1 // Copyright (C) 2018-2019 Intel Corporation
2 // SPDX-License-Identifier: Apache-2.0
5 #ifndef PUGIXML_VERSION
6 // Define version macro; evaluates to major * 100 + minor so that it's safe to use in less-than comparisons
7 # define PUGIXML_VERSION 170
10 // Include user configuration file (this can define various configuration macros)
11 #include "pugiconfig.hpp"
13 #ifndef HEADER_PUGIXML_HPP
14 #define HEADER_PUGIXML_HPP
16 // Include stddef.h for size_t and ptrdiff_t
19 // Include exception header for XPath
20 #if !defined(PUGIXML_NO_XPATH) && !defined(PUGIXML_NO_EXCEPTIONS)
24 // Include STL headers
25 #ifndef PUGIXML_NO_STL
31 // Macro for deprecated features
32 #ifndef PUGIXML_DEPRECATED
33 # if defined(__GNUC__)
34 # define PUGIXML_DEPRECATED __attribute__((deprecated))
35 # elif defined(_MSC_VER) && _MSC_VER >= 1300
36 # define PUGIXML_DEPRECATED __declspec(deprecated)
38 # define PUGIXML_DEPRECATED
42 // If no API is defined, assume default
47 // If no API for classes is defined, assume default
49 # define PUGIXML_CLASS PUGIXML_API
52 // If no API for functions is defined, assume default
53 #ifndef PUGIXML_FUNCTION
54 # define PUGIXML_FUNCTION PUGIXML_API
57 // If the platform is known to have long long support, enable long long functions
58 #ifndef PUGIXML_HAS_LONG_LONG
59 # if __cplusplus >= 201103
60 # define PUGIXML_HAS_LONG_LONG
61 # elif defined(_MSC_VER) && _MSC_VER >= 1400
62 # define PUGIXML_HAS_LONG_LONG
66 // Character interface macros
67 #ifdef PUGIXML_WCHAR_MODE
68 # define PUGIXML_TEXT(t) L ## t
69 # define PUGIXML_CHAR wchar_t
71 # define PUGIXML_TEXT(t) t
72 # define PUGIXML_CHAR char
77 // Character type used for all internal storage and operations; depends on PUGIXML_WCHAR_MODE
78 typedef PUGIXML_CHAR char_t;
80 #ifndef PUGIXML_NO_STL
81 // String type used for operations that work with STL string; depends on PUGIXML_WCHAR_MODE
82 typedef std::basic_string<PUGIXML_CHAR, std::char_traits<PUGIXML_CHAR>, std::allocator<PUGIXML_CHAR> > string_t;
86 // The PugiXML namespace
92 node_null, // Empty (null) node handle
93 node_document, // A document tree's absolute root
94 node_element, // Element tag, i.e. '<node/>'
95 node_pcdata, // Plain character data, i.e. 'text'
96 node_cdata, // Character data, i.e. '<![CDATA[text]]>'
97 node_comment, // Comment tag, i.e. '<!-- text -->'
98 node_pi, // Processing instruction, i.e. '<?name?>'
99 node_declaration, // Document declaration, i.e. '<?xml version="1.0"?>'
100 node_doctype // Document type declaration, i.e. '<!DOCTYPE doc>'
105 // Minimal parsing mode (equivalent to turning all other flags off).
106 // Only elements and PCDATA sections are added to the DOM tree, no text conversions are performed.
107 const unsigned int parse_minimal = 0x0000;
109 // This flag determines if processing instructions (node_pi) are added to the DOM tree. This flag is off by default.
110 const unsigned int parse_pi = 0x0001;
112 // This flag determines if comments (node_comment) are added to the DOM tree. This flag is off by default.
113 const unsigned int parse_comments = 0x0002;
115 // This flag determines if CDATA sections (node_cdata) are added to the DOM tree. This flag is on by default.
116 const unsigned int parse_cdata = 0x0004;
118 // This flag determines if plain character data (node_pcdata) that consist only of whitespace are added to the DOM tree.
119 // This flag is off by default; turning it on usually results in slower parsing and more memory consumption.
120 const unsigned int parse_ws_pcdata = 0x0008;
122 // This flag determines if character and entity references are expanded during parsing. This flag is on by default.
123 const unsigned int parse_escapes = 0x0010;
125 // This flag determines if EOL characters are normalized (converted to #xA) during parsing. This flag is on by default.
126 const unsigned int parse_eol = 0x0020;
128 // This flag determines if attribute values are normalized using CDATA normalization rules during parsing. This flag is on by default.
129 const unsigned int parse_wconv_attribute = 0x0040;
131 // This flag determines if attribute values are normalized using NMTOKENS normalization rules during parsing. This flag is off by default.
132 const unsigned int parse_wnorm_attribute = 0x0080;
134 // This flag determines if document declaration (node_declaration) is added to the DOM tree. This flag is off by default.
135 const unsigned int parse_declaration = 0x0100;
137 // This flag determines if document type declaration (node_doctype) is added to the DOM tree. This flag is off by default.
138 const unsigned int parse_doctype = 0x0200;
140 // This flag determines if plain character data (node_pcdata) that is the only child of the parent node and that consists only
141 // of whitespace is added to the DOM tree.
142 // This flag is off by default; turning it on may result in slower parsing and more memory consumption.
143 const unsigned int parse_ws_pcdata_single = 0x0400;
145 // This flag determines if leading and trailing whitespace is to be removed from plain character data. This flag is off by default.
146 const unsigned int parse_trim_pcdata = 0x0800;
148 // 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
149 // is a valid document. This flag is off by default.
