/*! \file xml.hpp \brief XML input and output archives */ /* Copyright (c) 2014, Randolph Voorhies, Shane Grant All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. * Neither the name of cereal nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL RANDOLPH VOORHIES OR SHANE GRANT BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #ifndef CEREAL_ARCHIVES_XML_HPP_ #define CEREAL_ARCHIVES_XML_HPP_ #include #include #include #include #include #include #include #include #include #include #include #include namespace cereal { namespace xml_detail { #ifndef CEREAL_XML_STRING_VALUE //! The default name for the root node in a cereal xml archive. /*! You can define CEREAL_XML_STRING_VALUE to be different assuming you do so before this file is included. */ #define CEREAL_XML_STRING_VALUE "cereal" #endif // CEREAL_XML_STRING_VALUE //! The name given to the root node in a cereal xml archive static const char * CEREAL_XML_STRING = CEREAL_XML_STRING_VALUE; //! Returns true if the character is whitespace inline bool isWhitespace( char c ) { return c == ' ' || c == '\t' || c == '\n' || c == '\r'; } } // ###################################################################### //! An output archive designed to save data to XML /*! This archive uses RapidXML to build an in memory XML tree of the data it serializes before outputting it to its stream upon destruction. The envisioned way of using this archive is in an RAII fashion, letting the automatic destruction of the object cause the flush to its stream. XML archives provides a human readable output but at decreased performance (both in time and space) compared to binary archives. XML benefits greatly from name-value pairs, which if present, will name the nodes in the output. If these are not present, each level of the output tree will be given an automatically generated delimited name. The precision of the output archive controls the number of decimals output for floating point numbers and should be sufficiently large (i.e. at least 20) if there is a desire to have binary equality between the numbers output and those read in. In general you should expect a loss of precision when going from floating point to text and back. XML archives can optionally print the type of everything they serialize, which adds an attribute to each node. XML archives do not output the size information for any dynamically sized structure and instead infer it from the number of children for a node. This means that data can be hand edited for dynamic sized structures and will still be readable. This is accomplished through the cereal::SizeTag object, which will also add an attribute to its parent field. \ingroup Archives */ class XMLOutputArchive : public OutputArchive, public traits::TextArchive { public: /*! @name Common Functionality Common use cases for directly interacting with an XMLOutputArchive */ //! @{ //! A class containing various advanced options for the XML archive class Options { public: //! Default options static Options Default(){ return Options(); } //! Default options with no indentation static Options NoIndent(){ return Options( std::numeric_limits::max_digits10, false ); } //! Specify specific options for the XMLOutputArchive /*! @param precision The precision used for floating point numbers @param indent Whether to indent each line of XML @param outputType Whether to output the type of each serialized object as an attribute */ explicit Options( int precision = std::numeric_limits::max_digits10, bool indent = true, bool outputType = false ) : itsPrecision( precision ), itsIndent( indent ), itsOutputType( outputType ) { } private: friend class XMLOutputArchive; int itsPrecision; bool itsIndent; bool itsOutputType; }; //! Construct, outputting to the provided stream upon destruction /*! @param stream The stream to output to. Note that XML is only guaranteed to flush its output to the stream upon destruction. @param options The XML specific options to use. See the Options struct for the values of default parameters */ XMLOutputArchive( std::ostream & stream, Options const & options = Options::Default() ) : OutputArchive(this), itsStream(stream), itsOutputType( options.itsOutputType ), itsIndent( options.itsIndent ) { // rapidxml will delete all allocations when xml_document is cleared auto node = itsXML.allocate_node( rapidxml::node_declaration ); node->append_attribute( itsXML.allocate_attribute( "version", "1.0" ) ); node->append_attribute( itsXML.allocate_attribute( "encoding", "utf-8" ) ); itsXML.append_node( node ); // allocate root node auto root = itsXML.allocate_node( rapidxml::node_element, xml_detail::CEREAL_XML_STRING ); itsXML.append_node( root ); itsNodes.emplace( root ); // set attributes on the streams itsStream << std::boolalpha; itsStream.precision( options.itsPrecision ); itsOS << std::boolalpha; itsOS.precision( options.itsPrecision ); } //! Destructor, flushes the XML ~XMLOutputArchive() { const int flags = itsIndent ? 