1 // Copyright (C) 2018 Intel Corporation
3 // SPDX-License-Identifier: Apache-2.0
6 #ifndef PUGIXML_VERSION
7 // Define version macro; evaluates to major * 100 + minor so that it's safe to use in less-than comparisons
8 # define PUGIXML_VERSION 170
11 // Include user configuration file (this can define various configuration macros)
12 #include "pugiconfig.hpp"
14 #ifndef HEADER_PUGIXML_HPP
15 #define HEADER_PUGIXML_HPP
17 // Include stddef.h for size_t and ptrdiff_t
20 // Include exception header for XPath
21 #if !defined(PUGIXML_NO_XPATH) && !defined(PUGIXML_NO_EXCEPTIONS)
25 // Include STL headers
26 #ifndef PUGIXML_NO_STL
32 // Macro for deprecated features
33 #ifndef PUGIXML_DEPRECATED
34 # if defined(__GNUC__)
35 # define PUGIXML_DEPRECATED __attribute__((deprecated))
36 # elif defined(_MSC_VER) && _MSC_VER >= 1300
37 # define PUGIXML_DEPRECATED __declspec(deprecated)
39 # define PUGIXML_DEPRECATED
43 // If no API is defined, assume default
48 // If no API for classes is defined, assume default
50 # define PUGIXML_CLASS PUGIXML_API
53 // If no API for functions is defined, assume default
54 #ifndef PUGIXML_FUNCTION
55 # define PUGIXML_FUNCTION PUGIXML_API
58 // If the platform is known to have long long support, enable long long functions
59 #ifndef PUGIXML_HAS_LONG_LONG
60 # if __cplusplus >= 201103
61 # define PUGIXML_HAS_LONG_LONG
62 # elif defined(_MSC_VER) && _MSC_VER >= 1400
63 # define PUGIXML_HAS_LONG_LONG
67 // Character interface macros
68 #ifdef PUGIXML_WCHAR_MODE
69 # define PUGIXML_TEXT(t) L ## t
70 # define PUGIXML_CHAR wchar_t
72 # define PUGIXML_TEXT(t) t
73 # define PUGIXML_CHAR char
78 // Character type used for all internal storage and operations; depends on PUGIXML_WCHAR_MODE
79 typedef PUGIXML_CHAR char_t;
81 #ifndef PUGIXML_NO_STL
82 // String type used for operations that work with STL string; depends on PUGIXML_WCHAR_MODE
83 typedef std::basic_string<PUGIXML_CHAR, std::char_traits<PUGIXML_CHAR>, std::allocator<PUGIXML_CHAR> > string_t;
87 // The PugiXML namespace
93 node_null, // Empty (null) node handle
94 node_document, // A document tree's absolute root
95 node_element, // Element tag, i.e. '<node/>'
96 node_pcdata, // Plain character data, i.e. 'text'
97 node_cdata, // Character data, i.e. '<![CDATA[text]]>'
98 node_comment, // Comment tag, i.e. '<!-- text -->'
99 node_pi, // Processing instruction, i.e. '<?name?>'
100 node_declaration, // Document declaration, i.e. '<?xml version="1.0"?>'
101 node_doctype // Document type declaration, i.e. '<!DOCTYPE doc>'
106 // Minimal parsing mode (equivalent to turning all other flags off).
107 // Only elements and PCDATA sections are added to the DOM tree, no text conversions are performed.
108 const unsigned int parse_minimal = 0x0000;
110 // This flag determines if processing instructions (node_pi) are added to the DOM tree. This flag is off by default.
111 const unsigned int parse_pi = 0x0001;
113 // This flag determines if comments (node_comment) are added to the DOM tree. This flag is off by default.
114 const unsigned int parse_comments = 0x0002;
116 // This flag determines if CDATA sections (node_cdata) are added to the DOM tree. This flag is on by default.
117 const unsigned int parse_cdata = 0x0004;
119 // This flag determines if plain character data (node_pcdata) that consist only of whitespace are added to the DOM tree.
120 // This flag is off by default; turning it on usually results in slower parsing and more memory consumption.
121 const unsigned int parse_ws_pcdata = 0x0008;
123 // This flag determines if character and entity references are expanded during parsing. This flag is on by default.
124 const unsigned int parse_escapes = 0x0010;
126 // This flag determines if EOL characters are normalized (converted to #xA) during parsing. This flag is on by default.
127 const unsigned int parse_eol = 0x0020;
129 // This flag determines if attribute values are normalized using CDATA normalization rules during parsing. This flag is on by default.
130 const unsigned int parse_wconv_attribute = 0x0040;
132 // This flag determines if attribute values are normalized using NMTOKENS normalization rules during parsing. This flag is off by default.
133 const unsigned int parse_wnorm_attribute = 0x0080;
135 // This flag determines if document declaration (node_declaration) is added to the DOM tree. This flag is off by default.
136 const unsigned int parse_declaration = 0x0100;
138 // This flag determines if document type declaration (node_doctype) is added to the DOM tree. This flag is off by default.
139 const unsigned int parse_doctype = 0x0200;
141 // This flag determines if plain character data (node_pcdata) that is the only child of the parent node and that consists only
142 // of whitespace is added to the DOM tree.
143 // This flag is off by default; turning it on may result in slower parsing and more memory consumption.
144 const unsigned int parse_ws_pcdata_single = 0x0400;
146 // This flag determines if leading and trailing whitespace is to be removed from plain character data. This flag is off by default.
147 const unsigned int parse_trim_pcdata = 0x0800;
149 // This flag determines if plain character data that does not have a parent node is added to the DOM tree, and if an empty document
150 // is a valid document. This flag is off by default.
151 const unsigned int parse_fragment = 0x1000;
153 // This flag determines if plain character data is be stored in the parent element's value. This significantly changes the structure of
154 // the document; this flag is only recommended for parsing documents with many PCDATA nodes in memory-constrained environments.