150 const unsigned int parse_fragment = 0x1000;
152 // This flag determines if plain character data is be stored in the parent element's value. This significantly changes the structure of
153 // the document; this flag is only recommended for parsing documents with many PCDATA nodes in memory-constrained environments.
154 // This flag is off by default.
155 const unsigned int parse_embed_pcdata = 0x2000;
157 // The default parsing mode.
158 // Elements, PCDATA and CDATA sections are added to the DOM tree, character/reference entities are expanded,
159 // End-of-Line characters are normalized, attribute values are normalized using CDATA normalization rules.
160 const unsigned int parse_default = parse_cdata | parse_escapes | parse_wconv_attribute | parse_eol;
162 // The full parsing mode.
163 // Nodes of all types are added to the DOM tree, character/reference entities are expanded,
164 // End-of-Line characters are normalized, attribute values are normalized using CDATA normalization rules.
165 const unsigned int parse_full = parse_default | parse_pi | parse_comments | parse_declaration | parse_doctype;
167 // These flags determine the encoding of input data for XML document
170 encoding_auto, // Auto-detect input encoding using BOM or < / <? detection; use UTF8 if BOM is not found
171 encoding_utf8, // UTF8 encoding
172 encoding_utf16_le, // Little-endian UTF16
173 encoding_utf16_be, // Big-endian UTF16
174 encoding_utf16, // UTF16 with native endianness
175 encoding_utf32_le, // Little-endian UTF32
176 encoding_utf32_be, // Big-endian UTF32
177 encoding_utf32, // UTF32 with native endianness
178 encoding_wchar, // The same encoding wchar_t has (either UTF16 or UTF32)
184 // 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.
185 const unsigned int format_indent = 0x01;
187 // Write encoding-specific BOM to the output stream. This flag is off by default.
188 const unsigned int format_write_bom = 0x02;
190 // Use raw output mode (no indentation and no line breaks are written). This flag is off by default.
191 const unsigned int format_raw = 0x04;
193 // Omit default XML declaration even if there is no declaration in the document. This flag is off by default.
194 const unsigned int format_no_declaration = 0x08;
196 // Don't escape attribute values and PCDATA contents. This flag is off by default.
197 const unsigned int format_no_escapes = 0x10;
199 // 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.
200 const unsigned int format_save_file_text = 0x20;
202 // Write every attribute on a new line with appropriate indentation. This flag is off by default.
203 const unsigned int format_indent_attributes = 0x40;
205 // The default set of formatting flags.
206 // Nodes are indented depending on their depth in DOM tree, a default declaration is output if document has none.
207 const unsigned int format_default = format_indent;
209 // Forward declarations
210 struct xml_attribute_struct;
211 struct xml_node_struct;
213 class xml_node_iterator;
214 class xml_attribute_iterator;
215 class xml_named_node_iterator;
217 class xml_tree_walker;
219 struct xml_parse_result;
225 #ifndef PUGIXML_NO_XPATH
227 class xpath_node_set;
229 class xpath_variable_set;
232 // Range-based for loop support
233 template <typename It> class xml_object_range
236 typedef It const_iterator;
239 xml_object_range(It b, It e): _begin(b), _end(e)
243 It begin() const { return _begin; }
244 It end() const { return _end; }
250 // Writer interface for node printing (see xml_node::print)
251 class PUGIXML_CLASS xml_writer
254 virtual ~xml_writer() {}
256 // Write memory chunk into stream/file/whatever
257 virtual void write(const void* data, size_t size) = 0;
260 // xml_writer implementation for FILE*
261 class PUGIXML_CLASS xml_writer_file: public xml_writer
264 // Construct writer from a FILE* object; void* is used to avoid header dependencies on stdio
265 xml_writer_file(void* file);
267 virtual void write(const void* data, size_t size);
273 #ifndef PUGIXML_NO_STL
274 // xml_writer implementation for streams
275 class PUGIXML_CLASS xml_writer_stream: public xml_writer
278 // Construct writer from an output stream object
279 xml_writer_stream(std::basic_ostream<char, std::char_traits<char> >& stream);
280 xml_writer_stream(std::basic_ostream<wchar_t, std::char_traits<wchar_t> >& stream);
282 virtual void write(const void* data, size_t size);
285 std::basic_ostream<char, std::char_traits<char> >* narrow_stream;
286 std::basic_ostream<wchar_t, std::char_traits<wchar_t> >* wide_stream;
290 // A light-weight handle for manipulating attributes in DOM tree
291 class PUGIXML_CLASS xml_attribute
293 friend class xml_attribute_iterator;
294 friend class xml_node;
297 xml_attribute_struct* _attr;
299 typedef void (*unspecified_bool_type)(xml_attribute***);
302 // Default constructor. Constructs an empty attribute.