0x0 : rapidxml::print_no_indenting; rapidxml::print( itsStream, itsXML, flags ); itsXML.clear(); } //! Saves some binary data, encoded as a base64 string, with an optional name /*! This can be called directly by users and it will automatically create a child node for the current XML node, populate it with a base64 encoded string, and optionally name it. The node will be finished after it has been populated. */ void saveBinaryValue( const void * data, size_t size, const char * name = nullptr ) { itsNodes.top().name = name; startNode(); auto base64string = base64::encode( reinterpret_cast( data ), size ); saveValue( base64string ); if( itsOutputType ) itsNodes.top().node->append_attribute( itsXML.allocate_attribute( "type", "cereal binary data" ) ); finishNode(); }; //! @} /*! @name Internal Functionality Functionality designed for use by those requiring control over the inner mechanisms of the XMLOutputArchive */ //! @{ //! Creates a new node that is a child of the node at the top of the stack /*! Nodes will be given a name that has either been pre-set by a name value pair, or generated based upon a counter unique to the parent node. If you want to give a node a specific name, use setNextName prior to calling startNode. The node will then be pushed onto the node stack. */ void startNode() { // generate a name for this new node const auto nameString = itsNodes.top().getValueName(); // allocate strings for all of the data in the XML object auto namePtr = itsXML.allocate_string( nameString.data(), nameString.length() + 1 ); // insert into the XML auto node = itsXML.allocate_node( rapidxml::node_element, namePtr, nullptr, nameString.size() ); itsNodes.top().node->append_node( node ); itsNodes.emplace( node ); } //! Designates the most recently added node as finished void finishNode() { itsNodes.pop(); } //! Sets the name for the next node created with startNode void setNextName( const char * name ) { itsNodes.top().name = name; } //! Saves some data, encoded as a string, into the current top level node /*! The data will be be named with the most recent name if one exists, otherwise it will be given some default delimited value that depends upon the parent node */ template inline void saveValue( T const & value ) { itsOS.clear(); itsOS.seekp( 0, std::ios::beg ); itsOS << value << std::ends; const auto strValue = itsOS.str(); // If the first or last character is a whitespace, add xml:space attribute // the string always contains a '\0' added by std::ends, so the last character is at len-2 and an 'empty' // string has a length of 1 or lower const auto len = strValue.length(); if ( len > 1 && ( xml_detail::isWhitespace( strValue[0] ) || xml_detail::isWhitespace( strValue[len - 2] ) ) ) { itsNodes.top().node->append_attribute( itsXML.allocate_attribute( "xml:space", "preserve" ) ); } // allocate strings for all of the data in the XML object auto dataPtr = itsXML.allocate_string( itsOS.str().c_str(), itsOS.str().length() + 1 ); // insert into the XML itsNodes.top().node->append_node( itsXML.allocate_node( rapidxml::node_data, nullptr, dataPtr ) ); } //! Overload for uint8_t prevents them from being serialized as characters void saveValue( uint8_t const & value ) { saveValue( static_cast( value ) ); } //! Overload for int8_t prevents them from being serialized as characters void saveValue( int8_t const & value ) { saveValue( static_cast( value ) ); } //! Causes the type to be appended as an attribute to the most recently made node if output type is set to true template inline void insertType() { if( !itsOutputType ) return; // generate a name for this new node const auto nameString = util::demangledName(); // allocate strings for all of the data in the XML object auto namePtr = itsXML.allocate_string( nameString.data(), nameString.length() + 1 ); itsNodes.top().node->append_attribute( itsXML.allocate_attribute( "type", namePtr ) ); } //! Appends an attribute to the current top level node void appendAttribute( const char * name, const char * value ) { auto namePtr = itsXML.allocate_string( name ); auto valuePtr = itsXML.allocate_string( value ); itsNodes.top().node->append_attribute( itsXML.allocate_attribute( namePtr, valuePtr ) ); } protected: //! A struct that contains metadata about a node struct NodeInfo { NodeInfo( rapidxml::xml_node<> * n = nullptr, const char * nm = nullptr ) : node( n ), counter( 0 ), name( nm ) { } rapidxml::xml_node<> * node; //!