155 // This flag is off by default.
156 const unsigned int parse_embed_pcdata = 0x2000;
158 // The default parsing mode.
159 // Elements, PCDATA and CDATA sections are added to the DOM tree, character/reference entities are expanded,
160 // End-of-Line characters are normalized, attribute values are normalized using CDATA normalization rules.
161 const unsigned int parse_default = parse_cdata | parse_escapes | parse_wconv_attribute | parse_eol;
163 // The full parsing mode.
164 // Nodes of all types are added to the DOM tree, character/reference entities are expanded,
165 // End-of-Line characters are normalized, attribute values are normalized using CDATA normalization rules.
166 const unsigned int parse_full = parse_default | parse_pi | parse_comments | parse_declaration | parse_doctype;
168 // These flags determine the encoding of input data for XML document
171 encoding_auto, // Auto-detect input encoding using BOM or < / <? detection; use UTF8 if BOM is not found
172 encoding_utf8, // UTF8 encoding
173 encoding_utf16_le, // Little-endian UTF16
174 encoding_utf16_be, // Big-endian UTF16
175 encoding_utf16, // UTF16 with native endianness
176 encoding_utf32_le, // Little-endian UTF32
177 encoding_utf32_be, // Big-endian UTF32
178 encoding_utf32, // UTF32 with native endianness
179 encoding_wchar, // The same encoding wchar_t has (either UTF16 or UTF32)
185 // Indent the nodes that are written to output stream with as many indentation strings as deep the node is in DOM tree. This flag is on by default.
186 const unsigned int format_indent = 0x01;
188 // Write encoding-specific BOM to the output stream. This flag is off by default.
189 const unsigned int format_write_bom = 0x02;
191 // Use raw output mode (no indentation and no line breaks are written). This flag is off by default.
192 const unsigned int format_raw = 0x04;
194 // Omit default XML declaration even if there is no declaration in the document. This flag is off by default.
195 const unsigned int format_no_declaration = 0x08;
197 // Don't escape attribute values and PCDATA contents. This flag is off by default.
198 const unsigned int format_no_escapes = 0x10;
200 // Open file using text mode in xml_document::save_file. This enables special character (i.e. new-line) conversions on some systems. This flag is off by default.
201 const unsigned int format_save_file_text = 0x20;
203 // Write every attribute on a new line with appropriate indentation. This flag is off by default.
204 const unsigned int format_indent_attributes = 0x40;
206 // The default set of formatting flags.
207 // Nodes are indented depending on their depth in DOM tree, a default declaration is output if document has none.
208 const unsigned int format_default = format_indent;
210 // Forward declarations
211 struct xml_attribute_struct;
212 struct xml_node_struct;
214 class xml_node_iterator;
215 class xml_attribute_iterator;
216 class xml_named_node_iterator;
218 class xml_tree_walker;
220 struct xml_parse_result;
226 #ifndef PUGIXML_NO_XPATH
228 class xpath_node_set;
230 class xpath_variable_set;
233 // Range-based for loop support
234 template <typename It> class xml_object_range
237 typedef It const_iterator;
240 xml_object_range(It b, It e): _begin(b), _end(e)
244 It begin() const { return _begin; }
245 It end() const { return _end; }
251 // Writer interface for node printing (see xml_node::print)
252 class PUGIXML_CLASS xml_writer
255 virtual ~xml_writer() {}
257 // Write memory chunk into stream/file/whatever
258 virtual void write(const void* data, size_t size) = 0;
261 // xml_writer implementation for FILE*
262 class PUGIXML_CLASS xml_writer_file: public xml_writer
265 // Construct writer from a FILE* object; void* is used to avoid header dependencies on stdio
266 xml_writer_file(void* file);
268 virtual void write(const void* data, size_t size);
274 #ifndef PUGIXML_NO_STL
275 // xml_writer implementation for streams
276 class PUGIXML_CLASS xml_writer_stream: public xml_writer
279 // Construct writer from an output stream object
280 xml_writer_stream(std::basic_ostream<char, std::char_traits<char> >& stream);
281 xml_writer_stream(std::basic_ostream<wchar_t, std::char_traits<wchar_t> >& stream);
283 virtual void write(const void* data, size_t size);
286 std::basic_ostream<char, std::char_traits<char> >* narrow_stream;
287 std::basic_ostream<wchar_t, std::char_traits<wchar_t> >* wide_stream;
291 // A light-weight handle for manipulating attributes in DOM tree
292 class PUGIXML_CLASS xml_attribute
294 friend class xml_attribute_iterator;
295 friend class xml_node;
298 xml_attribute_struct* _attr;
300 typedef void (*unspecified_bool_type)(xml_attribute***);
303 // Default constructor. Constructs an empty attribute.