305 // Constructs attribute from internal pointer
306 explicit xml_attribute(xml_attribute_struct* attr);
308 // Safe bool conversion operator
309 operator unspecified_bool_type() const;
311 // Borland C++ workaround
312 bool operator!() const;
314 // Comparison operators (compares wrapped attribute pointers)
315 bool operator==(const xml_attribute& r) const;
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;
322 // Check if attribute is empty
325 // Get attribute name/value, or "" if attribute is empty
326 const char_t* name() const;
327 const char_t* value() const;
329 // Get attribute value, or the default value if attribute is empty
330 const char_t* as_string(const char_t* def = PUGIXML_TEXT("")) const;
332 // Get attribute value as a number, or the default value if conversion did not succeed or attribute is empty
333 int as_int(int def = 0) const;
334 unsigned int as_uint(unsigned int def = 0) const;
335 double as_double(double def = 0) const;
336 float as_float(float def = 0) const;
338 #ifdef PUGIXML_HAS_LONG_LONG
339 long long as_llong(long long def = 0) const;
340 unsigned long long as_ullong(unsigned long long def = 0) const;
343 // Get attribute value as bool (returns true if first character is in '1tTyY' set), or the default value if attribute is empty
344 bool as_bool(bool def = false) const;
346 // Set attribute name/value (returns false if attribute is empty or there is not enough memory)
347 bool set_name(const char_t* rhs);
348 bool set_value(const char_t* rhs);
350 // Set attribute value with type conversion (numbers are converted to strings, boolean is converted to "true"/"false")
351 bool set_value(int rhs);
352 bool set_value(unsigned int rhs);
353 bool set_value(long rhs);
354 bool set_value(unsigned long rhs);
355 bool set_value(double rhs);
356 bool set_value(float rhs);
357 bool set_value(bool rhs);
359 #ifdef PUGIXML_HAS_LONG_LONG
360 bool set_value(long long rhs);
361 bool set_value(unsigned long long rhs);
364 // Set attribute value (equivalent to set_value without error checking)
365 xml_attribute& operator=(const char_t* rhs);
366 xml_attribute& operator=(int rhs);
367 xml_attribute& operator=(unsigned int rhs);
368 xml_attribute& operator=(long rhs);
369 xml_attribute& operator=(unsigned long rhs);
370 xml_attribute& operator=(double rhs);
371 xml_attribute& operator=(float rhs);
372 xml_attribute& operator=(bool rhs);
374 #ifdef PUGIXML_HAS_LONG_LONG
375 xml_attribute& operator=(long long rhs);
376 xml_attribute& operator=(unsigned long long rhs);
379 // Get next/previous attribute in the attribute list of the parent node
380 xml_attribute next_attribute() const;
381 xml_attribute previous_attribute() const;
383 // Get hash value (unique for handles to the same object)
384 size_t hash_value() const;
386 // Get internal pointer
387 xml_attribute_struct* internal_object() const;
391 // Borland C++ workaround
392 bool PUGIXML_FUNCTION operator&&(const xml_attribute& lhs, bool rhs);
393 bool PUGIXML_FUNCTION operator||(const xml_attribute& lhs, bool rhs);
396 // A light-weight handle for manipulating nodes in DOM tree
397 class PUGIXML_CLASS xml_node
399 friend class xml_attribute_iterator;
400 friend class xml_node_iterator;
401 friend class xml_named_node_iterator;
404 xml_node_struct* _root;
406 typedef void (*unspecified_bool_type)(xml_node***);
409 // Default constructor. Constructs an empty node.
412 // Constructs node from internal pointer
413 explicit xml_node(xml_node_struct* p);
415 // Safe bool conversion operator
416 operator unspecified_bool_type() const;
418 // Borland C++ workaround
419 bool operator!() const;
421 // Comparison operators (compares wrapped node pointers)
422 bool operator==(const xml_node& r) const;
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;
429 // Check if node is empty.
433 xml_node_type type() const;
435 // Get node name, or "" if node is empty or it has no name
436 const char_t* name() const;
438 // Get node value, or "" if node is empty or it has no value
439 // Note: For <node>text</node> node.value() does not return "text"! Use child_value() or text() methods to access text inside nodes.
440 const char_t* value() const;
442 // Get attribute list
443 xml_attribute first_attribute() const;
444 xml_attribute last_attribute() const;
447 xml_node first_child() const;
448 xml_node last_child() const;
450 // Get next/previous sibling in the children list of the parent node
451 xml_node next_sibling() const;
452 xml_node previous_sibling() const;
455 xml_node parent() const;
457 // Get root of DOM tree this node belongs to
458 xml_node root() const;
460 // Get text object for the current node
461 xml_text text() const;
463 // Get child, attribute or next/previous sibling with the specified name
464 xml_node child(const char_t* name) const;
465 xml_attribute attribute(const char_t* name) const;
466 xml_node next_sibling(const char_t* name) const;
467 xml_node previous_sibling(const char_t* name) const;
469 // Get attribute, starting the search from a hint (and updating hint so that searching for a sequence of attributes is fast)
470 xml_attribute attribute(const char_t* name, xml_attribute& hint) const;
472 // Get child value of current node; that is, value of the first child node of type PCDATA/CDATA
473 const char_t* child_value() const;
475 // Get child value of child with specified name. Equivalent to child(name).child_value().
476 const char_t* child_value(const char_t* name) const;
478 // Set node name/value (returns false if node is empty, there is not enough memory, or node can not have name/value)
479 bool set_name(const char_t* rhs);
480 bool set_value(const char_t* rhs);
482 // Add attribute with specified name. Returns added attribute, or empty attribute on errors.
483 xml_attribute append_attribute(const char_t* name);
484 xml_attribute prepend_attribute(const char_t* name);
485 xml_attribute insert_attribute_after(const char_t* name, const xml_attribute& attr);
486 xml_attribute insert_attribute_before(const char_t* name, const xml_attribute& attr);
488 // Add a copy of the specified attribute. Returns added attribute, or empty attribute on errors.
489 xml_attribute append_copy(const xml_attribute& proto);
490 xml_attribute prepend_copy(const xml_attribute& proto);
491 xml_attribute insert_copy_after(const xml_attribute& proto, const xml_attribute& attr);
492 xml_attribute insert_copy_before(const xml_attribute& proto, const xml_attribute& attr);
494 // Add child node with specified type. Returns added node, or empty node on errors.
