1 /* gmarkup.c - Simple XML-like parser
3 * Copyright 2000, 2003 Red Hat, Inc.
4 * Copyright 2007, 2008 Ryan Lortie <desrt@desrt.ca>
6 * GLib is free software; you can redistribute it and/or modify it
7 * under the terms of the GNU Lesser General Public License as
8 * published by the Free Software Foundation; either version 2 of the
9 * License, or (at your option) any later version.
11 * GLib is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with GLib; see the file COPYING.LIB. If not,
18 * write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 * Boston, MA 02111-1307, USA.
35 #include "gstrfuncs.h"
37 #include "gtestutils.h"
43 * @Title: Simple XML Subset Parser
44 * @Short_description: parses a subset of XML
45 * @See_also: <ulink url="http://www.w3.org/TR/REC-xml/">XML
46 * Specification</ulink>
48 * The "GMarkup" parser is intended to parse a simple markup format
49 * that's a subset of XML. This is a small, efficient, easy-to-use
50 * parser. It should not be used if you expect to interoperate with
51 * other applications generating full-scale XML. However, it's very
52 * useful for application data files, config files, etc. where you
53 * know your application will be the only one writing the file.
54 * Full-scale XML parsers should be able to parse the subset used by
55 * GMarkup, so you can easily migrate to full-scale XML at a later
56 * time if the need arises.
58 * GMarkup is not guaranteed to signal an error on all invalid XML;
59 * the parser may accept documents that an XML parser would not.
60 * However, XML documents which are not well-formed<footnote
61 * id="wellformed">Being wellformed is a weaker condition than being
62 * valid. See the <ulink url="http://www.w3.org/TR/REC-xml/">XML
63 * specification</ulink> for definitions of these terms.</footnote>
64 * are not considered valid GMarkup documents.
66 * Simplifications to XML include:
68 * <listitem>Only UTF-8 encoding is allowed</listitem>
69 * <listitem>No user-defined entities</listitem>
70 * <listitem>Processing instructions, comments and the doctype declaration
71 * are "passed through" but are not interpreted in any way</listitem>
72 * <listitem>No DTD or validation.</listitem>
75 * The markup format does support:
77 * <listitem>Elements</listitem>
78 * <listitem>Attributes</listitem>
79 * <listitem>5 standard entities:
80 * <literal>&amp; &lt; &gt; &quot; &apos;</literal>
82 * <listitem>Character references</listitem>
83 * <listitem>Sections marked as CDATA</listitem>
87 G_DEFINE_QUARK (g-markup-error-quark, g_markup_error)
92 STATE_AFTER_OPEN_ANGLE,
93 STATE_AFTER_CLOSE_ANGLE,
94 STATE_AFTER_ELISION_SLASH, /* the slash that obviates need for end element */
95 STATE_INSIDE_OPEN_TAG_NAME,
96 STATE_INSIDE_ATTRIBUTE_NAME,
97 STATE_AFTER_ATTRIBUTE_NAME,
98 STATE_BETWEEN_ATTRIBUTES,
99 STATE_AFTER_ATTRIBUTE_EQUALS_SIGN,
100 STATE_INSIDE_ATTRIBUTE_VALUE_SQ,
101 STATE_INSIDE_ATTRIBUTE_VALUE_DQ,
103 STATE_AFTER_CLOSE_TAG_SLASH,
104 STATE_INSIDE_CLOSE_TAG_NAME,
105 STATE_AFTER_CLOSE_TAG_NAME,
106 STATE_INSIDE_PASSTHROUGH,
112 const char *prev_element;
113 const GMarkupParser *prev_parser;
114 gpointer prev_user_data;
115 } GMarkupRecursionTracker;
117 struct _GMarkupParseContext
119 const GMarkupParser *parser;
121 volatile gint ref_count;
123 GMarkupParseFlags flags;
128 GMarkupParseState state;
131 GDestroyNotify dnotify;
133 /* A piece of character data or an element that
134 * hasn't "ended" yet so we haven't yet called
135 * the callback for it.
137 GString *partial_chunk;
138 GSList *spare_chunks;
141 GSList *tag_stack_gstr;
142 GSList *spare_list_nodes;
144 GString **attr_names;
145 GString **attr_values;
149 const gchar *current_text;
150 gssize current_text_len;
151 const gchar *current_text_end;
153 /* used to save the start of the last interesting thingy */
158 guint document_empty : 1;
160 guint awaiting_pop : 1;
163 /* subparser support */
164 GSList *subparser_stack; /* (GMarkupRecursionTracker *) */
165 const char *subparser_element;
166 gpointer held_user_data;
170 * Helpers to reduce our allocation overhead, we have
171 * a well defined allocation lifecycle.
174 get_list_node (GMarkupParseContext *context, gpointer data)
177 if (context->spare_list_nodes != NULL)
179 node = context->spare_list_nodes;
180 context->spare_list_nodes = g_slist_remove_link (context->spare_list_nodes, node);
183 node = g_slist_alloc();
189 free_list_node (GMarkupParseContext *context, GSList *node)
192 context->spare_list_nodes = g_slist_concat (node, context->spare_list_nodes);
196 string_blank (GString *string)
198 string->str[0] = '\0';
203 * g_markup_parse_context_new:
204 * @parser: a #GMarkupParser
205 * @flags: one or more #GMarkupParseFlags
206 * @user_data: user data to pass to #GMarkupParser functions
207 * @user_data_dnotify: user data destroy notifier called when
208 * the parse context is freed
210 * Creates a new parse context. A parse context is used to parse
211 * marked-up documents. You can feed any number of documents into
212 * a context, as long as no errors occur; once an error occurs,
213 * the parse context can't continue to parse text (you have to
214 * free it and create a new parse context).
216 * Return value: a new #GMarkupParseContext
218 GMarkupParseContext *
219 g_markup_parse_context_new (const GMarkupParser *parser,
220 GMarkupParseFlags flags,
222 GDestroyNotify user_data_dnotify)
224 GMarkupParseContext *context;
226 g_return_val_if_fail (parser != NULL, NULL);
228 context = g_new (GMarkupParseContext, 1);
230 context->ref_count = 1;
231 context->parser = parser;
232 context->flags = flags;
233 context->user_data = user_data;
234 context->dnotify = user_data_dnotify;
236 context->line_number = 1;
237 context->char_number = 1;
239 context->partial_chunk = NULL;
240 context->spare_chunks = NULL;
241 context->spare_list_nodes = NULL;
243 context->state = STATE_START;
244 context->tag_stack = NULL;
245 context->tag_stack_gstr = NULL;
246 context->attr_names = NULL;
247 context->attr_values = NULL;
248 context->cur_attr = -1;
249 context->alloc_attrs = 0;
251 context->current_text = NULL;
252 context->current_text_len = -1;
253 context->current_text_end = NULL;
255 context->start = NULL;
256 context->iter = NULL;
258 context->document_empty = TRUE;
259 context->parsing = FALSE;
261 context->awaiting_pop = FALSE;
262 context->subparser_stack = NULL;
263 context->subparser_element = NULL;
265 /* this is only looked at if awaiting_pop = TRUE. initialise anyway. */
266 context->held_user_data = NULL;
268 context->balance = 0;
274 * g_markup_parse_context_ref:
275 * @context: a #GMarkupParseContext
277 * Increases the reference count of @context.
279 * Returns: the same @context
283 GMarkupParseContext *
284 g_markup_parse_context_ref (GMarkupParseContext *context)
286 g_return_val_if_fail (context != NULL, NULL);
287 g_return_val_if_fail (context->ref_count > 0, NULL);
289 g_atomic_int_inc (&context->ref_count);
295 * g_markup_parse_context_unref:
296 * @context: a #GMarkupParseContext
298 * Decreases the reference count of @context. When its reference count
299 * drops to 0, it is freed.
304 g_markup_parse_context_unref (GMarkupParseContext *context)
306 g_return_if_fail (context != NULL);
307 g_return_if_fail (context->ref_count > 0);
309 if (g_atomic_int_dec_and_test (&context->ref_count))
310 g_markup_parse_context_free (context);
314 string_full_free (gpointer ptr)
316 g_string_free (ptr, TRUE);
319 static void clear_attributes (GMarkupParseContext *context);
322 * g_markup_parse_context_free:
323 * @context: a #GMarkupParseContext
325 * Frees a #GMarkupParseContext.
327 * This function can't be called from inside one of the
328 * #GMarkupParser functions or while a subparser is pushed.
