
+
+		case object::type::map: {
+			if (in_object.size() == 0) {
+				simple_append<CharT, ResultCharT>(out_string, u8"{}"sv);
+			}
+
+			simple_append<CharT, ResultCharT>(out_string, '{');
+
+			// Serialize all objects in map
+			for (auto& item : in_object.get<object::map_type>(s_null_map)) {
+				make_json_string<CharT, ResultCharT>(out_string, item.first);
+				simple_append<CharT, ResultCharT>(out_string, ':');
+				json_parser::serialize_impl<CharT, ResultCharT>(out_string, item.second);
+				simple_append<CharT, ResultCharT>(out_string, ',');
+			}
+
+			// Replace last comma with '}'
+			if constexpr (sizeof(CharT) == sizeof(ResultCharT)) {
+				out_string.back() = '}';
+			}
+			else if constexpr (std::is_same_v<ResultCharT, char>) {
+				out_string.erase(out_string.size() - sizeof(CharT));
+				simple_append<CharT, ResultCharT>(out_string, '}');
+			}
+			// else // not supported
+			return;
+		}
+
+		default:
+			throw std::invalid_argument{ "Invalid data type: " + std::to_string(static_cast<size_t>(in_object.type())) };
+	}
+}
+
 } // namespace jessilib

src/include/jessilib/serialize.hpp

@@ -29,13 +29,15 @@ public:
 };
 
 /** Deserialization */
-object deserialize_object(const std::u8string& in_data, const std::string& in_format);
-object deserialize_object(const std::vector<char8_t>& in_data, const std::string& in_format);
 object deserialize_object(std::u8string_view in_data, const std::string& in_format);
-object deserialize_object(std::istream& in_stream, const std::string& in_format);
+object deserialize_object(std::u16string_view in_data, const std::string& in_format);
+object deserialize_object(std::u32string_view in_data, const std::string& in_format);
+object deserialize_object(const std::vector<char8_t>& in_data, const std::string& in_format);
+//object deserialize_object(std::u8string_view in_data, const std::string& in_format);
+object deserialize_object(std::istream& in_stream, const std::string& in_format); // TODO: add encoding param
 
 /** Serialization */
-std::u8string serialize_object(const object& in_object, const std::string& in_format);
-void serialize_object(std::ostream& in_stream, const object& in_object, const std::string& in_format);
+std::u8string serialize_object(const object& in_object, const std::string& in_format); // TODO: templatize?
+void serialize_object(std::ostream& in_stream, const object& in_object, const std::string& in_format); // TODO: add encoding param
 
 } // namespace jessilib

src/include/jessilib/unicode_base.hpp

@@ -169,6 +169,67 @@ struct unicode_traits<wchar_t> : std::true_type {
 template<typename CharT>
 using encode_buffer_type = CharT[unicode_traits<CharT>::max_units_per_codepoint];
 
+// enum representing the character encodings I intend to support
+enum class encoding {
+	utf_8, // The most common and arguably superior encoding for files and networking protocols not in straight ASCII
+	utf_16,
+	utf_32,
+	wchar, // essentially only really for std::wcout / std::wcout
+	multibyte // essentially only really for std::cout / std::cin
+};
+
+template<encoding EncodingV>
+struct encoding_info;
+
+template<>
+struct encoding_info<encoding::utf_8> {
+	using data_type = char8_t;
+	static constexpr encoding text_encoding = encoding::utf_8;
+};
+
+template<>
+struct encoding_info<encoding::utf_16> {
+	using data_type = char16_t;
+	static constexpr encoding text_encoding = encoding::utf_16;
+};
+
+template<>
+struct encoding_info<encoding::utf_32> {
+	using data_type = char32_t;
+	static constexpr encoding text_encoding = encoding::utf_32;
+};
+
+template<>
+struct encoding_info<encoding::wchar> {
+	using data_type = wchar_t;
+	static constexpr encoding text_encoding = encoding::wchar;
+};
+
+template<>
+struct encoding_info<encoding::multibyte> {
+	using data_type = char;
+	static constexpr encoding text_encoding = encoding::multibyte;
+};
+
+template<typename CharT>
+struct default_encoding_info;
+
+template<>
+struct default_encoding_info<char8_t> : public encoding_info<encoding::utf_8> {
+};
+
+template<>
+struct default_encoding_info<char16_t> : public encoding_info<encoding::utf_16> {
+};
+
+template<>
+struct default_encoding_info<char32_t> : public encoding_info<encoding::utf_32> {
+};
+
+template<>
+struct default_encoding_info<wchar_t> : public encoding_info<encoding::wchar> {
+};
+
 /** single-unit helper utilities */
 char32_t fold(char32_t in_codepoint); // Folds codepoint for case-insensitive checks (not for human output)
 constexpr int as_base(char32_t in_character, unsigned int base); // The value represented by in_character in terms of base if valid, -1 otherwise