495 xml_node append_child(xml_node_type type = node_element);
496 xml_node prepend_child(xml_node_type type = node_element);
497 xml_node insert_child_after(xml_node_type type, const xml_node& node);
498 xml_node insert_child_before(xml_node_type type, const xml_node& node);
500 // Add child element with specified name. Returns added node, or empty node on errors.
501 xml_node append_child(const char_t* name);
502 xml_node prepend_child(const char_t* name);
503 xml_node insert_child_after(const char_t* name, const xml_node& node);
504 xml_node insert_child_before(const char_t* name, const xml_node& node);
506 // Add a copy of the specified node as a child. Returns added node, or empty node on errors.
507 xml_node append_copy(const xml_node& proto);
508 xml_node prepend_copy(const xml_node& proto);
509 xml_node insert_copy_after(const xml_node& proto, const xml_node& node);
510 xml_node insert_copy_before(const xml_node& proto, const xml_node& node);
512 // Move the specified node to become a child of this node. Returns moved node, or empty node on errors.
513 xml_node append_move(const xml_node& moved);
514 xml_node prepend_move(const xml_node& moved);
515 xml_node insert_move_after(const xml_node& moved, const xml_node& node);
516 xml_node insert_move_before(const xml_node& moved, const xml_node& node);
518 // Remove specified attribute
519 bool remove_attribute(const xml_attribute& a);
520 bool remove_attribute(const char_t* name);
522 // Remove specified child
523 bool remove_child(const xml_node& n);
524 bool remove_child(const char_t* name);
526 // Parses buffer as an XML document fragment and appends all nodes as children of the current node.
527 // Copies/converts the buffer, so it may be deleted or changed after the function returns.
528 // Note: append_buffer allocates memory that has the lifetime of the owning document; removing the appended nodes does not immediately reclaim that memory.
529 xml_parse_result append_buffer(const void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
531 // Find attribute using predicate. Returns first attribute for which predicate returned true.
532 template <typename Predicate> xml_attribute find_attribute(Predicate pred) const
534 if (!_root) return xml_attribute();
536 for (xml_attribute attrib = first_attribute(); attrib; attrib = attrib.next_attribute())
540 return xml_attribute();
543 // Find child node using predicate. Returns first child for which predicate returned true.
544 template <typename Predicate> xml_node find_child(Predicate pred) const
546 if (!_root) return xml_node();
548 for (xml_node node = first_child(); node; node = node.next_sibling())
555 // Find node from subtree using predicate. Returns first node from subtree (depth-first), for which predicate returned true.
556 template <typename Predicate> xml_node find_node(Predicate pred) const
558 if (!_root) return xml_node();
560 xml_node cur = first_child();
562 while (cur._root && cur._root != _root)
564 if (pred(cur)) return cur;
566 if (cur.first_child()) cur = cur.first_child();
567 else if (cur.next_sibling()) cur = cur.next_sibling();
570 while (!cur.next_sibling() && cur._root != _root) cur = cur.parent();
572 if (cur._root != _root) cur = cur.next_sibling();
579 // Find child node by attribute name/value
580 xml_node find_child_by_attribute(const char_t* name, const char_t* attr_name, const char_t* attr_value) const;
581 xml_node find_child_by_attribute(const char_t* attr_name, const char_t* attr_value) const;
583 #ifndef PUGIXML_NO_STL
584 // Get the absolute node path from root as a text string.
585 string_t path(char_t delimiter = '/') const;
588 // Search for a node by path consisting of node names and . or .. elements.
589 xml_node first_element_by_path(const char_t* path, char_t delimiter = '/') const;
591 // Recursively traverse subtree with xml_tree_walker
592 bool traverse(xml_tree_walker& walker);
594 #ifndef PUGIXML_NO_XPATH
595 // Select single node by evaluating XPath query. Returns first node from the resulting node set.
596 xpath_node select_node(const char_t* query, xpath_variable_set* variables = 0) const;
597 xpath_node select_node(const xpath_query& query) const;
599 // Select node set by evaluating XPath query
600 xpath_node_set select_nodes(const char_t* query, xpath_variable_set* variables = 0) const;
601 xpath_node_set select_nodes(const xpath_query& query) const;
603 // (deprecated: use select_node instead) Select single node by evaluating XPath query.
604 xpath_node select_single_node(const char_t* query, xpath_variable_set* variables = 0) const;
605 xpath_node select_single_node(const xpath_query& query) const;
609 // Print subtree using a writer object
610 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;
612 #ifndef PUGIXML_NO_STL
613 // Print subtree to stream
614 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;
615 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;
618 // Child nodes iterators
619 typedef xml_node_iterator iterator;
621 iterator begin() const;
622 iterator end() const;
624 // Attribute iterators
625 typedef xml_attribute_iterator attribute_iterator;
627 attribute_iterator attributes_begin() const;
628 attribute_iterator attributes_end() const;
630 // Range-based for support
631 xml_object_range<xml_node_iterator> children() const;
632 xml_object_range<xml_named_node_iterator> children(const char_t* name) const;
633 xml_object_range<xml_attribute_iterator> attributes() const;
635 // Get node offset in parsed file/string (in char_t units) for debugging purposes
636 ptrdiff_t offset_debug() const;
638 // Get hash value (unique for handles to the same object)
639 size_t hash_value() const;
641 // Get internal pointer
642 xml_node_struct* internal_object() const;
646 // Borland C++ workaround
647 bool PUGIXML_FUNCTION operator&&(const xml_node& lhs, bool rhs);
648 bool PUGIXML_FUNCTION operator||(const xml_node& lhs, bool rhs);
651 // A helper for working with text inside PCDATA nodes
652 class PUGIXML_CLASS xml_text
654 friend class xml_node;
656 xml_node_struct* _root;
658 typedef void (*unspecified_bool_type)(xml_text***);
660 explicit xml_text(xml_node_struct* root);
662 xml_node_struct* _data_new();
663 xml_node_struct* _data() const;
666 // Default constructor. Constructs an empty object.