331 g_markup_parse_context_free (GMarkupParseContext *context)
333 g_return_if_fail (context != NULL);
334 g_return_if_fail (!context->parsing);
335 g_return_if_fail (!context->subparser_stack);
336 g_return_if_fail (!context->awaiting_pop);
338 if (context->dnotify)
339 (* context->dnotify) (context->user_data);
341 clear_attributes (context);
342 g_free (context->attr_names);
343 g_free (context->attr_values);
345 g_slist_free_full (context->tag_stack_gstr, string_full_free);
346 g_slist_free (context->tag_stack);
348 g_slist_free_full (context->spare_chunks, string_full_free);
349 g_slist_free (context->spare_list_nodes);
351 if (context->partial_chunk)
352 g_string_free (context->partial_chunk, TRUE);
357 static void pop_subparser_stack (GMarkupParseContext *context);
360 mark_error (GMarkupParseContext *context,
363 context->state = STATE_ERROR;
365 if (context->parser->error)
366 (*context->parser->error) (context, error, context->user_data);
368 /* report the error all the way up to free all the user-data */
369 while (context->subparser_stack)
371 pop_subparser_stack (context);
372 context->awaiting_pop = FALSE; /* already been freed */
374 if (context->parser->error)
375 (*context->parser->error) (context, error, context->user_data);
380 set_error (GMarkupParseContext *context,
384 ...) G_GNUC_PRINTF (4, 5);
387 set_error_literal (GMarkupParseContext *context,
390 const gchar *message)
394 tmp_error = g_error_new_literal (G_MARKUP_ERROR, code, message);
396 g_prefix_error (&tmp_error,
397 _("Error on line %d char %d: "),
398 context->line_number,
399 context->char_number);
401 mark_error (context, tmp_error);
403 g_propagate_error (error, tmp_error);
407 set_error (GMarkupParseContext *context,
417 va_start (args, format);
418 s = g_strdup_vprintf (format, args);
421 /* Make sure that the GError message is valid UTF-8
422 * even if it is complaining about invalid UTF-8 in the markup
424 s_valid = _g_utf8_make_valid (s);
425 set_error_literal (context, error, code, s);
432 propagate_error (GMarkupParseContext *context,
436 if (context->flags & G_MARKUP_PREFIX_ERROR_POSITION)
437 g_prefix_error (&src,
438 _("Error on line %d char %d: "),
439 context->line_number,
440 context->char_number);
442 mark_error (context, src);
444 g_propagate_error (dest, src);
447 #define IS_COMMON_NAME_END_CHAR(c) \
448 ((c) == '=' || (c) == '/' || (c) == '>' || (c) == ' ')
451 slow_name_validate (GMarkupParseContext *context,
455 const gchar *p = name;
457 if (!g_utf8_validate (name, strlen (name), NULL))
459 set_error (context, error, G_MARKUP_ERROR_BAD_UTF8,
460 _("Invalid UTF-8 encoded text in name - not valid '%s'"), name);
464 if (!(g_ascii_isalpha (*p) ||
465 (!IS_COMMON_NAME_END_CHAR (*p) &&
468 g_unichar_isalpha (g_utf8_get_char (p))))))
470 set_error (context, error, G_MARKUP_ERROR_PARSE,
471 _("'%s' is not a valid name"), name);
475 for (p = g_utf8_next_char (name); *p != '\0'; p = g_utf8_next_char (p))
478 if (!(g_ascii_isalnum (*p) ||
479 (!IS_COMMON_NAME_END_CHAR (*p) &&
484 g_unichar_isalpha (g_utf8_get_char (p))))))
486 set_error (context, error, G_MARKUP_ERROR_PARSE,
487 _("'%s' is not a valid name: '%c'"), name, *p);
495 * Use me for elements, attributes etc.
498 name_validate (GMarkupParseContext *context,
505 /* name start char */
507 if (G_UNLIKELY (IS_COMMON_NAME_END_CHAR (*p) ||
508 !(g_ascii_isalpha (*p) || *p == '_' || *p == ':')))
511 for (mask = *p++; *p != '\0'; p++)
516 if (G_UNLIKELY (!(g_ascii_isalnum (*p) ||
517 (!IS_COMMON_NAME_END_CHAR (*p) &&
525 if (mask & 0x80) /* un-common / non-ascii */
531 return slow_name_validate (context, name, error);
535 text_validate (GMarkupParseContext *context,
540 if (!g_utf8_validate (p, len, NULL))
542 set_error (context, error, G_MARKUP_ERROR_BAD_UTF8,
543 _("Invalid UTF-8 encoded text in name - not valid '%s'"), p);
551 char_str (gunichar c,
555 g_unichar_to_utf8 (c, buf);
560 utf8_str (const gchar *utf8,
563 char_str (g_utf8_get_char (utf8), buf);
568 set_unescape_error (GMarkupParseContext *context,
570 const gchar *remaining_text,
578 gint remaining_newlines;
581 remaining_newlines = 0;
586 ++remaining_newlines;
590 va_start (args, format);
591 s = g_strdup_vprintf (format, args);
594 tmp_error = g_error_new (G_MARKUP_ERROR,
596 _("Error on line %d: %s"),
597 context->line_number - remaining_newlines,
602 mark_error (context, tmp_error);
604 g_propagate_error (error, tmp_error);
608 * re-write the GString in-place, unescaping anything that escaped.
609 * most XML does not contain entities, or escaping.
612 unescape_gstring_inplace (GMarkupParseContext *context,
620 gboolean normalize_attribute;
624 /* are we unescaping an attribute or not ? */
625 if (context->state == STATE_INSIDE_ATTRIBUTE_VALUE_SQ ||
626 context->state == STATE_INSIDE_ATTRIBUTE_VALUE_DQ)
627 normalize_attribute = TRUE;
629 normalize_attribute = FALSE;
632 * Meeks' theorum: unescaping can only shrink text.
633 * for < etc. this is obvious, for  more
634 * thought is required, but this is patently so.
637 for (from = to = string->str; *from != '\0'; from++, to++)
644 if (normalize_attribute && (*to == '\t' || *to == '\n'))
648 *to = normalize_attribute ? ' ' : '\n';
657 gboolean is_hex = FALSE;
669 /* digit is between start and p */
672 l = strtoul (from, &end, 16);
674 l = strtoul (from, &end, 10);
676 if (end == from || errno != 0)
678 set_unescape_error (context, error,
679 from, G_MARKUP_ERROR_PARSE,
680 _("Failed to parse '%-.*s', which "
681 "should have been a digit "
682 "inside a character reference "
683 "(ê for example) - perhaps "
684 "the digit is too large"),
688 else if (*end != ';')
690 set_unescape_error (context, error,
691 from, G_MARKUP_ERROR_PARSE,
692 _("Character reference did not end with a "
694 "most likely you used an ampersand "
695 "character without intending to start "
696 "an entity - escape ampersand as &"));
701 /* characters XML 1.1 permits */
702 if ((0 < l && l <= 0xD7FF) ||
703 (0xE000 <= l && l <= 0xFFFD) ||
704 (0x10000 <= l && l <= 0x10FFFF))
709 to += strlen (buf) - 1;
711 if (l >= 0x80) /* not ascii */
716 set_unescape_error (context, error,
717 from, G_MARKUP_ERROR_PARSE,
718 _("Character reference '%-.*s' does not "
719 "encode a permitted character"),
726 else if (strncmp (from, "lt;", 3) == 0)
731 else if (strncmp (from, "gt;", 3) == 0)
736 else if (strncmp (from, "amp;", 4) == 0)
741 else if (strncmp (from, "quot;", 5) == 0)
746 else if (strncmp (from, "apos;", 5) == 0)
754 set_unescape_error (context, error,
755 from, G_MARKUP_ERROR_PARSE,
756 _("Empty entity '&;' seen; valid "
757 "entities are: & " < > '"));
760 const char *end = strchr (from, ';');
762 set_unescape_error (context, error,
763 from, G_MARKUP_ERROR_PARSE,
764 _("Entity name '%-.*s' is not known"),
767 set_unescape_error (context, error,
768 from, G_MARKUP_ERROR_PARSE,
769 _("Entity did not end with a semicolon; "
770 "most likely you used an ampersand "
771 "character without intending to start "
772 "an entity - escape ampersand as &"));
779 g_assert (to - string->str <= string->len);
780 if (to - string->str != string->len)
781 g_string_truncate (string, to - string->str);
783 *is_ascii = !(mask & 0x80);
788 static inline gboolean
789 advance_char (GMarkupParseContext *context)
792 context->char_number++;
794 if (G_UNLIKELY (context->iter == context->current_text_end))
797 else if (G_UNLIKELY (*context->iter == '\n'))
799 context->line_number++;
800 context->char_number = 1;
806 static inline gboolean
809 return c == ' ' || c == '\t' || c == '\n' || c == '\r';
813 skip_spaces (GMarkupParseContext *context)
817 if (!xml_isspace (*context->iter))
820 while (advance_char (context));
824 advance_to_name_end (GMarkupParseContext *context)
828 if (IS_COMMON_NAME_END_CHAR (*(context->iter)))
830 if (xml_isspace (*(context->iter)))
833 while (advance_char (context));
837 release_chunk (GMarkupParseContext *context, GString *str)