src/test/parser.cpp

@@ -30,14 +30,63 @@ using namespace std::literals;
 class test_parser : public parser {
 public:
 	/** deserialize/serialize overrides */
-	virtual object deserialize(std::u8string_view in_data) override {
-		return deserialize_impl(in_data);
+	object deserialize_bytes(bytes_view_type in_data, encoding in_write_encoding) override {
+		std::u8string u8_string;
+
+		switch (in_write_encoding) {
+			case encoding::utf_8:
+				u8_string = string_view_cast<char8_t>(in_data);
+				break;
+			case encoding::utf_16:
+				u8_string = jessilib::string_cast<char8_t>(string_view_cast<char16_t>(in_data));
+				break;
+			case encoding::utf_32:
+				u8_string = jessilib::string_cast<char8_t>(string_view_cast<char32_t>(in_data));
+				break;
+			case encoding::wchar:
+				u8_string = jessilib::string_cast<char8_t>(string_view_cast<wchar_t>(in_data));
+				break;
+			case encoding::multibyte:
+				u8_string = mbstring_to_ustring<char8_t>(string_view_cast<char>(in_data)).second;
+				break;
+		}
+
+		return deserialize_impl(std::u8string_view{ u8_string });
 	}
 
-	virtual std::u8string serialize(const object& in_object) override {
+	std::string serialize_bytes(const object& in_object, encoding in_write_encoding) override {
+		std::u8string u8_serialized = serialize_impl(in_object);
+
+		switch (in_write_encoding) {
+			case encoding::utf_8:
+				return { u8_serialized.begin(), u8_serialized.end() };
+			case encoding::utf_16: {
+				auto casted = string_cast<char16_t>(u8_serialized);
+				return { reinterpret_cast<const char*>(casted.data()), casted.size() * sizeof(char16_t) };
+			}
+			case encoding::utf_32: {
+				auto casted = string_cast<char32_t>(u8_serialized);
+				return { reinterpret_cast<const char*>(casted.data()), casted.size() * sizeof(char32_t) };
+			}
+			case encoding::wchar: {
+				auto casted = string_cast<wchar_t>(u8_serialized);
+				return { reinterpret_cast<const char*>(casted.data()), casted.size() * sizeof(wchar_t) };
+			}
+			case encoding::multibyte:
+				return ustring_to_mbstring(u8_serialized).second;
+		}
+
+		return {};
+	}
+
+	virtual std::u8string serialize_u8(const object& in_object) override {
 		return serialize_impl(in_object);
 	}
 
+	std::u16string serialize_u16(const object& in_object) override { return string_cast<char16_t>(serialize_u8(in_object)); }
+	std::u32string serialize_u32(const object& in_object) override { return string_cast<char32_t>(serialize_u8(in_object)); }
+	std::wstring serialize_w(const object& in_object) override { return string_cast<wchar_t>(serialize_u8(in_object)); }
+
 	/** helpers */
 	static void reset() {
 		serialize_impl = &serialize_default;
@@ -54,7 +103,7 @@ public:
 	}
 
 	static object deserialize_default(std::u8string_view in_data) {
-		return object{ string_view_cast<char8_t>(in_data) };
+		return object{ in_data };
 	}
 
 	/** static members */

src/test/parsers/json.cpp

@@ -25,26 +25,26 @@ using namespace std::literals;
 TEST(JsonParser, serialize_null) {
 	json_parser parser;
 
-	EXPECT_EQ(parser.serialize({}), u8"null");
+	EXPECT_EQ(parser.serialize<char8_t>({}), u8"null");
 }
 
 TEST(JsonParser, serialize_boolean) {
 	json_parser parser;
 
-	EXPECT_EQ(parser.serialize(true), u8"true");
-	EXPECT_EQ(parser.serialize(false), u8"false");
+	EXPECT_EQ(parser.serialize<char8_t>(true), u8"true");
+	EXPECT_EQ(parser.serialize<char8_t>(false), u8"false");
 }
 
 TEST(JsonParser, serialize_integer) {
 	json_parser parser;
 
-	EXPECT_EQ(parser.serialize(1234), u8"1234");
+	EXPECT_EQ(parser.serialize<char8_t>(1234), u8"1234");
 }
 
 TEST(JsonParser, serialize_decimal) {
 	json_parser parser;
-	EXPECT_DOUBLE_EQ(std::atof(reinterpret_cast<const char*>(parser.serialize(12.34).c_str())), 12.34);
-	EXPECT_DOUBLE_EQ(std::atof(reinterpret_cast<const char*>(parser.serialize(1234.0).c_str())), 1234.0);
+	EXPECT_DOUBLE_EQ(std::atof(reinterpret_cast<const char*>(parser.serialize<char8_t>(12.34).c_str())), 12.34);
+	EXPECT_DOUBLE_EQ(std::atof(reinterpret_cast<const char*>(parser.serialize<char8_t>(1234.0).c_str())), 1234.0);
 }
 