669 // Safe bool conversion operator
670 operator unspecified_bool_type() const;
672 // Borland C++ workaround
673 bool operator!() const;
675 // Check if text object is empty
678 // Get text, or "" if object is empty
679 const char_t* get() const;
681 // Get text, or the default value if object is empty
682 const char_t* as_string(const char_t* def = PUGIXML_TEXT("")) const;
684 // Get text as a number, or the default value if conversion did not succeed or object is empty
685 int as_int(int def = 0) const;
686 unsigned int as_uint(unsigned int def = 0) const;
687 double as_double(double def = 0) const;
688 float as_float(float def = 0) const;
690 #ifdef PUGIXML_HAS_LONG_LONG
691 long long as_llong(long long def = 0) const;
692 unsigned long long as_ullong(unsigned long long def = 0) const;
695 // Get text as bool (returns true if first character is in '1tTyY' set), or the default value if object is empty
696 bool as_bool(bool def = false) const;
698 // Set text (returns false if object is empty or there is not enough memory)
699 bool set(const char_t* rhs);
701 // Set text with type conversion (numbers are converted to strings, boolean is converted to "true"/"false")
703 bool set(unsigned int rhs);
705 bool set(unsigned long rhs);
706 bool set(double rhs);
710 #ifdef PUGIXML_HAS_LONG_LONG
711 bool set(long long rhs);
712 bool set(unsigned long long rhs);
715 // Set text (equivalent to set without error checking)
716 xml_text& operator=(const char_t* rhs);
717 xml_text& operator=(int rhs);
718 xml_text& operator=(unsigned int rhs);
719 xml_text& operator=(long rhs);
720 xml_text& operator=(unsigned long rhs);
721 xml_text& operator=(double rhs);
722 xml_text& operator=(float rhs);
723 xml_text& operator=(bool rhs);
725 #ifdef PUGIXML_HAS_LONG_LONG
726 xml_text& operator=(long long rhs);
727 xml_text& operator=(unsigned long long rhs);
730 // Get the data node (node_pcdata or node_cdata) for this object
731 xml_node data() const;
735 // Borland C++ workaround
736 bool PUGIXML_FUNCTION operator&&(const xml_text& lhs, bool rhs);
737 bool PUGIXML_FUNCTION operator||(const xml_text& lhs, bool rhs);
740 // Child node iterator (a bidirectional iterator over a collection of xml_node)
741 class PUGIXML_CLASS xml_node_iterator
743 friend class xml_node;
746 mutable xml_node _wrap;
749 xml_node_iterator(xml_node_struct* ref, xml_node_struct* parent);
753 typedef ptrdiff_t difference_type;
754 typedef xml_node value_type;
755 typedef xml_node* pointer;
756 typedef xml_node& reference;
758 #ifndef PUGIXML_NO_STL
759 typedef std::bidirectional_iterator_tag iterator_category;
762 // Default constructor
765 // Construct an iterator which points to the specified node
766 xml_node_iterator(const xml_node& node);
768 // Iterator operators
769 bool operator==(const xml_node_iterator& rhs) const;
770 bool operator!=(const xml_node_iterator& rhs) const;
772 xml_node& operator*() const;
773 xml_node* operator->() const;
775 const xml_node_iterator& operator++();
776 xml_node_iterator operator++(int);
778 const xml_node_iterator& operator--();
779 xml_node_iterator operator--(int);
782 // Attribute iterator (a bidirectional iterator over a collection of xml_attribute)
783 class PUGIXML_CLASS xml_attribute_iterator
785 friend class xml_node;
788 mutable xml_attribute _wrap;
791 xml_attribute_iterator(xml_attribute_struct* ref, xml_node_struct* parent);
795 typedef ptrdiff_t difference_type;
796 typedef xml_attribute value_type;
797 typedef xml_attribute* pointer;
798 typedef xml_attribute& reference;
800 #ifndef PUGIXML_NO_STL
801 typedef std::bidirectional_iterator_tag iterator_category;
804 // Default constructor
805 xml_attribute_iterator();
807 // Construct an iterator which points to the specified attribute
808 xml_attribute_iterator(const xml_attribute& attr, const xml_node& parent);
810 // Iterator operators
811 bool operator==(const xml_attribute_iterator& rhs) const;
812 bool operator!=(const xml_attribute_iterator& rhs) const;
814 xml_attribute& operator*() const;
815 xml_attribute* operator->() const;
817 const xml_attribute_iterator& operator++();
818 xml_attribute_iterator operator++(int);
820 const xml_attribute_iterator& operator--();
821 xml_attribute_iterator operator--(int);
824 // Named node range helper
825 class PUGIXML_CLASS xml_named_node_iterator
827 friend class xml_node;
831 typedef ptrdiff_t difference_type;
832 typedef xml_node value_type;
833 typedef xml_node* pointer;
834 typedef xml_node& reference;
836 #ifndef PUGIXML_NO_STL
837 typedef std::bidirectional_iterator_tag iterator_category;
840 // Default constructor
841 xml_named_node_iterator();
843 // Construct an iterator which points to the specified node
844 xml_named_node_iterator(const xml_node& node, const char_t* name);
846 // Iterator operators
847 bool operator==(const xml_named_node_iterator& rhs) const;
848 bool operator!=(const xml_named_node_iterator& rhs) const;
850 xml_node& operator*() const;
851 xml_node* operator->() const;
853 const xml_named_node_iterator& operator++();
854 xml_named_node_iterator operator++(int);
856 const xml_named_node_iterator& operator--();
857 xml_named_node_iterator operator--(int);
