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);
408 set_error (GMarkupParseContext *context,
418 va_start (args, format);
419 s = g_strdup_vprintf (format, args);
422 /* Make sure that the GError message is valid UTF-8
423 * even if it is complaining about invalid UTF-8 in the markup
425 s_valid = _g_utf8_make_valid (s);
426 set_error_literal (context, error, code, s);
433 propagate_error (GMarkupParseContext *context,
437 if (context->flags & G_MARKUP_PREFIX_ERROR_POSITION)
438 g_prefix_error (&src,
439 _("Error on line %d char %d: "),
440 context->line_number,
441 context->char_number);
443 mark_error (context, src);
445 g_propagate_error (dest, src);
448 #define IS_COMMON_NAME_END_CHAR(c) \
449 ((c) == '=' || (c) == '/' || (c) == '>' || (c) == ' ')
452 slow_name_validate (GMarkupParseContext *context,
456 const gchar *p = name;
458 if (!g_utf8_validate (name, strlen (name), NULL))
460 set_error (context, error, G_MARKUP_ERROR_BAD_UTF8,
461 _("Invalid UTF-8 encoded text in name - not valid '%s'"), name);
465 if (!(g_ascii_isalpha (*p) ||
466 (!IS_COMMON_NAME_END_CHAR (*p) &&
469 g_unichar_isalpha (g_utf8_get_char (p))))))
471 set_error (context, error, G_MARKUP_ERROR_PARSE,
472 _("'%s' is not a valid name"), name);
476 for (p = g_utf8_next_char (name); *p != '\0'; p = g_utf8_next_char (p))
479 if (!(g_ascii_isalnum (*p) ||
480 (!IS_COMMON_NAME_END_CHAR (*p) &&
485 g_unichar_isalpha (g_utf8_get_char (p))))))
487 set_error (context, error, G_MARKUP_ERROR_PARSE,
488 _("'%s' is not a valid name: '%c'"), name, *p);
496 * Use me for elements, attributes etc.
499 name_validate (GMarkupParseContext *context,
506 /* name start char */
508 if (G_UNLIKELY (IS_COMMON_NAME_END_CHAR (*p) ||
509 !(g_ascii_isalpha (*p) || *p == '_' || *p == ':')))
512 for (mask = *p++; *p != '\0'; p++)
517 if (G_UNLIKELY (!(g_ascii_isalnum (*p) ||
518 (!IS_COMMON_NAME_END_CHAR (*p) &&
526 if (mask & 0x80) /* un-common / non-ascii */
532 return slow_name_validate (context, name, error);
536 text_validate (GMarkupParseContext *context,
541 if (!g_utf8_validate (p, len, NULL))
543 set_error (context, error, G_MARKUP_ERROR_BAD_UTF8,
544 _("Invalid UTF-8 encoded text in name - not valid '%s'"), p);
552 char_str (gunichar c,
556 g_unichar_to_utf8 (c, buf);
561 utf8_str (const gchar *utf8,
564 char_str (g_utf8_get_char (utf8), buf);
570 set_unescape_error (GMarkupParseContext *context,
572 const gchar *remaining_text,
580 gint remaining_newlines;
583 remaining_newlines = 0;
588 ++remaining_newlines;
592 va_start (args, format);
593 s = g_strdup_vprintf (format, args);
596 tmp_error = g_error_new (G_MARKUP_ERROR,
598 _("Error on line %d: %s"),
599 context->line_number - remaining_newlines,
604 mark_error (context, tmp_error);
606 g_propagate_error (error, tmp_error);
610 * re-write the GString in-place, unescaping anything that escaped.
611 * most XML does not contain entities, or escaping.
614 unescape_gstring_inplace (GMarkupParseContext *context,
622 gboolean normalize_attribute;
626 /* are we unescaping an attribute or not ? */
627 if (context->state == STATE_INSIDE_ATTRIBUTE_VALUE_SQ ||
628 context->state == STATE_INSIDE_ATTRIBUTE_VALUE_DQ)
629 normalize_attribute = TRUE;
631 normalize_attribute = FALSE;
634 * Meeks' theorum: unescaping can only shrink text.
635 * for < etc. this is obvious, for  more
636 * thought is required, but this is patently so.