306 // Constructs attribute from internal pointer
307 explicit xml_attribute(xml_attribute_struct* attr);
309 // Safe bool conversion operator
310 operator unspecified_bool_type() const;
312 // Borland C++ workaround
313 bool operator!() const;
315 // Comparison operators (compares wrapped attribute pointers)
316 bool operator==(const xml_attribute& r) const;
317 bool operator!=(const xml_attribute& r) const;
318 bool operator<(const xml_attribute& r) const;
319 bool operator>(const xml_attribute& r) const;
320 bool operator<=(const xml_attribute& r) const;
321 bool operator>=(const xml_attribute& r) const;
323 // Check if attribute is empty
326 // Get attribute name/value, or "" if attribute is empty
327 const char_t* name() const;
328 const char_t* value() const;
330 // Get attribute value, or the default value if attribute is empty
331 const char_t* as_string(const char_t* def = PUGIXML_TEXT("")) const;
333 // Get attribute value as a number, or the default value if conversion did not succeed or attribute is empty
334 int as_int(int def = 0) const;
335 unsigned int as_uint(unsigned int def = 0) const;
336 double as_double(double def = 0) const;
337 float as_float(float def = 0) const;
339 #ifdef PUGIXML_HAS_LONG_LONG
340 long long as_llong(long long def = 0) const;
341 unsigned long long as_ullong(unsigned long long def = 0) const;
344 // Get attribute value as bool (returns true if first character is in '1tTyY' set), or the default value if attribute is empty
345 bool as_bool(bool def = false) const;
347 // Set attribute name/value (returns false if attribute is empty or there is not enough memory)
348 bool set_name(const char_t* rhs);
349 bool set_value(const char_t* rhs);
351 // Set attribute value with type conversion (numbers are converted to strings, boolean is converted to "true"/"false")
352 bool set_value(int rhs);
353 bool set_value(unsigned int rhs);
354 bool set_value(long rhs);
355 bool set_value(unsigned long rhs);
356 bool set_value(double rhs);
357 bool set_value(float rhs);
358 bool set_value(bool rhs);
360 #ifdef PUGIXML_HAS_LONG_LONG
361 bool set_value(long long rhs);
362 bool set_value(unsigned long long rhs);
365 // Set attribute value (equivalent to set_value without error checking)
366 xml_attribute& operator=(const char_t* rhs);
367 xml_attribute& operator=(int rhs);
368 xml_attribute& operator=(unsigned int rhs);
369 xml_attribute& operator=(long rhs);
370 xml_attribute& operator=(unsigned long rhs);
371 xml_attribute& operator=(double rhs);
372 xml_attribute& operator=(float rhs);
373 xml_attribute& operator=(bool rhs);
375 #ifdef PUGIXML_HAS_LONG_LONG
376 xml_attribute& operator=(long long rhs);
377 xml_attribute& operator=(unsigned long long rhs);
380 // Get next/previous attribute in the attribute list of the parent node
381 xml_attribute next_attribute() const;
382 xml_attribute previous_attribute() const;
384 // Get hash value (unique for handles to the same object)
385 size_t hash_value() const;
387 // Get internal pointer
388 xml_attribute_struct* internal_object() const;
392 // Borland C++ workaround
393 bool PUGIXML_FUNCTION operator&&(const xml_attribute& lhs, bool rhs);
394 bool PUGIXML_FUNCTION operator||(const xml_attribute& lhs, bool rhs);
397 // A light-weight handle for manipulating nodes in DOM tree
398 class PUGIXML_CLASS xml_node
400 friend class xml_attribute_iterator;
401 friend class xml_node_iterator;
402 friend class xml_named_node_iterator;
405 xml_node_struct* _root;
407 typedef void (*unspecified_bool_type)(xml_node***);
410 // Default constructor. Constructs an empty node.
413 // Constructs node from internal pointer
414 explicit xml_node(xml_node_struct* p);
416 // Safe bool conversion operator
417 operator unspecified_bool_type() const;
419 // Borland C++ workaround
420 bool operator!() const;
422 // Comparison operators (compares wrapped node pointers)
423 bool operator==(const xml_node& r) const;
424 bool operator!=(const xml_node& r) const;
425 bool operator<(const xml_node& r) const;
426 bool operator>(const xml_node& r) const;
427 bool operator<=(const xml_node& r) const;
428 bool operator>=(const xml_node& r) const;
430 // Check if node is empty.
434 xml_node_type type() const;
436 // Get node name, or "" if node is empty or it has no name
437 const char_t* name() const;
439 // Get node value, or "" if node is empty or it has no value
440 // Note: For <node>text</node> node.value() does not return "text"! Use child_value() or text() methods to access text inside nodes.
441 const char_t* value() const;
443 // Get attribute list
444 xml_attribute first_attribute() const;
445 xml_attribute last_attribute() const;
448 xml_node first_child() const;
449 xml_node last_child() const;
451 // Get next/previous sibling in the children list of the parent node
452 xml_node next_sibling() const;
453 xml_node previous_sibling() const;
456 xml_node parent() const;
458 // Get root of DOM tree this node belongs to
459 xml_node root() const;
461 // Get text object for the current node
462 xml_text text() const;
464 // Get child, attribute or next/previous sibling with the specified name
465 xml_node child(const char_t* name) const;
466 xml_attribute attribute(const char_t* name) const;
467 xml_node next_sibling(const char_t* name) const;
468 xml_node previous_sibling(const char_t* name) const;
470 // Get attribute, starting the search from a hint (and updating hint so that searching for a sequence of attributes is fast)
471 xml_attribute attribute(const char_t* name, xml_attribute& hint) const;
473 // Get child value of current node; that is, value of the first child node of type PCDATA/CDATA
474 const char_t* child_value() const;
476 // Get child value of child with specified name. Equivalent to child(name).child_value().
477 const char_t* child_value(const char_t* name) const;
479 // Set node name/value (returns false if node is empty, there is not enough memory, or node can not have name/value)
480 bool set_name(const char_t* rhs);
481 bool set_value(const char_t* rhs);
483 // Add attribute with specified name. Returns added attribute, or empty attribute on errors.
484 xml_attribute append_attribute(const char_t* name);
485 xml_attribute prepend_attribute(const char_t* name);
486 xml_attribute insert_attribute_after(const char_t* name, const xml_attribute& attr);
487 xml_attribute insert_attribute_before(const char_t* name, const xml_attribute& attr);
489 // Add a copy of the specified attribute. Returns added attribute, or empty attribute on errors.
490 xml_attribute append_copy(const xml_attribute& proto);
491 xml_attribute prepend_copy(const xml_attribute& proto);
492 xml_attribute insert_copy_after(const xml_attribute& proto, const xml_attribute& attr);
493 xml_attribute insert_copy_before(const xml_attribute& proto, const xml_attribute& attr);
495 // Add child node with specified type. Returns added node, or empty node on errors.