842 if (str->allocated_len > 256)
843 { /* large strings are unusual and worth freeing */
844 g_string_free (str, TRUE);
848 node = get_list_node (context, str);
849 context->spare_chunks = g_slist_concat (node, context->spare_chunks);
853 add_to_partial (GMarkupParseContext *context,
854 const gchar *text_start,
855 const gchar *text_end)
857 if (context->partial_chunk == NULL)
858 { /* allocate a new chunk to parse into */
860 if (context->spare_chunks != NULL)
862 GSList *node = context->spare_chunks;
863 context->spare_chunks = g_slist_remove_link (context->spare_chunks, node);
864 context->partial_chunk = node->data;
865 free_list_node (context, node);
868 context->partial_chunk = g_string_sized_new (MAX (28, text_end - text_start));
871 if (text_start != text_end)
872 g_string_insert_len (context->partial_chunk, -1,
873 text_start, text_end - text_start);
877 truncate_partial (GMarkupParseContext *context)
879 if (context->partial_chunk != NULL)
880 string_blank (context->partial_chunk);
883 static inline const gchar*
884 current_element (GMarkupParseContext *context)
886 return context->tag_stack->data;
890 pop_subparser_stack (GMarkupParseContext *context)
892 GMarkupRecursionTracker *tracker;
894 g_assert (context->subparser_stack);
896 tracker = context->subparser_stack->data;
898 context->awaiting_pop = TRUE;
899 context->held_user_data = context->user_data;
901 context->user_data = tracker->prev_user_data;
902 context->parser = tracker->prev_parser;
903 context->subparser_element = tracker->prev_element;
904 g_slice_free (GMarkupRecursionTracker, tracker);
906 context->subparser_stack = g_slist_delete_link (context->subparser_stack,
907 context->subparser_stack);
911 push_partial_as_tag (GMarkupParseContext *context)
913 GString *str = context->partial_chunk;
914 /* sadly, this is exported by gmarkup_get_element_stack as-is */
915 context->tag_stack = g_slist_concat (get_list_node (context, str->str), context->tag_stack);
916 context->tag_stack_gstr = g_slist_concat (get_list_node (context, str), context->tag_stack_gstr);
917 context->partial_chunk = NULL;
921 pop_tag (GMarkupParseContext *context)
923 GSList *nodea, *nodeb;
925 nodea = context->tag_stack;
926 nodeb = context->tag_stack_gstr;
927 release_chunk (context, nodeb->data);
928 context->tag_stack = g_slist_remove_link (context->tag_stack, nodea);
929 context->tag_stack_gstr = g_slist_remove_link (context->tag_stack_gstr, nodeb);
930 free_list_node (context, nodea);
931 free_list_node (context, nodeb);
935 possibly_finish_subparser (GMarkupParseContext *context)
937 if (current_element (context) == context->subparser_element)
938 pop_subparser_stack (context);
942 ensure_no_outstanding_subparser (GMarkupParseContext *context)
944 if (context->awaiting_pop)
945 g_critical ("During the first end_element call after invoking a "
946 "subparser you must pop the subparser stack and handle "
947 "the freeing of the subparser user_data. This can be "
948 "done by calling the end function of the subparser. "
949 "Very probably, your program just leaked memory.");
951 /* let valgrind watch the pointer disappear... */
952 context->held_user_data = NULL;
953 context->awaiting_pop = FALSE;
957 current_attribute (GMarkupParseContext *context)
959 g_assert (context->cur_attr >= 0);
960 return context->attr_names[context->cur_attr]->str;
964 add_attribute (GMarkupParseContext *context, GString *str)
966 if (context->cur_attr + 2 >= context->alloc_attrs)
968 context->alloc_attrs += 5; /* silly magic number */
969 context->attr_names = g_realloc (context->attr_names, sizeof(GString*)*context->alloc_attrs);
970 context->attr_values = g_realloc (context->attr_values, sizeof(GString*)*context->alloc_attrs);
973 context->attr_names[context->cur_attr] = str;
974 context->attr_values[context->cur_attr] = NULL;
975 context->attr_names[context->cur_attr+1] = NULL;
976 context->attr_values[context->cur_attr+1] = NULL;
980 clear_attributes (GMarkupParseContext *context)
982 /* Go ahead and free the attributes. */
983 for (; context->cur_attr >= 0; context->cur_attr--)
985 int pos = context->cur_attr;
986 release_chunk (context, context->attr_names[pos]);
987 release_chunk (context, context->attr_values[pos]);
988 context->attr_names[pos] = context->attr_values[pos] = NULL;
990 g_assert (context->cur_attr == -1);
991 g_assert (context->attr_names == NULL ||
992 context->attr_names[0] == NULL);
993 g_assert (context->attr_values == NULL ||
994 context->attr_values[0] == NULL);
997 /* This has to be a separate function to ensure the alloca's
998 * are unwound on exit - otherwise we grow & blow the stack
999 * with large documents
1002 emit_start_element (GMarkupParseContext *context,
1006 const gchar *start_name;
1007 const gchar **attr_names;
1008 const gchar **attr_values;
1011 attr_names = g_newa (const gchar *, context->cur_attr + 2);
1012 attr_values = g_newa (const gchar *, context->cur_attr + 2);
1013 for (i = 0; i < context->cur_attr + 1; i++)
1015 attr_names[i] = context->attr_names[i]->str;
1016 attr_values[i] = context->attr_values[i]->str;
1018 attr_names[i] = NULL;
1019 attr_values[i] = NULL;
1021 /* Call user callback for element start */
1023 start_name = current_element (context);
1025 if (context->parser->start_element &&
1026 name_validate (context, start_name, error))
1027 (* context->parser->start_element) (context,
1029 (const gchar **)attr_names,
1030 (const gchar **)attr_values,
1033 clear_attributes (context);
1035 if (tmp_error != NULL)
1036 propagate_error (context, error, tmp_error);
1040 * g_markup_parse_context_parse:
1041 * @context: a #GMarkupParseContext
1042 * @text: chunk of text to parse
1043 * @text_len: length of @text in bytes
1044 * @error: return location for a #GError
1046 * Feed some data to the #GMarkupParseContext.
1048 * The data need not be valid UTF-8; an error will be signaled if
1049 * it's invalid. The data need not be an entire document; you can
1050 * feed a document into the parser incrementally, via multiple calls
1051 * to this function. Typically, as you receive data from a network
1052 * connection or file, you feed each received chunk of data into this
1053 * function, aborting the process if an error occurs. Once an error
1054 * is reported, no further data may be fed to the #GMarkupParseContext;
1055 * all errors are fatal.
1057 * Return value: %FALSE if an error occurred, %TRUE on success
1060 g_markup_parse_context_parse (GMarkupParseContext *context,
1065 g_return_val_if_fail (context != NULL, FALSE);
1066 g_return_val_if_fail (text != NULL, FALSE);
1067 g_return_val_if_fail (context->state != STATE_ERROR, FALSE);
1068 g_return_val_if_fail (!context->parsing, FALSE);
1071 text_len = strlen (text);
1076 context->parsing = TRUE;
1079 context->current_text = text;
1080 context->current_text_len = text_len;
1081 context->current_text_end = context->current_text + text_len;
1082 context->iter = context->current_text;
1083 context->start = context->iter;
1085 while (context->iter != context->current_text_end)
1087 switch (context->state)
1090 /* Possible next state: AFTER_OPEN_ANGLE */
1092 g_assert (context->tag_stack == NULL);
1094 /* whitespace is ignored outside of any elements */
1095 skip_spaces (context);
1097 if (context->iter != context->current_text_end)
1099 if (*context->iter == '<')
1101 /* Move after the open angle */
1102 advance_char (context);
1104 context->state = STATE_AFTER_OPEN_ANGLE;
1106 /* this could start a passthrough */
1107 context->start = context->iter;
1109 /* document is now non-empty */
1110 context->document_empty = FALSE;
1114 set_error_literal (context,
1116 G_MARKUP_ERROR_PARSE,
1117 _("Document must begin with an element (e.g. <book>)"));
1122 case STATE_AFTER_OPEN_ANGLE:
1123 /* Possible next states: INSIDE_OPEN_TAG_NAME,
1124 * AFTER_CLOSE_TAG_SLASH, INSIDE_PASSTHROUGH
1126 if (*context->iter == '?' ||
1127 *context->iter == '!')
