Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion data/include/join/json.hpp
Original file line number Diff line number Diff line change
Expand Up @@ -880,7 +880,7 @@ namespace join
* @param data data to append to output stream.
* @param size data size.
*/
inline void append (const char* data, uint32_t size)
inline void append (const char* data, size_t size)
{
if (JOIN_UNLIKELY (size > sizeof (_buffer) - _size))
{
Expand Down
137 changes: 130 additions & 7 deletions data/include/join/pack.hpp
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,7 @@
#include <join/value_builder.hpp>
#include <join/value.hpp>
#include <join/utils.hpp>
#include <join/utf8.hpp>
#include <join/sax.hpp>

// C++.
Expand Down Expand Up @@ -691,47 +692,134 @@ namespace join
size_t _depth = 0;
};

/**
* @brief MessagePack deserialization mode.
*/
enum class PackReadMode : uint32_t
{
None = 0, /**< no read mode set. */
ValidateEncoding = 1L << 0, /**< validate UTF-8 encoding of strings. */
StopParsingOnDone = 1L << 1, /**< stop parsing on done. */
};

/**
* @brief perform binary AND on PackReadMode.
* @param a bitset.
* @param b other bitset.
* @return bitset result of binary AND on PackReadMode.
*/
constexpr PackReadMode operator& (PackReadMode a, PackReadMode b) noexcept
{
return PackReadMode (static_cast<uint32_t> (a) & static_cast<uint32_t> (b));
}

/**
* @brief perform binary OR on PackReadMode.
* @param a bitset.
* @param b other bitset.
* @return bitset result of binary OR on PackReadMode.
*/
constexpr PackReadMode operator| (PackReadMode a, PackReadMode b) noexcept
{
return PackReadMode (static_cast<uint32_t> (a) | static_cast<uint32_t> (b));
}

/**
* @brief perform binary XOR on PackReadMode.
* @param a bitset.
* @param b other bitset.
* @return bitset result of binary XOR on PackReadMode.
*/
constexpr PackReadMode operator^ (PackReadMode a, PackReadMode b) noexcept
{
return PackReadMode (static_cast<uint32_t> (a) ^ static_cast<uint32_t> (b));
}

/**
* @brief perform binary NOT on PackReadMode.
* @param a bitset.
* @return bitset result of binary NOT on PackReadMode.
*/
constexpr PackReadMode operator~(PackReadMode a) noexcept
{
return PackReadMode (~static_cast<uint32_t> (a));
}

/**
* @brief perform binary AND assignment on PackReadMode.
* @param a bitset.
* @param b other bitset.
* @return reference to updated bitset.
*/
constexpr const PackReadMode& operator&= (PackReadMode& a, PackReadMode b) noexcept
{
return a = a & b;
}

/**
* @brief perform binary OR assignment on PackReadMode.
* @param a bitset.
* @param b other bitset.
* @return reference to updated bitset.
*/
constexpr const PackReadMode& operator|= (PackReadMode& a, PackReadMode b) noexcept
{
return a = a | b;
}

/**
* @brief perform binary XOR assignment on PackReadMode.
* @param a bitset.
* @param b other bitset.
* @return reference to updated bitset.
*/
constexpr const PackReadMode& operator^= (PackReadMode& a, PackReadMode b) noexcept
{
return a = a ^ b;
}

/**
* @brief message pack reader class.
*/
class PackReader
template <PackReadMode ReadMode = PackReadMode::None>
class BasicPackReader
{
public:
/**
* @brief default constructor.
*/
PackReader () = default;
BasicPackReader () = default;

/**
* @brief copy constructor.
* @param other object to copy.
*/
PackReader (const PackReader& other) = delete;
BasicPackReader (const BasicPackReader& other) = delete;

/**
* @brief copy assignment.
* @param other object to copy.
* @return a reference of the current object.
*/
PackReader& operator= (const PackReader& other) = delete;
BasicPackReader& operator= (const BasicPackReader& other) = delete;

/**
* @brief move constructor.
* @param other object to move.
*/
PackReader (PackReader&& other) = delete;
BasicPackReader (BasicPackReader&& other) = delete;