496 xml_node append_child(xml_node_type type = node_element);
497 xml_node prepend_child(xml_node_type type = node_element);
498 xml_node insert_child_after(xml_node_type type, const xml_node& node);
499 xml_node insert_child_before(xml_node_type type, const xml_node& node);
501 // Add child element with specified name. Returns added node, or empty node on errors.
502 xml_node append_child(const char_t* name);
503 xml_node prepend_child(const char_t* name);
504 xml_node insert_child_after(const char_t* name, const xml_node& node);
505 xml_node insert_child_before(const char_t* name, const xml_node& node);
507 // Add a copy of the specified node as a child. Returns added node, or empty node on errors.
508 xml_node append_copy(const xml_node& proto);
509 xml_node prepend_copy(const xml_node& proto);
510 xml_node insert_copy_after(const xml_node& proto, const xml_node& node);
511 xml_node insert_copy_before(const xml_node& proto, const xml_node& node);
513 // Move the specified node to become a child of this node. Returns moved node, or empty node on errors.
514 xml_node append_move(const xml_node& moved);
515 xml_node prepend_move(const xml_node& moved);
516 xml_node insert_move_after(const xml_node& moved, const xml_node& node);
517 xml_node insert_move_before(const xml_node& moved, const xml_node& node);
519 // Remove specified attribute
520 bool remove_attribute(const xml_attribute& a);
521 bool remove_attribute(const char_t* name);
523 // Remove specified child
524 bool remove_child(const xml_node& n);
525 bool remove_child(const char_t* name);
527 // Parses buffer as an XML document fragment and appends all nodes as children of the current node.
528 // Copies/converts the buffer, so it may be deleted or changed after the function returns.
529 // Note: append_buffer allocates memory that has the lifetime of the owning document; removing the appended nodes does not immediately reclaim that memory.
530 xml_parse_result append_buffer(const void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
532 // Find attribute using predicate. Returns first attribute for which predicate returned true.
533 template <typename Predicate> xml_attribute find_attribute(Predicate pred) const
535 if (!_root) return xml_attribute();
537 for (xml_attribute attrib = first_attribute(); attrib; attrib = attrib.next_attribute())
541 return xml_attribute();
544 // Find child node using predicate. Returns first child for which predicate returned true.
545 template <typename Predicate> xml_node find_child(Predicate pred) const
547 if (!_root) return xml_node();
549 for (xml_node node = first_child(); node; node = node.next_sibling())
556 // Find node from subtree using predicate. Returns first node from subtree (depth-first), for which predicate returned true.
557 template <typename Predicate> xml_node find_node(Predicate pred) const
559 if (!_root) return xml_node();
561 xml_node cur = first_child();
563 while (cur._root && cur._root != _root)
565 if (pred(cur)) return cur;
567 if (cur.first_child()) cur = cur.first_child();
568 else if (cur.next_sibling()) cur = cur.next_sibling();
571 while (!cur.next_sibling() && cur._root != _root) cur = cur.parent();
573 if (cur._root != _root) cur = cur.next_sibling();
580 // Find child node by attribute name/value
581 xml_node find_child_by_attribute(const char_t* name, const char_t* attr_name, const char_t* attr_value) const;
582 xml_node find_child_by_attribute(const char_t* attr_name, const char_t* attr_value) const;
584 #ifndef PUGIXML_NO_STL
585 // Get the absolute node path from root as a text string.
586 string_t path(char_t delimiter = '/') const;
589 // Search for a node by path consisting of node names and . or .. elements.
590 xml_node first_element_by_path(const char_t* path, char_t delimiter = '/') const;
592 // Recursively traverse subtree with xml_tree_walker
593 bool traverse(xml_tree_walker& walker);
595 #ifndef PUGIXML_NO_XPATH
596 // Select single node by evaluating XPath query. Returns first node from the resulting node set.
597 xpath_node select_node(const char_t* query, xpath_variable_set* variables = 0) const;
598 xpath_node select_node(const xpath_query& query) const;
600 // Select node set by evaluating XPath query
601 xpath_node_set select_nodes(const char_t* query, xpath_variable_set* variables = 0) const;
602 xpath_node_set select_nodes(const xpath_query& query) const;
604 // (deprecated: use select_node instead) Select single node by evaluating XPath query.
605 xpath_node select_single_node(const char_t* query, xpath_variable_set* variables = 0) const;
606 xpath_node select_single_node(const xpath_query& query) const;
610 // Print subtree using a writer object
611 void print(xml_writer& writer, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto, unsigned int depth = 0) const;
613 #ifndef PUGIXML_NO_STL
614 // Print subtree to stream
615 void print(std::basic_ostream<char, std::char_traits<char> >& os, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto, unsigned int depth = 0) const;
616 void print(std::basic_ostream<wchar_t, std::char_traits<wchar_t> >& os, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, unsigned int depth = 0) const;
619 // Child nodes iterators
620 typedef xml_node_iterator iterator;
622 iterator begin() const;
623 iterator end() const;
625 // Attribute iterators
626 typedef xml_attribute_iterator attribute_iterator;
628 attribute_iterator attributes_begin() const;
629 attribute_iterator attributes_end() const;
631 // Range-based for support
632 xml_object_range<xml_node_iterator> children() const;
633 xml_object_range<xml_named_node_iterator> children(const char_t* name) const;
634 xml_object_range<xml_attribute_iterator> attributes() const;
636 // Get node offset in parsed file/string (in char_t units) for debugging purposes
637 ptrdiff_t offset_debug() const;
639 // Get hash value (unique for handles to the same object)
640 size_t hash_value() const;
642 // Get internal pointer
643 xml_node_struct* internal_object() const;
647 // Borland C++ workaround
648 bool PUGIXML_FUNCTION operator&&(const xml_node& lhs, bool rhs);
649 bool PUGIXML_FUNCTION operator||(const xml_node& lhs, bool rhs);
652 // A helper for working with text inside PCDATA nodes
653 class PUGIXML_CLASS xml_text
655 friend class xml_node;
657 xml_node_struct* _root;
659 typedef void (*unspecified_bool_type)(xml_text***);
661 explicit xml_text(xml_node_struct* root);
663 xml_node_struct* _data_new();
664 xml_node_struct* _data() const;
667 // Default constructor. Constructs an empty object.