1129 /* include < in the passthrough */
1130 const gchar *openangle = "<";
1131 add_to_partial (context, openangle, openangle + 1);
1132 context->start = context->iter;
1133 context->balance = 1;
1134 context->state = STATE_INSIDE_PASSTHROUGH;
1136 else if (*context->iter == '/')
1139 advance_char (context);
1141 context->state = STATE_AFTER_CLOSE_TAG_SLASH;
1143 else if (!IS_COMMON_NAME_END_CHAR (*(context->iter)))
1145 context->state = STATE_INSIDE_OPEN_TAG_NAME;
1147 /* start of tag name */
1148 context->start = context->iter;
1156 G_MARKUP_ERROR_PARSE,
1157 _("'%s' is not a valid character following "
1158 "a '<' character; it may not begin an "
1160 utf8_str (context->iter, buf));
1164 /* The AFTER_CLOSE_ANGLE state is actually sort of
1165 * broken, because it doesn't correspond to a range
1166 * of characters in the input stream as the others do,
1167 * and thus makes things harder to conceptualize
1169 case STATE_AFTER_CLOSE_ANGLE:
1170 /* Possible next states: INSIDE_TEXT, STATE_START */
1171 if (context->tag_stack == NULL)
1173 context->start = NULL;
1174 context->state = STATE_START;
1178 context->start = context->iter;
1179 context->state = STATE_INSIDE_TEXT;
1183 case STATE_AFTER_ELISION_SLASH:
1184 /* Possible next state: AFTER_CLOSE_ANGLE */
1187 /* We need to pop the tag stack and call the end_element
1188 * function, since this is the close tag
1190 GError *tmp_error = NULL;
1192 g_assert (context->tag_stack != NULL);
1194 possibly_finish_subparser (context);
1197 if (context->parser->end_element)
1198 (* context->parser->end_element) (context,
1199 current_element (context),
1203 ensure_no_outstanding_subparser (context);
1207 mark_error (context, tmp_error);
1208 g_propagate_error (error, tmp_error);
1212 if (*context->iter == '>')
1214 /* move after the close angle */
1215 advance_char (context);
1216 context->state = STATE_AFTER_CLOSE_ANGLE;
1224 G_MARKUP_ERROR_PARSE,
1225 _("Odd character '%s', expected a '>' character "
1226 "to end the empty-element tag '%s'"),
1227 utf8_str (context->iter, buf),
1228 current_element (context));
1235 case STATE_INSIDE_OPEN_TAG_NAME:
1236 /* Possible next states: BETWEEN_ATTRIBUTES */
1238 /* if there's a partial chunk then it's the first part of the
1239 * tag name. If there's a context->start then it's the start
1240 * of the tag name in current_text, the partial chunk goes
1241 * before that start though.
1243 advance_to_name_end (context);
1245 if (context->iter == context->current_text_end)
1247 /* The name hasn't necessarily ended. Merge with
1248 * partial chunk, leave state unchanged.
1250 add_to_partial (context, context->start, context->iter);
1254 /* The name has ended. Combine it with the partial chunk
1255 * if any; push it on the stack; enter next state.
1257 add_to_partial (context, context->start, context->iter);
1258 push_partial_as_tag (context);
1260 context->state = STATE_BETWEEN_ATTRIBUTES;
1261 context->start = NULL;
1265 case STATE_INSIDE_ATTRIBUTE_NAME:
1266 /* Possible next states: AFTER_ATTRIBUTE_NAME */
1268 advance_to_name_end (context);
1269 add_to_partial (context, context->start, context->iter);
1271 /* read the full name, if we enter the equals sign state
1272 * then add the attribute to the list (without the value),
1273 * otherwise store a partial chunk to be prepended later.
1275 if (context->iter != context->current_text_end)
1276 context->state = STATE_AFTER_ATTRIBUTE_NAME;
1279 case STATE_AFTER_ATTRIBUTE_NAME:
1280 /* Possible next states: AFTER_ATTRIBUTE_EQUALS_SIGN */
1282 skip_spaces (context);
1284 if (context->iter != context->current_text_end)
1286 /* The name has ended. Combine it with the partial chunk
1287 * if any; push it on the stack; enter next state.
1289 if (!name_validate (context, context->partial_chunk->str, error))
1292 add_attribute (context, context->partial_chunk);
1294 context->partial_chunk = NULL;
1295 context->start = NULL;
1297 if (*context->iter == '=')
1299 advance_char (context);
1300 context->state = STATE_AFTER_ATTRIBUTE_EQUALS_SIGN;
1308 G_MARKUP_ERROR_PARSE,
1309 _("Odd character '%s', expected a '=' after "
1310 "attribute name '%s' of element '%s'"),
1311 utf8_str (context->iter, buf),
1312 current_attribute (context),
1313 current_element (context));
1319 case STATE_BETWEEN_ATTRIBUTES:
1320 /* Possible next states: AFTER_CLOSE_ANGLE,
1321 * AFTER_ELISION_SLASH, INSIDE_ATTRIBUTE_NAME
1323 skip_spaces (context);
1325 if (context->iter != context->current_text_end)
1327 if (*context->iter == '/')
1329 advance_char (context);
1330 context->state = STATE_AFTER_ELISION_SLASH;
1332 else if (*context->iter == '>')
1334 advance_char (context);
1335 context->state = STATE_AFTER_CLOSE_ANGLE;
1337 else if (!IS_COMMON_NAME_END_CHAR (*(context->iter)))
1339 context->state = STATE_INSIDE_ATTRIBUTE_NAME;
1340 /* start of attribute name */
1341 context->start = context->iter;
1349 G_MARKUP_ERROR_PARSE,
1350 _("Odd character '%s', expected a '>' or '/' "
1351 "character to end the start tag of "
1352 "element '%s', or optionally an attribute; "
1353 "perhaps you used an invalid character in "
1354 "an attribute name"),
1355 utf8_str (context->iter, buf),
1356 current_element (context));
1359 /* If we're done with attributes, invoke
1360 * the start_element callback
1362 if (context->state == STATE_AFTER_ELISION_SLASH ||
1363 context->state == STATE_AFTER_CLOSE_ANGLE)
1364 emit_start_element (context, error);
1368 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN:
1369 /* Possible next state: INSIDE_ATTRIBUTE_VALUE_[SQ/DQ] */
1371 skip_spaces (context);
1373 if (context->iter != context->current_text_end)
1375 if (*context->iter == '"')
1377 advance_char (context);
1378 context->state = STATE_INSIDE_ATTRIBUTE_VALUE_DQ;
1379 context->start = context->iter;
1381 else if (*context->iter == '\'')
1383 advance_char (context);
1384 context->state = STATE_INSIDE_ATTRIBUTE_VALUE_SQ;
1385 context->start = context->iter;
1393 G_MARKUP_ERROR_PARSE,
1394 _("Odd character '%s', expected an open quote mark "
1395 "after the equals sign when giving value for "
1396 "attribute '%s' of element '%s'"),
1397 utf8_str (context->iter, buf),
1398 current_attribute (context),
1399 current_element (context));
1404 case STATE_INSIDE_ATTRIBUTE_VALUE_SQ:
1405 case STATE_INSIDE_ATTRIBUTE_VALUE_DQ:
1406 /* Possible next states: BETWEEN_ATTRIBUTES */
1410 if (context->state == STATE_INSIDE_ATTRIBUTE_VALUE_SQ)
1421 if (*context->iter == delim)
1424 while (advance_char (context));
1426 if (context->iter == context->current_text_end)
1428 /* The value hasn't necessarily ended. Merge with
1429 * partial chunk, leave state unchanged.