860 mutable xml_node _wrap;
864 xml_named_node_iterator(xml_node_struct* ref, xml_node_struct* parent, const char_t* name);
867 // Abstract tree walker class (see xml_node::traverse)
868 class PUGIXML_CLASS xml_tree_walker
870 friend class xml_node;
876 // Get current traversal depth
881 virtual ~xml_tree_walker();
883 // Callback that is called when traversal begins
884 virtual bool begin(xml_node& node);
886 // Callback that is called for each node traversed
887 virtual bool for_each(xml_node& node) = 0;
889 // Callback that is called when traversal ends
890 virtual bool end(xml_node& node);
893 // Parsing status, returned as part of xml_parse_result object
894 enum xml_parse_status
896 status_ok = 0, // No error
898 status_file_not_found, // File was not found during load_file()
899 status_io_error, // Error reading from file/stream
900 status_out_of_memory, // Could not allocate memory
901 status_internal_error, // Internal error occurred
903 status_unrecognized_tag, // Parser could not determine tag type
905 status_bad_pi, // Parsing error occurred while parsing document declaration/processing instruction
906 status_bad_comment, // Parsing error occurred while parsing comment
907 status_bad_cdata, // Parsing error occurred while parsing CDATA section
908 status_bad_doctype, // Parsing error occurred while parsing document type declaration
909 status_bad_pcdata, // Parsing error occurred while parsing PCDATA section
910 status_bad_start_element, // Parsing error occurred while parsing start element tag
911 status_bad_attribute, // Parsing error occurred while parsing element attribute
912 status_bad_end_element, // Parsing error occurred while parsing end element tag
913 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)
915 status_append_invalid_root, // Unable to append nodes since root type is not node_element or node_document (exclusive to xml_node::append_buffer)
917 status_no_document_element // Parsing resulted in a document without element nodes
921 struct PUGIXML_CLASS xml_parse_result
923 // Parsing status (see xml_parse_status)
924 xml_parse_status status;
926 // Last parsed offset (in char_t units from start of input data)
929 // Source document encoding
930 xml_encoding encoding;
932 // Default constructor, initializes object to failed state
935 // Cast to bool operator
936 operator bool() const;
938 // Get error description
939 const char* description() const;
942 // Document class (DOM tree root)
943 class PUGIXML_CLASS xml_document: public xml_node
950 // Non-copyable semantics
951 xml_document(const xml_document&);
952 xml_document& operator=(const xml_document&);
958 // Default constructor, makes empty document
961 // Destructor, invalidates all node/attribute handles to this document
964 // Removes all nodes, leaving the empty document
967 // Removes all nodes, then copies the entire contents of the specified document
968 void reset(const xml_document& proto);
970 #ifndef PUGIXML_NO_STL
971 // Load document from stream.
972 xml_parse_result load(std::basic_istream<char, std::char_traits<char> >& stream, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
973 xml_parse_result load(std::basic_istream<wchar_t, std::char_traits<wchar_t> >& stream, unsigned int options = parse_default);
976 // (deprecated: use load_string instead) Load document from zero-terminated string. No encoding conversions are applied.
977 xml_parse_result load(const char_t* contents, unsigned int options = parse_default);
979 // Load document from zero-terminated string. No encoding conversions are applied.
980 xml_parse_result load_string(const char_t* contents, unsigned int options = parse_default);
982 // Load document from file
983 xml_parse_result load_file(const char* path, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
984 xml_parse_result load_file(const wchar_t* path, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
986 // Load document from buffer. Copies/converts the buffer, so it may be deleted or changed after the function returns.
987 xml_parse_result load_buffer(const void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
989 // Load document from buffer, using the buffer for in-place parsing (the buffer is modified and used for storage of document data).
990 // You should ensure that buffer data will persist throughout the document's lifetime, and free the buffer memory manually once document is destroyed.
991 xml_parse_result load_buffer_inplace(void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
993 // Load document from buffer, using the buffer for in-place parsing (the buffer is modified and used for storage of document data).
994 // 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).
995 xml_parse_result load_buffer_inplace_own(void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto);
997 // Save XML document to writer (semantics is slightly different from xml_node::print, see documentation for details).
998 void save(xml_writer& writer, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const;
1000 #ifndef PUGIXML_NO_STL
1001 // Save XML document to stream (semantics is slightly different from xml_node::print, see documentation for details).