/**
* @brief move assignment.
* @param other object to move.
* @return a reference of the current object.
*/
PackReader& operator= (PackReader&& other) = delete;
BasicPackReader& operator= (BasicPackReader&& other) = delete;

/**
* @brief destroy instance.
*/
~PackReader () = default;
~BasicPackReader () = default;

/**
* @brief deserialize a document.
Expand Down Expand Up @@ -818,6 +906,11 @@ namespace join
{
if (JOIN_LIKELY (readValue (document, handler, 0) == 0))
{
if constexpr ((ReadMode & PackReadMode::StopParsingOnDone) != PackReadMode::None)
{
return 0;
}

if (JOIN_LIKELY (document.peek () == std::char_traits<char>::eof ()))
{
return 0;
Expand Down Expand Up @@ -1155,6 +1248,10 @@ namespace join
}

std::string_view output (document.tell (), len);
if (JOIN_UNLIKELY (validateString (output) == -1))
{
return -1;
}
document.seek (document.tell () + len);
return isKey ? handler.setKey (output) : handler.setString (output);
}
Expand All @@ -1167,11 +1264,34 @@ namespace join
return -1;
}

if (JOIN_UNLIKELY (validateString (output) == -1))
{
return -1;
}

int result = isKey ? handler.setKey (output) : handler.setString (output);
cache = std::move (output);
return result;
}

/**
* @brief validate the UTF-8 encoding of a string when required by the read mode.
* @param value string to validate.
* @return 0 on success, -1 otherwise.
*/
static int validateString ([[maybe_unused]] std::string_view value) noexcept
{
if constexpr ((ReadMode & PackReadMode::ValidateEncoding) != PackReadMode::None)
{
if (JOIN_UNLIKELY (utf8::validate (value.data (), value.data () + value.size ()) == -1))
{
join::lastError = make_error_code (SaxErrc::InvalidEncoding);
return -1;
}
}
return 0;
}

/**
* @brief parse binary data.
* @param document document to parse.
Expand Down Expand Up @@ -1581,6 +1701,9 @@ namespace join
return ((c >= 0x80) && (c <= 0x8f)) || (c == 0xde) || (c == 0xdf);
}
};

/// MessagePack reader.
using PackReader = BasicPackReader<PackReadMode::None>;
}

#endif
1 change: 1 addition & 0 deletions data/include/join/sax_error.hpp
Original file line number Diff line number Diff line change
Expand Up @@ -43,6 +43,7 @@ namespace join
InvalidParent, /**< Parent not an array nor an object. */
InvalidKey, /**< Key is invalid. */
InvalidValue, /**< Value is invalid. */
InvalidEncoding, /**< Character encoding is invalid. */
ExtraData, /**< Extra data detected. */
Truncated /**< Data is truncated. */
};
Expand Down
2 changes: 2 additions & 0 deletions data/src/sax_error.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -53,6 +53,8 @@ std::string SaxCategory::message (int code) const
return "key is invalid";
case SaxErrc::InvalidValue:
return "value is invalid";
case SaxErrc::InvalidEncoding:
return "character encoding is invalid";
case SaxErrc::ExtraData:
return "extra data detected";
case SaxErrc::Truncated:
Expand Down
52 changes: 52 additions & 0 deletions data/tests/pack_reader_test.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -209,6 +209,25 @@ TEST (PackReader, deserialize)
builder.startArray ();
}
ASSERT_EQ (PackReader ().deserialize (std::string (1, '\x80'), builder), -1);

using Stopping = join::BasicPackReader<join::PackReadMode::StopParsingOnDone>;
std::stringstream messages (std::string ({'\x01', '\x02', '\xc3'}));
ASSERT_EQ (Stopping ().deserialize (messages, value), 0) << join::lastError.message ();
ASSERT_EQ (value, 1);
ASSERT_EQ (Stopping ().deserialize (messages, value), 0) << join::lastError.message ();
ASSERT_EQ (value, 2);
ASSERT_EQ (Stopping ().deserialize (messages, value), 0) << join::lastError.message ();
ASSERT_EQ (value, true);
ASSERT_EQ (Stopping ().deserialize (messages, value), -1);
ASSERT_EQ (join::lastError, SaxErrc::Truncated);
ASSERT_EQ (Stopping ().deserialize (std::string ({'\x00', '\x00'}), value), 0) << join::lastError.message ();