1431 add_to_partial (context, context->start, context->iter);
1436 /* The value has ended at the quote mark. Combine it
1437 * with the partial chunk if any; set it for the current
1440 add_to_partial (context, context->start, context->iter);
1442 g_assert (context->cur_attr >= 0);
1444 if (unescape_gstring_inplace (context, context->partial_chunk, &is_ascii, error) &&
1445 (is_ascii || text_validate (context, context->partial_chunk->str,
1446 context->partial_chunk->len, error)))
1448 /* success, advance past quote and set state. */
1449 context->attr_values[context->cur_attr] = context->partial_chunk;
1450 context->partial_chunk = NULL;
1451 advance_char (context);
1452 context->state = STATE_BETWEEN_ATTRIBUTES;
1453 context->start = NULL;
1456 truncate_partial (context);
1460 case STATE_INSIDE_TEXT:
1461 /* Possible next states: AFTER_OPEN_ANGLE */
1464 if (*context->iter == '<')
1467 while (advance_char (context));
1469 /* The text hasn't necessarily ended. Merge with
1470 * partial chunk, leave state unchanged.
1473 add_to_partial (context, context->start, context->iter);
1475 if (context->iter != context->current_text_end)
1479 /* The text has ended at the open angle. Call the text
1482 if (unescape_gstring_inplace (context, context->partial_chunk, &is_ascii, error) &&
1483 (is_ascii || text_validate (context, context->partial_chunk->str,
1484 context->partial_chunk->len, error)))
1486 GError *tmp_error = NULL;
1488 if (context->parser->text)
1489 (*context->parser->text) (context,
1490 context->partial_chunk->str,
1491 context->partial_chunk->len,
1495 if (tmp_error == NULL)
1497 /* advance past open angle and set state. */
1498 advance_char (context);
1499 context->state = STATE_AFTER_OPEN_ANGLE;
1500 /* could begin a passthrough */
1501 context->start = context->iter;
1504 propagate_error (context, error, tmp_error);
1507 truncate_partial (context);
1511 case STATE_AFTER_CLOSE_TAG_SLASH:
1512 /* Possible next state: INSIDE_CLOSE_TAG_NAME */
1513 if (!IS_COMMON_NAME_END_CHAR (*(context->iter)))
1515 context->state = STATE_INSIDE_CLOSE_TAG_NAME;
1517 /* start of tag name */
1518 context->start = context->iter;
1526 G_MARKUP_ERROR_PARSE,
1527 _("'%s' is not a valid character following "
1528 "the characters '</'; '%s' may not begin an "
1530 utf8_str (context->iter, buf),
1531 utf8_str (context->iter, buf));
1535 case STATE_INSIDE_CLOSE_TAG_NAME:
1536 /* Possible next state: AFTER_CLOSE_TAG_NAME */
1537 advance_to_name_end (context);
1538 add_to_partial (context, context->start, context->iter);
1540 if (context->iter != context->current_text_end)
1541 context->state = STATE_AFTER_CLOSE_TAG_NAME;
1544 case STATE_AFTER_CLOSE_TAG_NAME:
1545 /* Possible next state: AFTER_CLOSE_TAG_SLASH */
1547 skip_spaces (context);
1549 if (context->iter != context->current_text_end)
1551 GString *close_name;
1553 close_name = context->partial_chunk;
1554 context->partial_chunk = NULL;
1556 if (*context->iter != '>')
1562 G_MARKUP_ERROR_PARSE,
1563 _("'%s' is not a valid character following "
1564 "the close element name '%s'; the allowed "
1565 "character is '>'"),
1566 utf8_str (context->iter, buf),
1569 else if (context->tag_stack == NULL)
1573 G_MARKUP_ERROR_PARSE,
1574 _("Element '%s' was closed, no element "
1575 "is currently open"),
1578 else if (strcmp (close_name->str, current_element (context)) != 0)
1582 G_MARKUP_ERROR_PARSE,
1583 _("Element '%s' was closed, but the currently "
1584 "open element is '%s'"),
1586 current_element (context));
1591 advance_char (context);
1592 context->state = STATE_AFTER_CLOSE_ANGLE;
1593 context->start = NULL;
1595 possibly_finish_subparser (context);
1597 /* call the end_element callback */
1599 if (context->parser->end_element)
1600 (* context->parser->end_element) (context,
1605 ensure_no_outstanding_subparser (context);
1609 propagate_error (context, error, tmp_error);
1611 context->partial_chunk = close_name;
1612 truncate_partial (context);
1616 case STATE_INSIDE_PASSTHROUGH:
1617 /* Possible next state: AFTER_CLOSE_ANGLE */
1620 if (*context->iter == '<')
1622 if (*context->iter == '>')
1628 add_to_partial (context, context->start, context->iter);
1629 context->start = context->iter;
1631 str = context->partial_chunk->str;
1632 len = context->partial_chunk->len;
1634 if (str[1] == '?' && str[len - 1] == '?')
1636 if (strncmp (str, "<!--", 4) == 0 &&
1637 strcmp (str + len - 2, "--") == 0)
1639 if (strncmp (str, "<![CDATA[", 9) == 0 &&
1640 strcmp (str + len - 2, "]]") == 0)
1642 if (strncmp (str, "<!DOCTYPE", 9) == 0 &&
1643 context->balance == 0)
1647 while (advance_char (context));
1649 if (context->iter == context->current_text_end)
1651 /* The passthrough hasn't necessarily ended. Merge with
1652 * partial chunk, leave state unchanged.
1654 add_to_partial (context, context->start, context->iter);
1658 /* The passthrough has ended at the close angle. Combine
1659 * it with the partial chunk if any. Call the passthrough
1660 * callback. Note that the open/close angles are
1661 * included in the text of the passthrough.
1663 GError *tmp_error = NULL;
1665 advance_char (context); /* advance past close angle */
1666 add_to_partial (context, context->start, context->iter);
1668 if (context->flags & G_MARKUP_TREAT_CDATA_AS_TEXT &&
1669 strncmp (context->partial_chunk->str, "<![CDATA[", 9) == 0)
1671 if (context->parser->text &&
1672 text_validate (context,
1673 context->partial_chunk->str + 9,
1674 context->partial_chunk->len - 12,
1676 (*context->parser->text) (context,
1677 context->partial_chunk->str + 9,
1678 context->partial_chunk->len - 12,
1682 else if (context->parser->passthrough &&
1683 text_validate (context,
1684 context->partial_chunk->str,
1685 context->partial_chunk->len,
1687 (*context->parser->passthrough) (context,
1688 context->partial_chunk->str,
1689 context->partial_chunk->len,
1693 truncate_partial (context);
1695 if (tmp_error == NULL)
1697 context->state = STATE_AFTER_CLOSE_ANGLE;
1698 context->start = context->iter; /* could begin text */
1701 propagate_error (context, error, tmp_error);
1710 g_assert_not_reached ();
1716 context->parsing = FALSE;
1718 return context->state != STATE_ERROR;
1722 * g_markup_parse_context_end_parse:
1723 * @context: a #GMarkupParseContext
1724 * @error: return location for a #GError
1726 * Signals to the #GMarkupParseContext that all data has been
1727 * fed into the parse context with g_markup_parse_context_parse().
1729 * This function reports an error if the document isn't complete,
1730 * for example if elements are still open.