670 // Safe bool conversion operator
671 operator unspecified_bool_type() const;
673 // Borland C++ workaround
674 bool operator!() const;
676 // Check if text object is empty
679 // Get text, or "" if object is empty
680 const char_t* get() const;
682 // Get text, or the default value if object is empty
683 const char_t* as_string(const char_t* def = PUGIXML_TEXT("")) const;
685 // Get text as a number, or the default value if conversion did not succeed or object is empty
686 int as_int(int def = 0) const;
687 unsigned int as_uint(unsigned int def = 0) const;
688 double as_double(double def = 0) const;
689 float as_float(float def = 0) const;
691 #ifdef PUGIXML_HAS_LONG_LONG
692 long long as_llong(long long def = 0) const;
693 unsigned long long as_ullong(unsigned long long def = 0) const;
696 // Get text as bool (returns true if first character is in '1tTyY' set), or the default value if object is empty
697 bool as_bool(bool def = false) const;
699 // Set text (returns false if object is empty or there is not enough memory)
700 bool set(const char_t* rhs);
702 // Set text with type conversion (numbers are converted to strings, boolean is converted to "true"/"false")
704 bool set(unsigned int rhs);
706 bool set(unsigned long rhs);
707 bool set(double rhs);
711 #ifdef PUGIXML_HAS_LONG_LONG
712 bool set(long long rhs);
713 bool set(unsigned long long rhs);
716 // Set text (equivalent to set without error checking)
717 xml_text& operator=(const char_t* rhs);
718 xml_text& operator=(int rhs);
719 xml_text& operator=(unsigned int rhs);
720 xml_text& operator=(long rhs);
721 xml_text& operator=(unsigned long rhs);
722 xml_text& operator=(double rhs);
723 xml_text& operator=(float rhs);
724 xml_text& operator=(bool rhs);
726 #ifdef PUGIXML_HAS_LONG_LONG
727 xml_text& operator=(long long rhs);
728 xml_text& operator=(unsigned long long rhs);
731 // Get the data node (node_pcdata or node_cdata) for this object
732 xml_node data() const;
736 // Borland C++ workaround
737 bool PUGIXML_FUNCTION operator&&(const xml_text& lhs, bool rhs);
738 bool PUGIXML_FUNCTION operator||(const xml_text& lhs, bool rhs);
741 // Child node iterator (a bidirectional iterator over a collection of xml_node)
742 class PUGIXML_CLASS xml_node_iterator
744 friend class xml_node;
747 mutable xml_node _wrap;
750 xml_node_iterator(xml_node_struct* ref, xml_node_struct* parent);
754 typedef ptrdiff_t difference_type;
755 typedef xml_node value_type;
756 typedef xml_node* pointer;
757 typedef xml_node& reference;
759 #ifndef PUGIXML_NO_STL
760 typedef std::bidirectional_iterator_tag iterator_category;
763 // Default constructor
766 // Construct an iterator which points to the specified node
767 xml_node_iterator(const xml_node& node);
769 // Iterator operators
770 bool operator==(const xml_node_iterator& rhs) const;
771 bool operator!=(const xml_node_iterator& rhs) const;
773 xml_node& operator*() const;
774 xml_node* operator->() const;
776 const xml_node_iterator& operator++();
777 xml_node_iterator operator++(int);
779 const xml_node_iterator& operator--();
780 xml_node_iterator operator--(int);
783 // Attribute iterator (a bidirectional iterator over a collection of xml_attribute)
784 class PUGIXML_CLASS xml_attribute_iterator
786 friend class xml_node;
789 mutable xml_attribute _wrap;
792 xml_attribute_iterator(xml_attribute_struct* ref, xml_node_struct* parent);
796 typedef ptrdiff_t difference_type;
797 typedef xml_attribute value_type;
798 typedef xml_attribute* pointer;
799 typedef xml_attribute& reference;
801 #ifndef PUGIXML_NO_STL
802 typedef std::bidirectional_iterator_tag iterator_category;
805 // Default constructor
806 xml_attribute_iterator();
808 // Construct an iterator which points to the specified attribute
809 xml_attribute_iterator(const xml_attribute& attr, const xml_node& parent);
811 // Iterator operators
812 bool operator==(const xml_attribute_iterator& rhs) const;
813 bool operator!=(const xml_attribute_iterator& rhs) const;
815 xml_attribute& operator*() const;
816 xml_attribute* operator->() const;
818 const xml_attribute_iterator& operator++();
819 xml_attribute_iterator operator++(int);
821 const xml_attribute_iterator& operator--();
822 xml_attribute_iterator operator--(int);
825 // Named node range helper
826 class PUGIXML_CLASS xml_named_node_iterator
828 friend class xml_node;
832 typedef ptrdiff_t difference_type;
833 typedef xml_node value_type;
834 typedef xml_node* pointer;
835 typedef xml_node& reference;
837 #ifndef PUGIXML_NO_STL
838 typedef std::bidirectional_iterator_tag iterator_category;
841 // Default constructor
842 xml_named_node_iterator();
844 // Construct an iterator which points to the specified node
845 xml_named_node_iterator(const xml_node& node, const char_t* name);
847 // Iterator operators
848 bool operator==(const xml_named_node_iterator& rhs) const;
849 bool operator!=(const xml_named_node_iterator& rhs) const;
851 xml_node& operator*() const;
852 xml_node* operator->() const;
854 const xml_named_node_iterator& operator++();
855 xml_named_node_iterator operator++(int);
857 const xml_named_node_iterator& operator--();
858 xml_named_node_iterator operator--(int);
861 mutable xml_node _wrap;
865 xml_named_node_iterator(xml_node_struct* ref, xml_node_struct* parent, const char_t* name);