< A pointer to this node size_t counter; //!< The counter for naming child nodes const char * name; //!< The name for the next child node //! Gets the name for the next child node created from this node /*! The name will be automatically generated using the counter if a name has not been previously set. If a name has been previously set, that name will be returned only once */ std::string getValueName() { if( name ) { auto n = name; name = nullptr; return {n}; } else return "value" + std::to_string( counter++ ) + "\0"; } }; // NodeInfo //! @} private: std::ostream & itsStream; //!< The output stream rapidxml::xml_document<> itsXML; //!< The XML document std::stack itsNodes; //!< A stack of nodes added to the document std::ostringstream itsOS; //!< Used to format strings internally bool itsOutputType; //!< Controls whether type information is printed bool itsIndent; //!< Controls whether indenting is used }; // XMLOutputArchive // ###################################################################### //! An output archive designed to load data from XML /*! This archive uses RapidXML to build an in memory XML tree of the data in the stream it is given before loading any types serialized. Input XML should have been produced by the XMLOutputArchive. Data can only be added to dynamically sized containers - the input archive will determine their size by looking at the number of child nodes. Data that did not originate from an XMLOutputArchive is not officially supported, but may be possible to use if properly formatted. The XMLInputArchive does not require that nodes are loaded in the same order they were saved by XMLOutputArchive. Using name value pairs (NVPs), it is possible to load in an out of order fashion or otherwise skip/select specific nodes to load. The default behavior of the input archive is to read sequentially starting with the first node and exploring its children. When a given NVP does not match the read in name for a node, the archive will search for that node at the current level and load it if it exists. After loading an out of order node, the archive will then proceed back to loading sequentially from its new position. Consider this simple example where loading of some data is skipped: @code{cpp} // imagine the input file has someData(1-9) saved in order at the top level node ar( someData1, someData2, someData3 ); // XML loads in the order it sees in the file ar( cereal::make_nvp( "hello", someData6 ) ); // NVP given does not // match expected NVP name, so we search // for the given NVP and load that value ar( someData7, someData8, someData9 ); // with no NVP given, loading resumes at its // current location, proceeding sequentially @endcode \ingroup Archives */ class XMLInputArchive : public InputArchive, public traits::TextArchive { public: /*! @name Common Functionality Common use cases for directly interacting with an XMLInputArchive */ //! @{ //! Construct, reading in from the provided stream /*! Reads in an entire XML document from some stream and parses it as soon as serialization starts @param stream The stream to read from. Can be a stringstream or a file. */ XMLInputArchive( std::istream & stream ) : InputArchive( this ), itsData( std::istreambuf_iterator( stream ), std::istreambuf_iterator() ) { try { itsData.push_back('\0'); // rapidxml will do terrible things without the data being null terminated itsXML.parse( reinterpret_cast( itsData.data() ) ); } catch( rapidxml::parse_error const & ) { //std::cerr << "-----Original-----" << std::endl; //stream.seekg(0); //std::cout << std::string( std::istreambuf_iterator( stream ), std::istreambuf_iterator() ) << std::endl; //std::cerr << "-----Error-----" << std::endl; //std::cerr << e.what() << std::endl; //std::cerr << e.where() << std::endl; throw Exception("XML Parsing failed - likely due to invalid characters or invalid naming"); } // Parse the root auto root = itsXML.first_node( xml_detail::CEREAL_XML_STRING ); if( root == nullptr ) throw Exception("Could not detect cereal root