639 for (from = to = string->str; *from != '\0'; from++, to++)
646 if (normalize_attribute && (*to == '\t' || *to == '\n'))
650 *to = normalize_attribute ? ' ' : '\n';
659 gboolean is_hex = FALSE;
671 /* digit is between start and p */
674 l = strtoul (from, &end, 16);
676 l = strtoul (from, &end, 10);
678 if (end == from || errno != 0)
680 set_unescape_error (context, error,
681 from, G_MARKUP_ERROR_PARSE,
682 _("Failed to parse '%-.*s', which "
683 "should have been a digit "
684 "inside a character reference "
685 "(ê for example) - perhaps "
686 "the digit is too large"),
687 (int)(end - from), from);
690 else if (*end != ';')
692 set_unescape_error (context, error,
693 from, G_MARKUP_ERROR_PARSE,
694 _("Character reference did not end with a "
696 "most likely you used an ampersand "
697 "character without intending to start "
698 "an entity - escape ampersand as &"));
703 /* characters XML 1.1 permits */
704 if ((0 < l && l <= 0xD7FF) ||
705 (0xE000 <= l && l <= 0xFFFD) ||
706 (0x10000 <= l && l <= 0x10FFFF))
711 to += strlen (buf) - 1;
713 if (l >= 0x80) /* not ascii */
718 set_unescape_error (context, error,
719 from, G_MARKUP_ERROR_PARSE,
720 _("Character reference '%-.*s' does not "
721 "encode a permitted character"),
722 (int)(end - from), from);
728 else if (strncmp (from, "lt;", 3) == 0)
733 else if (strncmp (from, "gt;", 3) == 0)
738 else if (strncmp (from, "amp;", 4) == 0)
743 else if (strncmp (from, "quot;", 5) == 0)
748 else if (strncmp (from, "apos;", 5) == 0)
756 set_unescape_error (context, error,
757 from, G_MARKUP_ERROR_PARSE,
758 _("Empty entity '&;' seen; valid "
759 "entities are: & " < > '"));
762 const char *end = strchr (from, ';');
764 set_unescape_error (context, error,
765 from, G_MARKUP_ERROR_PARSE,
766 _("Entity name '%-.*s' is not known"),
767 (int)(end - from), from);
769 set_unescape_error (context, error,
770 from, G_MARKUP_ERROR_PARSE,
771 _("Entity did not end with a semicolon; "
772 "most likely you used an ampersand "
773 "character without intending to start "
774 "an entity - escape ampersand as &"));
781 g_assert (to - string->str <= string->len);
782 if (to - string->str != string->len)
783 g_string_truncate (string, to - string->str);
785 *is_ascii = !(mask & 0x80);
790 static inline gboolean
791 advance_char (GMarkupParseContext *context)
794 context->char_number++;
796 if (G_UNLIKELY (context->iter == context->current_text_end))
799 else if (G_UNLIKELY (*context->iter == '\n'))
801 context->line_number++;
802 context->char_number = 1;
808 static inline gboolean
811 return c == ' ' || c == '\t' || c == '\n' || c == '\r';
815 skip_spaces (GMarkupParseContext *context)
819 if (!xml_isspace (*context->iter))
822 while (advance_char (context));
826 advance_to_name_end (GMarkupParseContext *context)
830 if (IS_COMMON_NAME_END_CHAR (*(context->iter)))
832 if (xml_isspace (*(context->iter)))
835 while (advance_char (context));
839 release_chunk (GMarkupParseContext *context, GString *str)
844 if (str->allocated_len > 256)
845 { /* large strings are unusual and worth freeing */
846 g_string_free (str, TRUE);
850 node = get_list_node (context, str);
851 context->spare_chunks = g_slist_concat (node, context->spare_chunks);
855 add_to_partial (GMarkupParseContext *context,
856 const gchar *text_start,
857 const gchar *text_end)
859 if (context->partial_chunk == NULL)
860 { /* allocate a new chunk to parse into */
862 if (context->spare_chunks != NULL)
864 GSList *node = context->spare_chunks;
865 context->spare_chunks = g_slist_remove_link (context->spare_chunks, node);
866 context->partial_chunk = node->data;
867 free_list_node (context, node);
870 context->partial_chunk = g_string_sized_new (MAX (28, text_end - text_start));
873 if (text_start != text_end)
874 g_string_insert_len (context->partial_chunk, -1,
875 text_start, text_end - text_start);
879 truncate_partial (GMarkupParseContext *context)
881 if (context->partial_chunk != NULL)
882 string_blank (context->partial_chunk);
885 static inline const gchar*
886 current_element (GMarkupParseContext *context)
888 return context->tag_stack->data;
892 pop_subparser_stack (GMarkupParseContext *context)
894 GMarkupRecursionTracker *tracker;
896 g_assert (context->subparser_stack);
898 tracker = context->subparser_stack->data;
900 context->awaiting_pop = TRUE;
901 context->held_user_data = context->user_data;
903 context->user_data = tracker->prev_user_data;
904 context->parser = tracker->prev_parser;
905 context->subparser_element = tracker->prev_element;
906 g_slice_free (GMarkupRecursionTracker, tracker);
908 context->subparser_stack = g_slist_delete_link (context->subparser_stack,
909 context->subparser_stack);