868 // Abstract tree walker class (see xml_node::traverse)
869 class PUGIXML_CLASS xml_tree_walker
871 friend class xml_node;
877 // Get current traversal depth
882 virtual ~xml_tree_walker();
884 // Callback that is called when traversal begins
885 virtual bool begin(xml_node& node);
887 // Callback that is called for each node traversed
888 virtual bool for_each(xml_node& node) = 0;
890 // Callback that is called when traversal ends
891 virtual bool end(xml_node& node);
894 // Parsing status, returned as part of xml_parse_result object
895 enum xml_parse_status
897 status_ok = 0, // No error
899 status_file_not_found, // File was not found during load_file()
900 status_io_error, // Error reading from file/stream
901 status_out_of_memory, // Could not allocate memory
902 status_internal_error, // Internal error occurred
904 status_unrecognized_tag, // Parser could not determine tag type
906 status_bad_pi, // Parsing error occurred while parsing document declaration/processing instruction
907 status_bad_comment, // Parsing error occurred while parsing comment
908 status_bad_cdata, // Parsing error occurred while parsing CDATA section
909 status_bad_doctype, // Parsing error occurred while parsing document type declaration
910 status_bad_pcdata, // Parsing error occurred while parsing PCDATA section
911 status_bad_start_element, // Parsing error occurred while parsing start element tag
912 status_bad_attribute, // Parsing error occurred while parsing element attribute
913 status_bad_end_element, // Parsing error occurred while parsing end element tag
914 status_end_element_mismatch,// There was a mismatch of start-end tags (closing tag had incorrect name, some tag was not closed or there was an excessive closing tag)
916 status_append_invalid_root, // Unable to append nodes since root type is not node_element or node_document (exclusive to xml_node::append_buffer)
918 status_no_document_element // Parsing resulted in a document without element nodes
922 struct PUGIXML_CLASS xml_parse_result
924 // Parsing status (see xml_parse_status)
925 xml_parse_status status;
927 // Last parsed offset (in char_t units from start of input data)
930 // Source document encoding
931 xml_encoding encoding;
933 // Default constructor, initializes object to failed state
936 // Cast to bool operator
937 operator bool() const;
939 // Get error description
940 const char* description() const;
943 // Document class (DOM tree root)
944 class PUGIXML_CLASS xml_document: public xml_node
951 // Non-copyable semantics
952 xml_document(const xml_document&);
953 xml_document& operator=(const xml_document&);
959 // Default constructor, makes empty document
962 // Destructor, invalidates all node/attribute handles to this document
965 // Removes all nodes, leaving the empty document
968 // Removes all nodes, then copies the entire contents of the specified document
969 void reset(const xml_document& proto);
971 #ifndef PUGIXML_NO_STL
972 // Load document from stream.
973 xml_parse_result load(std::basic_istream<char, std::char_traits<char> >& stream, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
974 xml_parse_result load(std::basic_istream<wchar_t, std::char_traits<wchar_t> >& stream, unsigned int options = parse_default);
977 // (deprecated: use load_string instead) Load document from zero-terminated string. No encoding conversions are applied.
978 xml_parse_result load(const char_t* contents, unsigned int options = parse_default);
980 // Load document from zero-terminated string. No encoding conversions are applied.
981 xml_parse_result load_string(const char_t* contents, unsigned int options = parse_default);
983 // Load document from file
984 xml_parse_result load_file(const char* path, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
985 xml_parse_result load_file(const wchar_t* path, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
987 // Load document from buffer. Copies/converts the buffer, so it may be deleted or changed after the function returns.
988 xml_parse_result load_buffer(const void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
990 // Load document from buffer, using the buffer for in-place parsing (the buffer is modified and used for storage of document data).
991 // You should ensure that buffer data will persist throughout the document's lifetime, and free the buffer memory manually once document is destroyed.
992 xml_parse_result load_buffer_inplace(void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
994 // Load document from buffer, using the buffer for in-place parsing (the buffer is modified and used for storage of document data).
995 // You should allocate the buffer with pugixml allocation function; document will free the buffer when it is no longer needed (you can't use it anymore).
996 xml_parse_result load_buffer_inplace_own(void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
998 // Save XML document to writer (semantics is slightly different from xml_node::print, see documentation for details).
999 void save(xml_writer& writer, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const;
1001 #ifndef PUGIXML_NO_STL
1002 // Save XML document to stream (semantics is slightly different from xml_node::print, see documentation for details).