using Strict = join::BasicPackReader<join::PackReadMode::ValidateEncoding | join::PackReadMode::StopParsingOnDone>;
std::stringstream strict (std::string ({'\xa1', 'a', '\xa1', '\xff'}));
ASSERT_EQ (Strict ().deserialize (strict, value), 0) << join::lastError.message ();
ASSERT_EQ (value.getString (), "a");
ASSERT_EQ (Strict ().deserialize (strict, value), -1);
ASSERT_EQ (join::lastError, SaxErrc::InvalidEncoding);
}

/**
Expand Down Expand Up @@ -1570,6 +1589,39 @@ TEST (PackReader, key)
ASSERT_TRUE (value.contains (-10));
}

/**
* @brief Test MessagePack parsing with UTF-8 validation.
*/
TEST (PackReader, encoding)
{
using Validator = join::BasicPackReader<join::PackReadMode::ValidateEncoding>;

for (const std::string& document : {std::string ({'\xa2', '\xc3', '\xa9'}), std::string ({'\xc4', '\x01', '\xff'})})
{
std::stringstream stream (document);
join::SaxHandler handler;
ASSERT_EQ (Validator ().deserialize (document, handler), 0) << join::lastError.message ();
ASSERT_EQ (Validator ().deserialize (stream, handler), 0) << join::lastError.message ();
}

for (const std::string& document : {std::string ({'\xa1', '\xff'}), std::string ({'\x81', '\xa1', '\xff', '\x01'})})
{
std::stringstream stream (document);
join::SaxHandler handler;
ASSERT_EQ (Validator ().deserialize (document, handler), -1);
ASSERT_EQ (join::lastError, SaxErrc::InvalidEncoding);
ASSERT_EQ (Validator ().deserialize (stream, handler), -1);
ASSERT_EQ (join::lastError, SaxErrc::InvalidEncoding);
}

Value value;
ASSERT_EQ (PackReader ().deserialize (std::string ({'\xa1', '\xff'}), value), 0) << join::lastError.message ();
ASSERT_EQ (value.getString (), "\xff");
ASSERT_EQ (value.deserialize<Validator> (std::string ({'\xa4', 'I', 'E', 'T', 'F'})), 0)
<< join::lastError.message ();
ASSERT_EQ (value.getString (), "IETF");
}

/**
* @brief main function.
*/
Expand Down
5 changes: 4 additions & 1 deletion data/tests/sax_error_test.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -50,6 +50,8 @@ TEST (SaxCategory, message)
"parent not an array nor an object");
EXPECT_STREQ (SaxCategory ().message (static_cast<int> (SaxErrc::InvalidKey)).c_str (), "key is invalid");
EXPECT_STREQ (SaxCategory ().message (static_cast<int> (SaxErrc::InvalidValue)).c_str (), "value is invalid");
EXPECT_STREQ (SaxCategory ().message (static_cast<int> (SaxErrc::InvalidEncoding)).c_str (),
"character encoding is invalid");
EXPECT_STREQ (SaxCategory ().message (static_cast<int> (SaxErrc::ExtraData)).c_str (), "extra data detected");
EXPECT_STREQ (SaxCategory ().message (static_cast<int> (SaxErrc::Truncated)).c_str (), "data is truncated");
}
Expand All @@ -64,7 +66,8 @@ TEST (SaxCategory, default_error_condition)
EXPECT_EQ (SaxCategory ().default_error_condition (2).message (), "parent not an array nor an object");
EXPECT_EQ (SaxCategory ().default_error_condition (3).message (), "key is invalid");
EXPECT_EQ (SaxCategory ().default_error_condition (4).message (), "value is invalid");
EXPECT_EQ (SaxCategory ().default_error_condition (5).message (), "extra data detected");
EXPECT_EQ (SaxCategory ().default_error_condition (5).message (), "character encoding is invalid");
EXPECT_EQ (SaxCategory ().default_error_condition (6).message (), "extra data detected");
}

/**
Expand Down
Loading