node - likely due to empty or invalid input"); else itsNodes.emplace( root ); } //! Loads some binary data, encoded as a base64 string, optionally specified by some name /*! This will automatically start and finish a node to load the data, and can be called directly by users. Note that this follows the same ordering rules specified in the class description in regards to loading in/out of order */ void loadBinaryValue( void * data, size_t size, const char * name = nullptr ) { setNextName( name ); startNode(); std::string encoded; loadValue( encoded ); auto decoded = base64::decode( encoded ); if( size != decoded.size() ) throw Exception("Decoded binary data size does not match specified size"); std::memcpy( data, decoded.data(), decoded.size() ); finishNode(); }; //! @} /*! @name Internal Functionality Functionality designed for use by those requiring control over the inner mechanisms of the XMLInputArchive */ //! @{ //! Prepares to start reading the next node /*! This places the next node to be parsed onto the nodes stack. By default our strategy is to start with the document root node and then recursively iterate through all children in the order they show up in the document. We don't need to know NVPs do to this; we'll just blindly load in the order things appear in. We check to see if the specified NVP matches what the next automatically loaded node is. If they match, we just continue as normal, going in order. If they don't match, we attempt to find a node named after the NVP that is being loaded. If that NVP does not exist, we throw an exception. */ void startNode() { auto next = itsNodes.top().child; // By default we would move to the next child node auto const expectedName = itsNodes.top().name; // this is the expected name from the NVP, if provided // If we were given an NVP name, look for it in the current level of the document. // We only need to do this if either we have exhausted the siblings of the current level or // the NVP name does not match the name of the node we would normally read next if( expectedName && ( next == nullptr || std::strcmp( next->name(), expectedName ) != 0 ) ) { next = itsNodes.top().search( expectedName ); if( next == nullptr ) throw Exception("XML Parsing failed - provided NVP not found"); } itsNodes.emplace( next ); } //! Finishes reading the current node void finishNode() { // remove current itsNodes.pop(); // advance parent itsNodes.top().advance(); // Reset name itsNodes.top().name = nullptr; } //! Retrieves the current node name //! will return @c nullptr if the node does not have a name const char * getNodeName() const { return itsNodes.top().node->name(); } //! Sets the name for the next node created with startNode void setNextName( const char * name ) { itsNodes.top().name = name; } //! Loads a bool from the current top node template ::value, std::is_same::value> = traits::sfinae> inline void loadValue( T & value ) { std::istringstream is( itsNodes.top().node->value() ); is.setf( std::ios::boolalpha ); is >> value; } //! Loads a char (signed or unsigned) from the current top node template ::value, !std::is_same::value, sizeof(T) == sizeof(char)> = traits::sfinae> inline void loadValue( T & value ) { value = *reinterpret_cast( itsNodes.top().node->value() ); } //! Load an int8_t from the current top node (ensures we parse entire number) void loadValue( int8_t & value ) { int32_t val; loadValue( val ); value = static_cast( val ); } //! Load a uint8_t from the current top node (ensures we parse entire number) void loadValue( uint8_t & value ) { uint32_t val; loadValue( val ); value = static_cast( val ); } //! Loads a type best represented as an unsigned long from the current top node template ::value, !std::is_same::value, !std::is_same::value, sizeof(T) < sizeof(long long)> = traits::sfinae> inline void loadValue( T & value ) { value = static_cast( std::stoul( itsNodes.top().node->value() ) ); } //! Loads a type best represented as an unsigned long long from the current top node template ::value, !std::is_same::value, sizeof(T) >= sizeof(long long)> = traits::sfinae> inline void loadValue( T & value ) { value = static_cast( std::stoull( itsNodes.top().node->value() ) ); } //! Loads a type best represented as an int from the current top node template ::value, !std::is_same::value, sizeof(T) <= sizeof(int)> = traits::sfinae> inline void loadValue( T & value ) { value = static_cast( std::stoi( itsNodes.top().node->value() ) ); } //! Loads a type