1002 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;
1003 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;
1007 bool save_file(const char* path, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const;
1008 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;
1010 // Get document element
1011 xml_node document_element() const;
1014 #ifndef PUGIXML_NO_XPATH
1015 // XPath query return type
1016 enum xpath_value_type
1018 xpath_type_none, // Unknown type (query failed to compile)
1019 xpath_type_node_set, // Node set (xpath_node_set)
1020 xpath_type_number, // Number
1021 xpath_type_string, // String
1022 xpath_type_boolean // Boolean
1025 // XPath parsing result
1026 struct PUGIXML_CLASS xpath_parse_result
1028 // Error message (0 if no error)
1031 // Last parsed offset (in char_t units from string start)
1034 // Default constructor, initializes object to failed state
1035 xpath_parse_result();
1037 // Cast to bool operator
1038 operator bool() const;
1040 // Get error description
1041 const char* description() const;
1044 // A single XPath variable
1045 class PUGIXML_CLASS xpath_variable
1047 friend class xpath_variable_set;
1050 xpath_value_type _type;
1051 xpath_variable* _next;
1053 xpath_variable(xpath_value_type type);
1055 // Non-copyable semantics
1056 xpath_variable(const xpath_variable&);
1057 xpath_variable& operator=(const xpath_variable&);
1060 // Get variable name
1061 const char_t* name() const;
1063 // Get variable type
1064 xpath_value_type type() const;
1066 // Get variable value; no type conversion is performed, default value (false, NaN, empty string, empty node set) is returned on type mismatch error
1067 bool get_boolean() const;
1068 double get_number() const;
1069 const char_t* get_string() const;
1070 const xpath_node_set& get_node_set() const;
1072 // Set variable value; no type conversion is performed, false is returned on type mismatch error
1073 bool set(bool value);
1074 bool set(double value);
1075 bool set(const char_t* value);
1076 bool set(const xpath_node_set& value);
1079 // A set of XPath variables
1080 class PUGIXML_CLASS xpath_variable_set
1083 xpath_variable* _data[64];
1085 void _assign(const xpath_variable_set& rhs);
1086 void _swap(xpath_variable_set& rhs);
1088 xpath_variable* _find(const char_t* name) const;
1090 static bool _clone(xpath_variable* var, xpath_variable** out_result);
1091 static void _destroy(xpath_variable* var);
1094 // Default constructor/destructor
1095 xpath_variable_set();
1096 ~xpath_variable_set();
1098 // Copy constructor/assignment operator
1099 xpath_variable_set(const xpath_variable_set& rhs);
1100 xpath_variable_set& operator=(const xpath_variable_set& rhs);
1102 #if __cplusplus >= 201103
1103 // Move semantics support
1104 xpath_variable_set(xpath_variable_set&& rhs);
1105 xpath_variable_set& operator=(xpath_variable_set&& rhs);
1108 // Add a new variable or get the existing one, if the types match
1109 xpath_variable* add(const char_t* name, xpath_value_type type);
1111 // Set value of an existing variable; no type conversion is performed, false is returned if there is no such variable or if types mismatch
1112 bool set(const char_t* name, bool value);
1113 bool set(const char_t* name, double value);
1114 bool set(const char_t* name, const char_t* value);
1115 bool set(const char_t* name, const xpath_node_set& value);
1117 // Get existing variable by name
1118 xpath_variable* get(const char_t* name);
1119 const xpath_variable* get(const char_t* name) const;
1122 // A compiled XPath query object
1123 class PUGIXML_CLASS xpath_query
1127 xpath_parse_result _result;
1129 typedef void (*unspecified_bool_type)(xpath_query***);
1131 // Non-copyable semantics
1132 xpath_query(const xpath_query&);
1133 xpath_query& operator=(const xpath_query&);
1136 // Construct a compiled object from XPath expression.
1137 // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on compilation errors.
1138 explicit xpath_query(const char_t* query, xpath_variable_set* variables = 0);
1146 #if __cplusplus >= 201103
1147 // Move semantics support
1148 xpath_query(xpath_query&& rhs);
1149 xpath_query& operator=(xpath_query&& rhs);
1152 // Get query expression return type
1153 xpath_value_type return_type() const;
1155 // Evaluate expression as boolean value in the specified context; performs type conversion if necessary.
1156 // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors.
1157 bool evaluate_boolean(const xpath_node& n) const;
1159 // Evaluate expression as double value in the specified context; performs type conversion if necessary.
1160 // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors.
1161 double evaluate_number(const xpath_node& n) const;
1163 #ifndef PUGIXML_NO_STL
1164 // Evaluate expression as string value in the specified context; performs type conversion if necessary.
1165 // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors.
1166 string_t evaluate_string(const xpath_node& n) const;
1169 // Evaluate expression as string value in the specified context; performs type conversion if necessary.
1170 // At most capacity characters are written to the destination buffer, full result size is returned (includes terminating zero).
1171 // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors.
1172 // If PUGIXML_NO_EXCEPTIONS is defined, returns empty set instead.
1173 size_t evaluate_string(char_t* buffer, size_t capacity, const xpath_node& n) const;
1175 // Evaluate expression as node set in the specified context.
1176 // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on type mismatch and std::bad_alloc on out of memory errors.
1177 // If PUGIXML_NO_EXCEPTIONS is defined, returns empty node set instead.
1178 xpath_node_set evaluate_node_set(const xpath_node& n) const;
1180 // Evaluate expression as node set in the specified context.
1181 // Return first node in document order, or empty node if node set is empty.
1182 // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on type mismatch and std::bad_alloc on out of memory errors.
1183 // If PUGIXML_NO_EXCEPTIONS is defined, returns empty node instead.