1732 * Return value: %TRUE on success, %FALSE if an error was set
1735 g_markup_parse_context_end_parse (GMarkupParseContext *context,
1738 g_return_val_if_fail (context != NULL, FALSE);
1739 g_return_val_if_fail (!context->parsing, FALSE);
1740 g_return_val_if_fail (context->state != STATE_ERROR, FALSE);
1742 if (context->partial_chunk != NULL)
1744 g_string_free (context->partial_chunk, TRUE);
1745 context->partial_chunk = NULL;
1748 if (context->document_empty)
1750 set_error_literal (context, error, G_MARKUP_ERROR_EMPTY,
1751 _("Document was empty or contained only whitespace"));
1755 context->parsing = TRUE;
1757 switch (context->state)
1763 case STATE_AFTER_OPEN_ANGLE:
1764 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1765 _("Document ended unexpectedly just after an open angle bracket '<'"));
1768 case STATE_AFTER_CLOSE_ANGLE:
1769 if (context->tag_stack != NULL)
1771 /* Error message the same as for INSIDE_TEXT */
1772 set_error (context, error, G_MARKUP_ERROR_PARSE,
1773 _("Document ended unexpectedly with elements still open - "
1774 "'%s' was the last element opened"),
1775 current_element (context));
1779 case STATE_AFTER_ELISION_SLASH:
1780 set_error (context, error, G_MARKUP_ERROR_PARSE,
1781 _("Document ended unexpectedly, expected to see a close angle "
1782 "bracket ending the tag <%s/>"), current_element (context));
1785 case STATE_INSIDE_OPEN_TAG_NAME:
1786 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1787 _("Document ended unexpectedly inside an element name"));
1790 case STATE_INSIDE_ATTRIBUTE_NAME:
1791 case STATE_AFTER_ATTRIBUTE_NAME:
1792 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1793 _("Document ended unexpectedly inside an attribute name"));
1796 case STATE_BETWEEN_ATTRIBUTES:
1797 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1798 _("Document ended unexpectedly inside an element-opening "
1802 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN:
1803 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1804 _("Document ended unexpectedly after the equals sign "
1805 "following an attribute name; no attribute value"));
1808 case STATE_INSIDE_ATTRIBUTE_VALUE_SQ:
1809 case STATE_INSIDE_ATTRIBUTE_VALUE_DQ:
1810 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1811 _("Document ended unexpectedly while inside an attribute "
1815 case STATE_INSIDE_TEXT:
1816 g_assert (context->tag_stack != NULL);
1817 set_error (context, error, G_MARKUP_ERROR_PARSE,
1818 _("Document ended unexpectedly with elements still open - "
1819 "'%s' was the last element opened"),
1820 current_element (context));
1823 case STATE_AFTER_CLOSE_TAG_SLASH:
1824 case STATE_INSIDE_CLOSE_TAG_NAME:
1825 case STATE_AFTER_CLOSE_TAG_NAME:
1826 set_error (context, error, G_MARKUP_ERROR_PARSE,
1827 _("Document ended unexpectedly inside the close tag for "
1828 "element '%s'"), current_element (context));
1831 case STATE_INSIDE_PASSTHROUGH:
1832 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1833 _("Document ended unexpectedly inside a comment or "
1834 "processing instruction"));
1839 g_assert_not_reached ();
1843 context->parsing = FALSE;
1845 return context->state != STATE_ERROR;
1849 * g_markup_parse_context_get_element:
1850 * @context: a #GMarkupParseContext
1852 * Retrieves the name of the currently open element.
1854 * If called from the start_element or end_element handlers this will
1855 * give the element_name as passed to those functions. For the parent
1856 * elements, see g_markup_parse_context_get_element_stack().
1858 * Returns: the name of the currently open element, or %NULL
1863 g_markup_parse_context_get_element (GMarkupParseContext *context)
1865 g_return_val_if_fail (context != NULL, NULL);
1867 if (context->tag_stack == NULL)
1870 return current_element (context);
1874 * g_markup_parse_context_get_element_stack:
1875 * @context: a #GMarkupParseContext
1877 * Retrieves the element stack from the internal state of the parser.
1879 * The returned #GSList is a list of strings where the first item is
1880 * the currently open tag (as would be returned by
1881 * g_markup_parse_context_get_element()) and the next item is its
1884 * This function is intended to be used in the start_element and
1885 * end_element handlers where g_markup_parse_context_get_element()
1886 * would merely return the name of the element that is being
1889 * Returns: the element stack, which must not be modified
1894 g_markup_parse_context_get_element_stack (GMarkupParseContext *context)
1896 g_return_val_if_fail (context != NULL, NULL);
1897 return context->tag_stack;
1901 * g_markup_parse_context_get_position:
1902 * @context: a #GMarkupParseContext
1903 * @line_number: (allow-none): return location for a line number, or %NULL
1904 * @char_number: (allow-none): return location for a char-on-line number, or %NULL
1906 * Retrieves the current line number and the number of the character on
1907 * that line. Intended for use in error messages; there are no strict
1908 * semantics for what constitutes the "current" line number other than
1909 * "the best number we could come up with for error messages."
1912 g_markup_parse_context_get_position (GMarkupParseContext *context,
1916 g_return_if_fail (context != NULL);
1919 *line_number = context->line_number;
1922 *char_number = context->char_number;
1926 * g_markup_parse_context_get_user_data:
1927 * @context: a #GMarkupParseContext
1929 * Returns the user_data associated with @context.
1931 * This will either be the user_data that was provided to
1932 * g_markup_parse_context_new() or to the most recent call
1933 * of g_markup_parse_context_push().
1935 * Returns: the provided user_data. The returned data belongs to
1936 * the markup context and will be freed when
1937 * g_markup_parse_context_free() is called.
1942 g_markup_parse_context_get_user_data (GMarkupParseContext *context)
1944 return context->user_data;
1948 * g_markup_parse_context_push:
1949 * @context: a #GMarkupParseContext
1950 * @parser: a #GMarkupParser
1951 * @user_data: user data to pass to #GMarkupParser functions
1953 * Temporarily redirects markup data to a sub-parser.
1955 * This function may only be called from the start_element handler of
1956 * a #GMarkupParser. It must be matched with a corresponding call to
1957 * g_markup_parse_context_pop() in the matching end_element handler
1958 * (except in the case that the parser aborts due to an error).
1960 * All tags, text and other data between the matching tags is
1961 * redirected to the subparser given by @parser. @user_data is used
1962 * as the user_data for that parser. @user_data is also passed to the
1963 * error callback in the event that an error occurs. This includes
1964 * errors that occur in subparsers of the subparser.
1966 * The end tag matching the start tag for which this call was made is
1967 * handled by the previous parser (which is given its own user_data)
1968 * which is why g_markup_parse_context_pop() is provided to allow "one
1969 * last access" to the @user_data provided to this function. In the
1970 * case of error, the @user_data provided here is passed directly to
1971 * the error callback of the subparser and g_markup_parse_context_pop()
1972 * should not be called. In either case, if @user_data was allocated
1973 * then it ought to be freed from both of these locations.
1975 * This function is not intended to be directly called by users
1976 * interested in invoking subparsers. Instead, it is intended to be
1977 * used by the subparsers themselves to implement a higher-level
1980 * As an example, see the following implementation of a simple
1981 * parser that counts the number of tags encountered.
1990 * counter_start_element (GMarkupParseContext *context,
1991 * const gchar *element_name,
1992 * const gchar **attribute_names,
1993 * const gchar **attribute_values,
1994 * gpointer user_data,
1997 * CounterData *data = user_data;
1999 * data->tag_count++;
2003 * counter_error (GMarkupParseContext *context,
2005 * gpointer user_data)
2007 * CounterData *data = user_data;
2009 * g_slice_free (CounterData, data);
2012 * static GMarkupParser counter_subparser =
2014 * counter_start_element,
2022 * In order to allow this parser to be easily used as a subparser, the
2023 * following interface is provided:
2027 * start_counting (GMarkupParseContext *context)
2029 * CounterData *data = g_slice_new (CounterData);
2031 * data->tag_count = 0;
2032 * g_markup_parse_context_push (context, &counter_subparser, data);
2036 * end_counting (GMarkupParseContext *context)
2038 * CounterData *data = g_markup_parse_context_pop (context);
2041 * result = data->tag_count;
2042 * g_slice_free (CounterData, data);
2048 * The subparser would then be used as follows:
2051 * static void start_element (context, element_name, ...)
2053 * if (strcmp (element_name, "count-these") == 0)
2054 * start_counting (context);
2056 * /* else, handle other tags... */
2059 * static void end_element (context, element_name, ...)
2061 * if (strcmp (element_name, "count-these") == 0)
2062 * g_print ("Counted %d tags\n", end_counting (context));
2064 * /* else, handle other tags... */
2071 g_markup_parse_context_push (GMarkupParseContext *context,
2072 const GMarkupParser *parser,
2075 GMarkupRecursionTracker *tracker;
2077 tracker = g_slice_new (GMarkupRecursionTracker);
2078 tracker->prev_element = context->subparser_element;
2079 tracker->prev_parser = context->parser;
2080 tracker->prev_user_data = context->user_data;
2082 context->subparser_element = current_element (context);
2083 context->parser = parser;
2084 context->user_data = user_data;
2086 context->subparser_stack = g_slist_prepend (context->subparser_stack,
2091 * g_markup_parse_context_pop:
2092 * @context: a #GMarkupParseContext
2094 * Completes the process of a temporary sub-parser redirection.
2096 * This function exists to collect the user_data allocated by a
2097 * matching call to g_markup_parse_context_push(). It must be called
2098 * in the end_element handler corresponding to the start_element
2099 * handler during which g_markup_parse_context_push() was called.
2100 * You must not call this function from the error callback -- the
2101 * @user_data is provided directly to the callback in that case.