best represented as a long from the current top node template ::value, (sizeof(T) > sizeof(int)), sizeof(T) <= sizeof(long)> = traits::sfinae> inline void loadValue( T & value ) { value = static_cast( std::stol( itsNodes.top().node->value() ) ); } //! Loads a type best represented as a long long from the current top node template ::value, (sizeof(T) > sizeof(long)), sizeof(T) <= sizeof(long long)> = traits::sfinae> inline void loadValue( T & value ) { value = static_cast( std::stoll( itsNodes.top().node->value() ) ); } //! Loads a type best represented as a float from the current top node void loadValue( float & value ) { try { value = std::stof( itsNodes.top().node->value() ); } catch( std::out_of_range const & ) { // special case for denormalized values std::istringstream is( itsNodes.top().node->value() ); is >> value; if( std::fpclassify( value ) != FP_SUBNORMAL ) throw; } } //! Loads a type best represented as a double from the current top node void loadValue( double & value ) { try { value = std::stod( itsNodes.top().node->value() ); } catch( std::out_of_range const & ) { // special case for denormalized values std::istringstream is( itsNodes.top().node->value() ); is >> value; if( std::fpclassify( value ) != FP_SUBNORMAL ) throw; } } //! Loads a type best represented as a long double from the current top node void loadValue( long double & value ) { try { value = std::stold( itsNodes.top().node->value() ); } catch( std::out_of_range const & ) { // special case for denormalized values std::istringstream is( itsNodes.top().node->value() ); is >> value; if( std::fpclassify( value ) != FP_SUBNORMAL ) throw; } } //! Loads a string from the current node from the current top node template inline void loadValue( std::basic_string & str ) { std::basic_istringstream is( itsNodes.top().node->value() ); str.assign( std::istreambuf_iterator( is ), std::istreambuf_iterator() ); } //! Loads the size of the current top node template inline void loadSize( T & value ) { value = getNumChildren( itsNodes.top().node ); } protected: //! Gets the number of children (usually interpreted as size) for the specified node static size_t getNumChildren( rapidxml::xml_node<> * node ) { size_t size = 0; node = node->first_node(); // get first child while( node != nullptr ) { ++size; node = node->next_sibling(); } return size; } //! A struct that contains metadata about a node /*! Keeps track of some top level node, its number of remaining children, and the current active child node */ struct NodeInfo { NodeInfo( rapidxml::xml_node<> * n = nullptr ) : node( n ), child( n->first_node() ), size( XMLInputArchive::getNumChildren( n ) ), name( nullptr ) { } //! Advances to the next sibling node of the child /*! If this is the last sibling child will be null after calling */ void advance() { if( size > 0 ) { --size; child = child->next_sibling(); } } //! Searches for a child with the given name in this node /*! @param searchName The name to search for (must be null terminated) @return The node if found, nullptr otherwise */ rapidxml::xml_node<> * search( const char * searchName ) { if( searchName ) { size_t new_size = XMLInputArchive::getNumChildren( node ); const size_t name_size = rapidxml::internal::measure( searchName ); for( auto new_child = node->first_node(); new_child != nullptr; new_child = new_child->next_sibling() ) { if( rapidxml::internal::compare( new_child->name(), new_child->name_size(), searchName, name_size, true ) ) { size = new_size; child = new_child; return new_child; } --new_size; } } return nullptr; } rapidxml::xml_node<> * node; //!< A pointer to this node rapidxml::xml_node<> * child; //!< A pointer to its current child size_t size; //!< The remaining number of children for this node const char * name; //!< The NVP name for next next child node }; // NodeInfo //! @} private: std::vector itsData; //!< The raw data loaded rapidxml::xml_document<> itsXML; //!< The XML document std::stack itsNodes; //!