 // necessary due to some sort of bug with EXPECT_EQ on MSVC
@@ -56,10 +56,15 @@ void expect_eq(LeftT in_left, RightT in_right) {
 TEST(JsonParser, serialize_string) {
 	json_parser parser;
 
-	EXPECT_EQ(parser.serialize(u8"text"), u8R"json("text")json");
-	expect_eq(parser.serialize(u8"\"text\""), u8R"json("\"text\"")json");
-	expect_eq(parser.serialize(u8"\"te\x01xt\""), u8R"json("\"te\u0001xt\"")json");
-	expect_eq(parser.serialize(u8"\"te\x10xt\""), u8R"json("\"te\u0010xt\"")json");
+	EXPECT_EQ(parser.serialize<char8_t>(u8"text"), u8R"json("text")json");
+	expect_eq(parser.serialize<char8_t>(u8"\"text\""), u8R"json("\"text\"")json");
+	expect_eq(parser.serialize<char8_t>(u8"\"te\x01xt\""), u8R"json("\"te\u0001xt\"")json");
+	expect_eq(parser.serialize<char8_t>(u8"\"te\x10xt\""), u8R"json("\"te\u0010xt\"")json");
+
+	EXPECT_EQ(parser.serialize<char8_t>(u8"text"), u8R"json("text")json");
+	EXPECT_EQ(parser.serialize<char16_t>(u8"text"), uR"json("text")json");
+	EXPECT_EQ(parser.serialize<char32_t>(u8"text"), UR"json("text")json");
+	EXPECT_EQ(parser.serialize<wchar_t>(u8"text"), LR"json("text")json");
 }
 
 TEST(JsonParser, serialize_array) {
@@ -71,7 +76,7 @@ TEST(JsonParser, serialize_array) {
 		object{}
 	};
 
-	EXPECT_EQ(parser.serialize(array),
+	EXPECT_EQ(parser.serialize<char8_t>(array),
 		u8R"json([true,1234,"text",null])json");
 }
 
@@ -84,7 +89,7 @@ TEST(JsonParser, serialize_map) {
 	obj[u8"some_string"] = u8"text";
 	obj[u8"some_null"];
 
-	EXPECT_EQ(parser.serialize(obj),
+	EXPECT_EQ(parser.serialize<char8_t>(obj),
 		u8R"json({"some_bool":true,"some_int":1234,"some_null":null,"some_string":"text"})json");
 }
 
@@ -143,14 +148,14 @@ TEST(JsonParser, deserialize_string) {
 TEST(JsonParser, deserialize_array) {
 	json_parser parser;
 
-	constexpr const char8_t* json_data = u8R"json([
+	constexpr std::u8string_view json_data = u8R"json([
 		true,
 		false,
 		1234,
 		12.34,
 		0.1234,
 		"text"
-	])json";
+	])json"sv;
 
 	auto array = parser.deserialize(json_data).get<std::vector<object>>();
 	ASSERT_EQ(array.size(), 6U);
@@ -165,7 +170,7 @@ TEST(JsonParser, deserialize_array) {
 TEST(JsonParser, deserialize_array_nested) {
 	json_parser parser;
 
-	constexpr const char8_t* json_data = u8R"json([
+	constexpr std::u8string_view json_data = u8R"json([
 		true,
 		false,
 		1234
@@ -178,7 +183,7 @@ TEST(JsonParser, deserialize_array_nested) {
 		12.34,
 		0.1234,
 		"text"
-	])json";
+	])json"sv;
 
 	auto array = parser.deserialize(json_data).get<std::vector<object>>();
 	ASSERT_EQ(array.size(), 9U);
@@ -211,14 +216,14 @@ TEST(JsonParser, deserialize_array_nested) {
 TEST(JsonParser, deserialize_map) {
 	json_parser parser;
 
-	constexpr const char8_t* json_data = u8R"json({
+	constexpr std::u8string_view json_data = u8R"json({
 		"some_true":true,
 		"some_false"	:  false,
 		"some_int": 1234,
 		"some_double" : 12.34,
 		"some_other_double" :0.1234,
 		"some_text" : "text"
-	})json";
+	})json"sv;
 
 	object obj = parser.deserialize(json_data);
 	EXPECT_EQ(obj.size(), 6U);
@@ -233,7 +238,7 @@ TEST(JsonParser, deserialize_map) {
 TEST(JsonParser, deserialize_map_nested) {
 	json_parser parser;
 
-	constexpr const char8_t* json_data = u8R"json({
+	constexpr std::u8string_view json_data = u8R"json({
 		"some_text" : "text",
 		"some_object" : {
 			"some_null_object": {}
@@ -246,7 +251,7 @@ TEST(JsonParser, deserialize_map_nested) {
 				"makes toot": true
 			}},
 		"some other text":" asdf "
-	})json";
+	})json"sv;
 
 	object obj = parser.deserialize(json_data);
 	EXPECT_EQ(obj.size(), 4U);