2103 * This function is not intended to be directly called by users
2104 * interested in invoking subparsers. Instead, it is intended to
2105 * be used by the subparsers themselves to implement a higher-level
2108 * Returns: the user data passed to g_markup_parse_context_push()
2113 g_markup_parse_context_pop (GMarkupParseContext *context)
2117 if (!context->awaiting_pop)
2118 possibly_finish_subparser (context);
2120 g_assert (context->awaiting_pop);
2122 context->awaiting_pop = FALSE;
2124 /* valgrind friendliness */
2125 user_data = context->held_user_data;
2126 context->held_user_data = NULL;
2132 append_escaped_text (GString *str,
2141 end = text + length;
2146 next = g_utf8_next_char (p);
2151 g_string_append (str, "&");
2155 g_string_append (str, "<");
2159 g_string_append (str, ">");
2163 g_string_append (str, "'");
2167 g_string_append (str, """);
2171 c = g_utf8_get_char (p);
2172 if ((0x1 <= c && c <= 0x8) ||
2173 (0xb <= c && c <= 0xc) ||
2174 (0xe <= c && c <= 0x1f) ||
2175 (0x7f <= c && c <= 0x84) ||
2176 (0x86 <= c && c <= 0x9f))
2177 g_string_append_printf (str, "&#x%x;", c);
2179 g_string_append_len (str, p, next - p);
2188 * g_markup_escape_text:
2189 * @text: some valid UTF-8 text
2190 * @length: length of @text in bytes, or -1 if the text is nul-terminated
2192 * Escapes text so that the markup parser will parse it verbatim.
2193 * Less than, greater than, ampersand, etc. are replaced with the
2194 * corresponding entities. This function would typically be used
2195 * when writing out a file to be parsed with the markup parser.
2197 * Note that this function doesn't protect whitespace and line endings
2198 * from being processed according to the XML rules for normalization
2199 * of line endings and attribute values.
2201 * Note also that this function will produce character references in
2202 * the range of &#x1; ... &#x1f; for all control sequences
2203 * except for tabstop, newline and carriage return. The character
2204 * references in this range are not valid XML 1.0, but they are
2205 * valid XML 1.1 and will be accepted by the GMarkup parser.
2207 * Return value: a newly allocated string with the escaped text
2210 g_markup_escape_text (const gchar *text,
2215 g_return_val_if_fail (text != NULL, NULL);
2218 length = strlen (text);
2220 /* prealloc at least as long as original text */
2221 str = g_string_sized_new (length);
2222 append_escaped_text (str, text, length);
2224 return g_string_free (str, FALSE);
2229 * @format: a printf-style format string
2230 * @after: location to store a pointer to the character after
2231 * the returned conversion. On a %NULL return, returns the
2232 * pointer to the trailing NUL in the string
2234 * Find the next conversion in a printf-style format string.
2235 * Partially based on code from printf-parser.c,
2236 * Copyright (C) 1999-2000, 2002-2003 Free Software Foundation, Inc.
2238 * Return value: pointer to the next conversion in @format,
2239 * or %NULL, if none.
2242 find_conversion (const char *format,
2245 const char *start = format;
2248 while (*start != '\0' && *start != '%')
2265 /* Test for positional argument. */
2266 if (*cp >= '0' && *cp <= '9')
2270 for (np = cp; *np >= '0' && *np <= '9'; np++)
2276 /* Skip the flags. */
2290 /* Skip the field width. */
2295 /* Test for positional argument. */
2296 if (*cp >= '0' && *cp <= '9')
2300 for (np = cp; *np >= '0' && *np <= '9'; np++)
2308 for (; *cp >= '0' && *cp <= '9'; cp++)
2312 /* Skip the precision. */
2318 /* Test for positional argument. */
2319 if (*cp >= '0' && *cp <= '9')
2323 for (np = cp; *np >= '0' && *np <= '9'; np++)
2331 for (; *cp >= '0' && *cp <= '9'; cp++)
2336 /* Skip argument type/size specifiers. */
2337 while (*cp == 'h' ||
2346 /* Skip the conversion character. */
2354 * g_markup_vprintf_escaped:
2355 * @format: printf() style format string
2356 * @args: variable argument list, similar to vprintf()
2358 * Formats the data in @args according to @format, escaping
2359 * all string and character arguments in the fashion
2360 * of g_markup_escape_text(). See g_markup_printf_escaped().
2362 * Return value: newly allocated result from formatting
2363 * operation. Free with g_free().
2368 g_markup_vprintf_escaped (const gchar *format,
2373 GString *result = NULL;
2374 gchar *output1 = NULL;
2375 gchar *output2 = NULL;
2376 const char *p, *op1, *op2;
2379 /* The technique here, is that we make two format strings that
2380 * have the identical conversions in the identical order to the
2381 * original strings, but differ in the text in-between. We
2382 * then use the normal g_strdup_vprintf() to format the arguments
2383 * with the two new format strings. By comparing the results,
2384 * we can figure out what segments of the output come from
2385 * the original format string, and what from the arguments,
2386 * and thus know what portions of the string to escape.
2388 * For instance, for:
2390 * g_markup_printf_escaped ("%s ate %d apples", "Susan & Fred", 5);
2392 * We form the two format strings "%sX%dX" and %sY%sY". The results
2393 * of formatting with those two strings are
2395 * "%sX%dX" => "Susan & FredX5X"
2396 * "%sY%dY" => "Susan & FredY5Y"
2398 * To find the span of the first argument, we find the first position
2399 * where the two arguments differ, which tells us that the first
2400 * argument formatted to "Susan & Fred". We then escape that
2401 * to "Susan & Fred" and join up with the intermediate portions
2402 * of the format string and the second argument to get
2403 * "Susan & Fred ate 5 apples".
2406 /* Create the two modified format strings
2408 format1 = g_string_new (NULL);
2409 format2 = g_string_new (NULL);
2414 const char *conv = find_conversion (p, &after);
2418 g_string_append_len (format1, conv, after - conv);
2419 g_string_append_c (format1, 'X');
2420 g_string_append_len (format2, conv, after - conv);
2421 g_string_append_c (format2, 'Y');
2426 /* Use them to format the arguments
2428 G_VA_COPY (args2, args);
2430 output1 = g_strdup_vprintf (format1->str, args);
2437 output2 = g_strdup_vprintf (format2->str, args2);
2442 result = g_string_new (NULL);
2444 /* Iterate through the original format string again,
2445 * copying the non-conversion portions and the escaped
2446 * converted arguments to the output string.
2454 const char *output_start;
2455 const char *conv = find_conversion (p, &after);
2458 if (!conv) /* The end, after points to the trailing \0 */
2460 g_string_append_len (result, p, after - p);
2464 g_string_append_len (result, p, conv - p);
2466 while (*op1 == *op2)
2472 escaped = g_markup_escape_text (output_start, op1 - output_start);
2473 g_string_append (result, escaped);
2482 g_string_free (format1, TRUE);
2483 g_string_free (format2, TRUE);
2488 return g_string_free (result, FALSE);
2494 * g_markup_printf_escaped:
2495 * @format: printf() style format string
2496 * @...: the arguments to insert in the format string
2498 * Formats arguments according to @format, escaping
2499 * all string and character arguments in the fashion
2500 * of g_markup_escape_text(). This is useful when you
2501 * want to insert literal strings into XML-style markup
2502 * output, without having to worry that the strings
2503 * might themselves contain markup.
2506 * const char *store = "Fortnum & Mason";
2507 * const char *item = "Tea";
2510 * output = g_markup_printf_escaped ("<purchase>"
2511 * "<store>%s</store>"
2512 * "<item>%s</item>"
2513 * "</purchase>",
2517 * Return value: newly allocated result from formatting
2518 * operation. Free with g_free().
2523 g_markup_printf_escaped (const gchar *format, ...)
2528 va_start (args, format);
2529 result = g_markup_vprintf_escaped (format, args);
2536 g_markup_parse_boolean (const char *string,
2539 char const * const falses[] = { "false", "f", "no", "n", "0" };
2540 char const * const trues[] = { "true", "t", "yes", "y", "1" };
2543 for (i = 0; i < G_N_ELEMENTS (falses); i++)
2545 if (g_ascii_strcasecmp (string, falses[i]) == 0)
2554 for (i = 0; i < G_N_ELEMENTS (trues); i++)
2556 if (g_ascii_strcasecmp (string, trues[i]) == 0)
2569 * GMarkupCollectType:
2570 * @G_MARKUP_COLLECT_INVALID: used to terminate the list of attributes
2572 * @G_MARKUP_COLLECT_STRING: collect the string pointer directly from
2573 * the attribute_values[] array. Expects a parameter of type (const
2574 * char **). If %G_MARKUP_COLLECT_OPTIONAL is specified and the
2575 * attribute isn't present then the pointer will be set to %NULL
2576 * @G_MARKUP_COLLECT_STRDUP: as with %G_MARKUP_COLLECT_STRING, but
2577 * expects a parameter of type (char **) and g_strdup()s the
2578 * returned pointer. The pointer must be freed with g_free()
2579 * @G_MARKUP_COLLECT_BOOLEAN: expects a parameter of type (gboolean *)
2580 * and parses the attribute value as a boolean. Sets %FALSE if the
2581 * attribute isn't present. Valid boolean values consist of
2582 * (case-insensitive) "false", "f", "no", "n", "0" and "true", "t",
2584 * @G_MARKUP_COLLECT_TRISTATE: as with %G_MARKUP_COLLECT_BOOLEAN, but
2585 * in the case of a missing attribute a value is set that compares
2586 * equal to neither %FALSE nor %TRUE G_MARKUP_COLLECT_OPTIONAL is
2588 * @G_MARKUP_COLLECT_OPTIONAL: can be bitwise ORed with the other fields.