1003 void save(std::basic_ostream<char, std::char_traits<char> >& stream, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const;
1004 void save(std::basic_ostream<wchar_t, std::char_traits<wchar_t> >& stream, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default) const;
1008 bool save_file(const char* path, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const;
1009 bool save_file(const wchar_t* path, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const;
1011 // Get document element
1012 xml_node document_element() const;
1015 #ifndef PUGIXML_NO_XPATH
1016 // XPath query return type
1017 enum xpath_value_type
1019 xpath_type_none, // Unknown type (query failed to compile)
1020 xpath_type_node_set, // Node set (xpath_node_set)
1021 xpath_type_number, // Number
1022 xpath_type_string, // String
1023 xpath_type_boolean // Boolean
1026 // XPath parsing result
1027 struct PUGIXML_CLASS xpath_parse_result
1029 // Error message (0 if no error)
1032 // Last parsed offset (in char_t units from string start)
1035 // Default constructor, initializes object to failed state
1036 xpath_parse_result();
1038 // Cast to bool operator
1039 operator bool() const;
1041 // Get error description
1042 const char* description() const;
1045 // A single XPath variable
1046 class PUGIXML_CLASS xpath_variable
1048 friend class xpath_variable_set;
1051 xpath_value_type _type;
1052 xpath_variable* _next;
1054 xpath_variable(xpath_value_type type);
1056 // Non-copyable semantics
1057 xpath_variable(const xpath_variable&);
1058 xpath_variable& operator=(const xpath_variable&);
1061 // Get variable name
1062 const char_t* name() const;
1064 // Get variable type
1065 xpath_value_type type() const;
1067 // Get variable value; no type conversion is performed, default value (false, NaN, empty string, empty node set) is returned on type mismatch error
1068 bool get_boolean() const;
1069 double get_number() const;
1070 const char_t* get_string() const;
1071 const xpath_node_set& get_node_set() const;
1073 // Set variable value; no type conversion is performed, false is returned on type mismatch error
1074 bool set(bool value);
1075 bool set(double value);
1076 bool set(const char_t* value);
1077 bool set(const xpath_node_set& value);
1080 // A set of XPath variables
1081 class PUGIXML_CLASS xpath_variable_set
1084 xpath_variable* _data[64];
1086 void _assign(const xpath_variable_set& rhs);
1087 void _swap(xpath_variable_set& rhs);
1089 xpath_variable* _find(const char_t* name) const;
1091 static bool _clone(xpath_variable* var, xpath_variable** out_result);
1092 static void _destroy(xpath_variable* var);
1095 // Default constructor/destructor
1096 xpath_variable_set();
1097 ~xpath_variable_set();
1099 // Copy constructor/assignment operator
1100 xpath_variable_set(const xpath_variable_set& rhs);
1101 xpath_variable_set& operator=(const xpath_variable_set& rhs);
1103 #if __cplusplus >= 201103
1104 // Move semantics support
1105 xpath_variable_set(xpath_variable_set&& rhs);
1106 xpath_variable_set& operator=(xpath_variable_set&& rhs);
1109 // Add a new variable or get the existing one, if the types match
1110 xpath_variable* add(const char_t* name, xpath_value_type type);
1112 // Set value of an existing variable; no type conversion is performed, false is returned if there is no such variable or if types mismatch
1113 bool set(const char_t* name, bool value);
1114 bool set(const char_t* name, double value);
1115 bool set(const char_t* name, const char_t* value);
1116 bool set(const char_t* name, const xpath_node_set& value);
1118 // Get existing variable by name
1119 xpath_variable* get(const char_t* name);
1120 const xpath_variable* get(const char_t* name) const;
1123 // A compiled XPath query object
1124 class PUGIXML_CLASS xpath_query
1128 xpath_parse_result _result;
1130 typedef void (*unspecified_bool_type)(xpath_query***);
1132 // Non-copyable semantics
1133 xpath_query(const xpath_query&);
1134 xpath_query& operator=(const xpath_query&);
1137 // Construct a compiled object from XPath expression.
1138 // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on compilation errors.
1139 explicit xpath_query(const char_t* query, xpath_variable_set* variables = 0);
1147 #if __cplusplus >= 201103
1148 // Move semantics support
1149 xpath_query(xpath_query&& rhs);
1150 xpath_query& operator=(xpath_query&& rhs);
1153 // Get query expression return type
1154 xpath_value_type return_type() const;
1156 // Evaluate expression as boolean value in the specified context; performs type conversion if necessary.
1157 // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors.
1158 bool evaluate_boolean(const xpath_node& n) const;
1160 // Evaluate expression as double value in the specified context; performs type conversion if necessary.
1161 // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors.
1162 double evaluate_number(const xpath_node& n) const;
1164 #ifndef PUGIXML_NO_STL
1165 // Evaluate expression as string value in the specified context; performs type conversion if necessary.
1166 // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors.
1167 string_t evaluate_string(const xpath_node& n) const;
1170 // Evaluate expression as string value in the specified context; performs type conversion if necessary.
1171 // At most capacity characters are written to the destination buffer, full result size is returned (includes terminating zero).
1172 // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors.
1173 // If PUGIXML_NO_EXCEPTIONS is defined, returns empty set instead.
1174 size_t evaluate_string(char_t* buffer, size_t capacity, const xpath_node& n) const;
1176 // Evaluate expression as node set in the specified context.
1177 // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on type mismatch and std::bad_alloc on out of memory errors.
1178 // If PUGIXML_NO_EXCEPTIONS is defined, returns empty node set instead.
1179 xpath_node_set evaluate_node_set(const xpath_node& n) const;
1181 // Evaluate expression as node set in the specified context.
1182 // Return first node in document order, or empty node if node set is empty.
1183 // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on type mismatch and std::bad_alloc on out of memory errors.
1184 // If PUGIXML_NO_EXCEPTIONS is defined, returns empty node instead.