< A stack of nodes read from the document }; // ###################################################################### // XMLArchive prologue and epilogue functions // ###################################################################### // ###################################################################### //! Prologue for NVPs for XML output archives /*! NVPs do not start or finish nodes - they just set up the names */ template inline void prologue( XMLOutputArchive &, NameValuePair const & ) { } //! Prologue for NVPs for XML input archives template inline void prologue( XMLInputArchive &, NameValuePair const & ) { } // ###################################################################### //! Epilogue for NVPs for XML output archives /*! NVPs do not start or finish nodes - they just set up the names */ template inline void epilogue( XMLOutputArchive &, NameValuePair const & ) { } //! Epilogue for NVPs for XML input archives template inline void epilogue( XMLInputArchive &, NameValuePair const & ) { } // ###################################################################### //! Prologue for SizeTags for XML output archives /*! SizeTags do not start or finish nodes */ template inline void prologue( XMLOutputArchive & ar, SizeTag const & ) { ar.appendAttribute( "size", "dynamic" ); } template inline void prologue( XMLInputArchive &, SizeTag const & ) { } //! Epilogue for SizeTags for XML output archives /*! SizeTags do not start or finish nodes */ template inline void epilogue( XMLOutputArchive &, SizeTag const & ) { } template inline void epilogue( XMLInputArchive &, SizeTag const & ) { } // ###################################################################### //! Prologue for all other types for XML output archives (except minimal types) /*! Starts a new node, named either automatically or by some NVP, that may be given data by the type about to be archived Minimal types do not start or end nodes */ template ::value || traits::has_minimal_output_serialization::value> = traits::sfinae> inline void prologue( XMLOutputArchive & ar, T const & ) { ar.startNode(); ar.insertType(); } //! Prologue for all other types for XML input archives (except minimal types) template ::value || traits::has_minimal_input_serialization::value> = traits::sfinae> inline void prologue( XMLInputArchive & ar, T const & ) { ar.startNode(); } // ###################################################################### //! Epilogue for all other types other for XML output archives (except minimal types) /*! Finishes the node created in the prologue Minimal types do not start or end nodes */ template ::value || traits::has_minimal_output_serialization::value> = traits::sfinae> inline void epilogue( XMLOutputArchive & ar, T const & ) { ar.finishNode(); } //! Epilogue for all other types other for XML output archives (except minimal types) template ::value || traits::has_minimal_input_serialization::value> = traits::sfinae> inline void epilogue( XMLInputArchive & ar, T const & ) { ar.finishNode(); } // ###################################################################### // Common XMLArchive serialization functions // ###################################################################### //! Saving NVP types to XML template inline void CEREAL_SAVE_FUNCTION_NAME( XMLOutputArchive & ar, NameValuePair const & t ) { ar.setNextName( t.name ); ar( t.value ); } //! Loading NVP types from XML template inline void CEREAL_LOAD_FUNCTION_NAME( XMLInputArchive & ar, NameValuePair & t ) { ar.setNextName( t.name ); ar( t.value ); } // ###################################################################### //! Saving SizeTags to XML template inline void CEREAL_SAVE_FUNCTION_NAME( XMLOutputArchive &, SizeTag const & ) { } //! Loading SizeTags from XML template inline void CEREAL_LOAD_FUNCTION_NAME( XMLInputArchive & ar, SizeTag & st ) { ar.loadSize( st.size ); } // ###################################################################### //! Saving for POD types to xml template ::value> = traits::sfinae> inline void CEREAL_SAVE_FUNCTION_NAME(XMLOutputArchive & ar, T const & t) { ar.saveValue( t ); } //! Loading for POD types from xml template ::value> = traits::sfinae> inline void CEREAL_LOAD_FUNCTION_NAME(XMLInputArchive & ar, T & t) { ar.loadValue( t ); } // ###################################################################### //! saving string to xml template inline void CEREAL_SAVE_FUNCTION_NAME(XMLOutputArchive & ar, std::basic_string const & str) { ar.saveValue( str ); } //! loading string from xml template inline void CEREAL_LOAD_FUNCTION_NAME(XMLInputArchive & ar, std::basic_string & str) { ar.loadValue( str ); } } // namespace cereal // register archives for polymorphic support CEREAL_REGISTER_ARCHIVE(cereal::XMLOutputArchive) CEREAL_REGISTER_ARCHIVE(cereal::XMLInputArchive) // tie input and output archives together CEREAL_SETUP_ARCHIVE_TRAITS(cereal::XMLInputArchive, cereal::XMLOutputArchive) #endif // CEREAL_ARCHIVES_XML_HPP_