2589 * If present, allows the attribute not to appear. A default value
2590 * is set depending on what value type is used
2592 * A mixed enumerated type and flags field. You must specify one type
2593 * (string, strdup, boolean, tristate). Additionally, you may optionally
2594 * bitwise OR the type with the flag %G_MARKUP_COLLECT_OPTIONAL.
2596 * It is likely that this enum will be extended in the future to
2597 * support other types.
2601 * g_markup_collect_attributes:
2602 * @element_name: the current tag name
2603 * @attribute_names: the attribute names
2604 * @attribute_values: the attribute values
2605 * @error: a pointer to a #GError or %NULL
2606 * @first_type: the #GMarkupCollectType of the first attribute
2607 * @first_attr: the name of the first attribute
2608 * @...: a pointer to the storage location of the first attribute
2609 * (or %NULL), followed by more types names and pointers, ending
2610 * with %G_MARKUP_COLLECT_INVALID
2612 * Collects the attributes of the element from the data passed to the
2613 * #GMarkupParser start_element function, dealing with common error
2614 * conditions and supporting boolean values.
2616 * This utility function is not required to write a parser but can save
2619 * The @element_name, @attribute_names, @attribute_values and @error
2620 * parameters passed to the start_element callback should be passed
2621 * unmodified to this function.
2623 * Following these arguments is a list of "supported" attributes to collect.
2624 * It is an error to specify multiple attributes with the same name. If any
2625 * attribute not in the list appears in the @attribute_names array then an
2626 * unknown attribute error will result.
2628 * The #GMarkupCollectType field allows specifying the type of collection
2629 * to perform and if a given attribute must appear or is optional.
2631 * The attribute name is simply the name of the attribute to collect.
2633 * The pointer should be of the appropriate type (see the descriptions
2634 * under #GMarkupCollectType) and may be %NULL in case a particular
2635 * attribute is to be allowed but ignored.
2637 * This function deals with issuing errors for missing attributes
2638 * (of type %G_MARKUP_ERROR_MISSING_ATTRIBUTE), unknown attributes
2639 * (of type %G_MARKUP_ERROR_UNKNOWN_ATTRIBUTE) and duplicate
2640 * attributes (of type %G_MARKUP_ERROR_INVALID_CONTENT) as well
2641 * as parse errors for boolean-valued attributes (again of type
2642 * %G_MARKUP_ERROR_INVALID_CONTENT). In all of these cases %FALSE
2643 * will be returned and @error will be set as appropriate.
2645 * Return value: %TRUE if successful
2650 g_markup_collect_attributes (const gchar *element_name,
2651 const gchar **attribute_names,
2652 const gchar **attribute_values,
2654 GMarkupCollectType first_type,
2655 const gchar *first_attr,
2658 GMarkupCollectType type;
2670 va_start (ap, first_attr);
2671 while (type != G_MARKUP_COLLECT_INVALID)
2676 mandatory = !(type & G_MARKUP_COLLECT_OPTIONAL);
2677 type &= (G_MARKUP_COLLECT_OPTIONAL - 1);
2679 /* tristate records a value != TRUE and != FALSE
2680 * for the case where the attribute is missing
2682 if (type == G_MARKUP_COLLECT_TRISTATE)
2685 for (i = 0; attribute_names[i]; i++)
2686 if (i >= 40 || !(collected & (G_GUINT64_CONSTANT(1) << i)))
2687 if (!strcmp (attribute_names[i], attr))
2690 /* ISO C99 only promises that the user can pass up to 127 arguments.
2691 * Subtracting the first 4 arguments plus the final NULL and dividing
2692 * by 3 arguments per collected attribute, we are left with a maximum
2693 * number of supported attributes of (127 - 5) / 3 = 40.
2695 * In reality, nobody is ever going to call us with anywhere close to
2696 * 40 attributes to collect, so it is safe to assume that if i > 40
2697 * then the user has given some invalid or repeated arguments. These
2698 * problems will be caught and reported at the end of the function.
2700 * We know at this point that we have an error, but we don't know
2701 * what error it is, so just continue...
2704 collected |= (G_GUINT64_CONSTANT(1) << i);
2706 value = attribute_values[i];
2708 if (value == NULL && mandatory)
2710 g_set_error (error, G_MARKUP_ERROR,
2711 G_MARKUP_ERROR_MISSING_ATTRIBUTE,
2712 "element '%s' requires attribute '%s'",
2713 element_name, attr);
2721 case G_MARKUP_COLLECT_STRING:
2723 const char **str_ptr;
2725 str_ptr = va_arg (ap, const char **);
2727 if (str_ptr != NULL)
2732 case G_MARKUP_COLLECT_STRDUP:
2736 str_ptr = va_arg (ap, char **);
2738 if (str_ptr != NULL)
2739 *str_ptr = g_strdup (value);
2743 case G_MARKUP_COLLECT_BOOLEAN:
2744 case G_MARKUP_COLLECT_TRISTATE:
2749 bool_ptr = va_arg (ap, gboolean *);
2751 if (bool_ptr != NULL)
2753 if (type == G_MARKUP_COLLECT_TRISTATE)
2754 /* constructivists rejoice!
2755 * neither false nor true...
2759 else /* G_MARKUP_COLLECT_BOOLEAN */
2765 if (!g_markup_parse_boolean (value, va_arg (ap, gboolean *)))
2767 g_set_error (error, G_MARKUP_ERROR,
2768 G_MARKUP_ERROR_INVALID_CONTENT,
2769 "element '%s', attribute '%s', value '%s' "
2770 "cannot be parsed as a boolean value",
2771 element_name, attr, value);
2781 g_assert_not_reached ();
2784 type = va_arg (ap, GMarkupCollectType);
2785 attr = va_arg (ap, const char *);
2790 /* ensure we collected all the arguments */
2791 for (i = 0; attribute_names[i]; i++)
2792 if ((collected & (G_GUINT64_CONSTANT(1) << i)) == 0)
2794 /* attribute not collected: could be caused by two things.
2796 * 1) it doesn't exist in our list of attributes
2797 * 2) it existed but was matched by a duplicate attribute earlier
2803 for (j = 0; j < i; j++)
2804 if (strcmp (attribute_names[i], attribute_names[j]) == 0)
2808 /* j is now the first occurrence of attribute_names[i] */
2810 g_set_error (error, G_MARKUP_ERROR,
2811 G_MARKUP_ERROR_UNKNOWN_ATTRIBUTE,
2812 "attribute '%s' invalid for element '%s'",
2813 attribute_names[i], element_name);
2815 g_set_error (error, G_MARKUP_ERROR,
2816 G_MARKUP_ERROR_INVALID_CONTENT,
2817 "attribute '%s' given multiple times for element '%s'",
2818 attribute_names[i], element_name);
2826 /* replay the above to free allocations */
2830 va_start (ap, first_attr);
2831 while (type != G_MARKUP_COLLECT_INVALID)
2835 ptr = va_arg (ap, gpointer);
2839 switch (type & (G_MARKUP_COLLECT_OPTIONAL - 1))
2841 case G_MARKUP_COLLECT_STRDUP:
2843 g_free (*(char **) ptr);
2845 case G_MARKUP_COLLECT_STRING:
2846 *(char **) ptr = NULL;
2849 case G_MARKUP_COLLECT_BOOLEAN:
2850 *(gboolean *) ptr = FALSE;
2853 case G_MARKUP_COLLECT_TRISTATE:
2854 *(gboolean *) ptr = -1;
2859 type = va_arg (ap, GMarkupCollectType);
2860 attr = va_arg (ap, const char *);