1184 xpath_node evaluate_node(const xpath_node& n) const;
1186 // Get parsing result (used to get compilation errors in PUGIXML_NO_EXCEPTIONS mode)
1187 const xpath_parse_result& result() const;
1189 // Safe bool conversion operator
1190 operator unspecified_bool_type() const;
1192 // Borland C++ workaround
1193 bool operator!() const;
1196 #ifndef PUGIXML_NO_EXCEPTIONS
1197 // XPath exception class
1198 class PUGIXML_CLASS xpath_exception: public std::exception
1201 xpath_parse_result _result;
1204 // Construct exception from parse result
1205 explicit xpath_exception(const xpath_parse_result& result);
1207 // Get error message
1208 virtual const char* what() const throw();
1211 const xpath_parse_result& result() const;
1215 // XPath node class (either xml_node or xml_attribute)
1216 class PUGIXML_CLASS xpath_node
1220 xml_attribute _attribute;
1222 typedef void (*unspecified_bool_type)(xpath_node***);
1225 // Default constructor; constructs empty XPath node
1228 // Construct XPath node from XML node/attribute
1229 xpath_node(const xml_node& node);
1230 xpath_node(const xml_attribute& attribute, const xml_node& parent);
1232 // Get node/attribute, if any
1233 xml_node node() const;
1234 xml_attribute attribute() const;
1236 // Get parent of contained node/attribute
1237 xml_node parent() const;
1239 // Safe bool conversion operator
1240 operator unspecified_bool_type() const;
1242 // Borland C++ workaround
1243 bool operator!() const;
1245 // Comparison operators
1246 bool operator==(const xpath_node& n) const;
1247 bool operator!=(const xpath_node& n) const;
1251 // Borland C++ workaround
1252 bool PUGIXML_FUNCTION operator&&(const xpath_node& lhs, bool rhs);
1253 bool PUGIXML_FUNCTION operator||(const xpath_node& lhs, bool rhs);
1256 // A fixed-size collection of XPath nodes
1257 class PUGIXML_CLASS xpath_node_set
1263 type_unsorted, // Not ordered
1264 type_sorted, // Sorted by document order (ascending)
1265 type_sorted_reverse // Sorted by document order (descending)
1268 // Constant iterator type
1269 typedef const xpath_node* const_iterator;
1271 // We define non-constant iterator to be the same as constant iterator so that various generic algorithms (i.e. boost foreach) work
1272 typedef const xpath_node* iterator;
1274 // Default constructor. Constructs empty set.
1277 // Constructs a set from iterator range; data is not checked for duplicates and is not sorted according to provided type, so be careful
1278 xpath_node_set(const_iterator begin, const_iterator end, type_t type = type_unsorted);
1283 // Copy constructor/assignment operator
1284 xpath_node_set(const xpath_node_set& ns);
1285 xpath_node_set& operator=(const xpath_node_set& ns);
1287 #if __cplusplus >= 201103
1288 // Move semantics support
1289 xpath_node_set(xpath_node_set&& rhs);
1290 xpath_node_set& operator=(xpath_node_set&& rhs);
1293 // Get collection type
1294 type_t type() const;
1296 // Get collection size
1297 size_t size() const;
1299 // Indexing operator
1300 const xpath_node& operator[](size_t index) const;
1302 // Collection iterators
1303 const_iterator begin() const;
1304 const_iterator end() const;
1306 // Sort the collection in ascending/descending order by document order
1307 void sort(bool reverse = false);
1309 // Get first node in the collection by document order
1310 xpath_node first() const;
1312 // Check if collection is empty
1318 xpath_node _storage;
1323 void _assign(const_iterator begin, const_iterator end, type_t type);
1324 void _move(xpath_node_set& rhs);
1328 #ifndef PUGIXML_NO_STL
1329 // Convert wide string to UTF8
1330 std::basic_string<char, std::char_traits<char>, std::allocator<char> > PUGIXML_FUNCTION as_utf8(const wchar_t* str);
1331 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);
1333 // Convert UTF8 to wide string
1334 std::basic_string<wchar_t, std::char_traits<wchar_t>, std::allocator<wchar_t> > PUGIXML_FUNCTION as_wide(const char* str);
1335 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);
1338 // Memory allocation function interface; returns pointer to allocated memory or NULL on failure
1339 typedef void* (*allocation_function)(size_t size);
1341 // Memory deallocation function interface
1342 typedef void (*deallocation_function)(void* ptr);
1344 // Override default memory management functions. All subsequent allocations/deallocations will be performed via supplied functions.
1345 void PUGIXML_FUNCTION set_memory_management_functions(allocation_function allocate, deallocation_function deallocate);
1347 // Get current memory management functions
1348 allocation_function PUGIXML_FUNCTION get_memory_allocation_function();
1349 deallocation_function PUGIXML_FUNCTION get_memory_deallocation_function();
1352 #if !defined(PUGIXML_NO_STL) && (defined(_MSC_VER) || defined(__ICC))
1355 // Workarounds for (non-standard) iterator category detection for older versions (MSVC7/IC8 and earlier)
1356 std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_node_iterator&);
1357 std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_attribute_iterator&);
1358 std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_named_node_iterator&);
1362 #if !defined(PUGIXML_NO_STL) && defined(__SUNPRO_CC)
1365 // Workarounds for (non-standard) iterator category detection
1366 std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_node_iterator&);
1367 std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_attribute_iterator&);
1368 std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_named_node_iterator&);
1374 // Make sure implementation is included in header-only mode
1375 // Use macro expansion in #include to work around QMake (QTBUG-11923)
1376 #if defined(PUGIXML_HEADER_ONLY) && !defined(PUGIXML_SOURCE)
1377 # define PUGIXML_SOURCE "pugixml.cpp"
1378 # include PUGIXML_SOURCE
1382 * Copyright (c) 2006-2016 Arseny Kapoulkine
1384 * Permission is hereby granted, free of charge, to any person
1385 * obtaining a copy of this software and associated documentation
1386 * files (the "Software"), to deal in the Software without
1387 * restriction, including without limitation the rights to use,
1388 * copy, modify, merge, publish, distribute, sublicense, and/or sell
1389 * copies of the Software, and to permit persons to whom the
1390 * Software is furnished to do so, subject to the following
1393 * The above copyright notice and this permission notice shall be
1394 * included in all copies or substantial portions of the Software.
1396 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
1397 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
1398 * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
1399 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
1400 * HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
1401 * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
1402 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
1403 * OTHER DEALINGS IN THE SOFTWARE.