1185 xpath_node evaluate_node(const xpath_node& n) const;
1187 // Get parsing result (used to get compilation errors in PUGIXML_NO_EXCEPTIONS mode)
1188 const xpath_parse_result& result() const;
1190 // Safe bool conversion operator
1191 operator unspecified_bool_type() const;
1193 // Borland C++ workaround
1194 bool operator!() const;
1197 #ifndef PUGIXML_NO_EXCEPTIONS
1198 // XPath exception class
1199 class PUGIXML_CLASS xpath_exception: public std::exception
1202 xpath_parse_result _result;
1205 // Construct exception from parse result
1206 explicit xpath_exception(const xpath_parse_result& result);
1208 // Get error message
1209 virtual const char* what() const throw();
1212 const xpath_parse_result& result() const;
1216 // XPath node class (either xml_node or xml_attribute)
1217 class PUGIXML_CLASS xpath_node
1221 xml_attribute _attribute;
1223 typedef void (*unspecified_bool_type)(xpath_node***);
1226 // Default constructor; constructs empty XPath node
1229 // Construct XPath node from XML node/attribute
1230 xpath_node(const xml_node& node);
1231 xpath_node(const xml_attribute& attribute, const xml_node& parent);
1233 // Get node/attribute, if any
1234 xml_node node() const;
1235 xml_attribute attribute() const;
1237 // Get parent of contained node/attribute
1238 xml_node parent() const;
1240 // Safe bool conversion operator
1241 operator unspecified_bool_type() const;
1243 // Borland C++ workaround
1244 bool operator!() const;
1246 // Comparison operators
1247 bool operator==(const xpath_node& n) const;
1248 bool operator!=(const xpath_node& n) const;
1252 // Borland C++ workaround
1253 bool PUGIXML_FUNCTION operator&&(const xpath_node& lhs, bool rhs);
1254 bool PUGIXML_FUNCTION operator||(const xpath_node& lhs, bool rhs);
1257 // A fixed-size collection of XPath nodes
1258 class PUGIXML_CLASS xpath_node_set
1264 type_unsorted, // Not ordered
1265 type_sorted, // Sorted by document order (ascending)
1266 type_sorted_reverse // Sorted by document order (descending)
1269 // Constant iterator type
1270 typedef const xpath_node* const_iterator;
1272 // We define non-constant iterator to be the same as constant iterator so that various generic algorithms (i.e. boost foreach) work
1273 typedef const xpath_node* iterator;
1275 // Default constructor. Constructs empty set.
1278 // Constructs a set from iterator range; data is not checked for duplicates and is not sorted according to provided type, so be careful
1279 xpath_node_set(const_iterator begin, const_iterator end, type_t type = type_unsorted);
1284 // Copy constructor/assignment operator
1285 xpath_node_set(const xpath_node_set& ns);
1286 xpath_node_set& operator=(const xpath_node_set& ns);
1288 #if __cplusplus >= 201103
1289 // Move semantics support
1290 xpath_node_set(xpath_node_set&& rhs);
1291 xpath_node_set& operator=(xpath_node_set&& rhs);
1294 // Get collection type
1295 type_t type() const;
1297 // Get collection size
1298 size_t size() const;
1300 // Indexing operator
1301 const xpath_node& operator[](size_t index) const;
1303 // Collection iterators
1304 const_iterator begin() const;
1305 const_iterator end() const;
1307 // Sort the collection in ascending/descending order by document order
1308 void sort(bool reverse = false);
1310 // Get first node in the collection by document order
1311 xpath_node first() const;
1313 // Check if collection is empty
1319 xpath_node _storage;
1324 void _assign(const_iterator begin, const_iterator end, type_t type);
1325 void _move(xpath_node_set& rhs);
1329 #ifndef PUGIXML_NO_STL
1330 // Convert wide string to UTF8
1331 std::basic_string<char, std::char_traits<char>, std::allocator<char> > PUGIXML_FUNCTION as_utf8(const wchar_t* str);
1332 std::basic_string<char, std::char_traits<char>, std::allocator<char> > PUGIXML_FUNCTION as_utf8(const std::basic_string<wchar_t, std::char_traits<wchar_t>, std::allocator<wchar_t> >& str);
1334 // Convert UTF8 to wide string
1335 std::basic_string<wchar_t, std::char_traits<wchar_t>, std::allocator<wchar_t> > PUGIXML_FUNCTION as_wide(const char* str);
1336 std::basic_string<wchar_t, std::char_traits<wchar_t>, std::allocator<wchar_t> > PUGIXML_FUNCTION as_wide(const std::basic_string<char, std::char_traits<char>, std::allocator<char> >& str);
1339 // Memory allocation function interface; returns pointer to allocated memory or NULL on failure
1340 typedef void* (*allocation_function)(size_t size);
1342 // Memory deallocation function interface
1343 typedef void (*deallocation_function)(void* ptr);
1345 // Override default memory management functions. All subsequent allocations/deallocations will be performed via supplied functions.
1346 void PUGIXML_FUNCTION set_memory_management_functions(allocation_function allocate, deallocation_function deallocate);
1348 // Get current memory management functions
1349 allocation_function PUGIXML_FUNCTION get_memory_allocation_function();
1350 deallocation_function PUGIXML_FUNCTION get_memory_deallocation_function();
1353 #if !defined(PUGIXML_NO_STL) && (defined(_MSC_VER) || defined(__ICC))
1356 // Workarounds for (non-standard) iterator category detection for older versions (MSVC7/IC8 and earlier)
1357 std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_node_iterator&);
1358 std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_attribute_iterator&);
1359 std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_named_node_iterator&);
1363 #if !defined(PUGIXML_NO_STL) && defined(__SUNPRO_CC)
1366 // Workarounds for (non-standard) iterator category detection
1367 std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_node_iterator&);
1368 std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_attribute_iterator&);
1369 std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_named_node_iterator&);
1375 // Make sure implementation is included in header-only mode
1376 // Use macro expansion in #include to work around QMake (QTBUG-11923)
1377 #if defined(PUGIXML_HEADER_ONLY) && !defined(PUGIXML_SOURCE)
1378 # define PUGIXML_SOURCE "pugixml.cpp"
1379 # include PUGIXML_SOURCE
1383 * Copyright (c) 2006-2016 Arseny Kapoulkine
1385 * Permission is hereby granted, free of charge, to any person
1386 * obtaining a copy of this software and associated documentation
1387 * files (the "Software"), to deal in the Software without
1388 * restriction, including without limitation the rights to use,
1389 * copy, modify, merge, publish, distribute, sublicense, and/or sell
1390 * copies of the Software, and to permit persons to whom the
1391 * Software is furnished to do so, subject to the following
1394 * The above copyright notice and this permission notice shall be
1395 * included in all copies or substantial portions of the Software.
1397 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
1398 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
1399 * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
1400 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
1401 * HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
1402 * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
1403 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
1404 * OTHER DEALINGS IN THE SOFTWARE.