913 push_partial_as_tag (GMarkupParseContext *context)
915 GString *str = context->partial_chunk;
916 /* sadly, this is exported by gmarkup_get_element_stack as-is */
917 context->tag_stack = g_slist_concat (get_list_node (context, str->str), context->tag_stack);
918 context->tag_stack_gstr = g_slist_concat (get_list_node (context, str), context->tag_stack_gstr);
919 context->partial_chunk = NULL;
923 pop_tag (GMarkupParseContext *context)
925 GSList *nodea, *nodeb;
927 nodea = context->tag_stack;
928 nodeb = context->tag_stack_gstr;
929 release_chunk (context, nodeb->data);
930 context->tag_stack = g_slist_remove_link (context->tag_stack, nodea);
931 context->tag_stack_gstr = g_slist_remove_link (context->tag_stack_gstr, nodeb);
932 free_list_node (context, nodea);
933 free_list_node (context, nodeb);
937 possibly_finish_subparser (GMarkupParseContext *context)
939 if (current_element (context) == context->subparser_element)
940 pop_subparser_stack (context);
944 ensure_no_outstanding_subparser (GMarkupParseContext *context)
946 if (context->awaiting_pop)
947 g_critical ("During the first end_element call after invoking a "
948 "subparser you must pop the subparser stack and handle "
949 "the freeing of the subparser user_data. This can be "
950 "done by calling the end function of the subparser. "
951 "Very probably, your program just leaked memory.");
953 /* let valgrind watch the pointer disappear... */
954 context->held_user_data = NULL;
955 context->awaiting_pop = FALSE;
959 current_attribute (GMarkupParseContext *context)
961 g_assert (context->cur_attr >= 0);
962 return context->attr_names[context->cur_attr]->str;
966 add_attribute (GMarkupParseContext *context, GString *str)
968 if (context->cur_attr + 2 >= context->alloc_attrs)
970 context->alloc_attrs += 5; /* silly magic number */
971 context->attr_names = g_realloc (context->attr_names, sizeof(GString*)*context->alloc_attrs);
972 context->attr_values = g_realloc (context->attr_values, sizeof(GString*)*context->alloc_attrs);
975 context->attr_names[context->cur_attr] = str;
976 context->attr_values[context->cur_attr] = NULL;
977 context->attr_names[context->cur_attr+1] = NULL;
978 context->attr_values[context->cur_attr+1] = NULL;
982 clear_attributes (GMarkupParseContext *context)
984 /* Go ahead and free the attributes. */
985 for (; context->cur_attr >= 0; context->cur_attr--)
987 int pos = context->cur_attr;
988 release_chunk (context, context->attr_names[pos]);
989 release_chunk (context, context->attr_values[pos]);
990 context->attr_names[pos] = context->attr_values[pos] = NULL;
992 g_assert (context->cur_attr == -1);
993 g_assert (context->attr_names == NULL ||
994 context->attr_names[0] == NULL);
995 g_assert (context->attr_values == NULL ||
996 context->attr_values[0] == NULL);
999 /* This has to be a separate function to ensure the alloca's
1000 * are unwound on exit - otherwise we grow & blow the stack
1001 * with large documents
1004 emit_start_element (GMarkupParseContext *context,
1008 const gchar *start_name;
1009 const gchar **attr_names;
1010 const gchar **attr_values;
1013 /* In case we want to ignore qualified tags and we see that we have
1014 * one here, we push a subparser. This will ignore all tags inside of
1015 * the qualified tag.
1017 * We deal with the end of the subparser from emit_end_element.
1019 if (context->flags & G_MARKUP_IGNORE_QUALIFIED && strchr (current_element (context), ':'))
1021 static const GMarkupParser ignore_parser;
1022 g_markup_parse_context_push (context, &ignore_parser, NULL);
1026 attr_names = g_newa (const gchar *, context->cur_attr + 2);
1027 attr_values = g_newa (const gchar *, context->cur_attr + 2);
1028 for (i = 0; i < context->cur_attr + 1; i++)
1030 /* Possibly omit qualified attribute names from the list */
1031 if ((context->flags & G_MARKUP_IGNORE_QUALIFIED) && strchr (context->attr_names[i]->str, ':'))
1034 attr_names[j] = context->attr_names[i]->str;
1035 attr_values[j] = context->attr_values[i]->str;
1038 attr_names[j] = NULL;
1039 attr_values[j] = NULL;
1041 /* Call user callback for element start */
1043 start_name = current_element (context);
1045 if (context->parser->start_element &&
1046 name_validate (context, start_name, error))
1047 (* context->parser->start_element) (context,
1049 (const gchar **)attr_names,
1050 (const gchar **)attr_values,
1053 clear_attributes (context);
1055 if (tmp_error != NULL)
1056 propagate_error (context, error, tmp_error);
1060 emit_end_element (GMarkupParseContext *context,
1063 /* We need to pop the tag stack and call the end_element
1064 * function, since this is the close tag
1066 GError *tmp_error = NULL;
1068 g_assert (context->tag_stack != NULL);
1070 possibly_finish_subparser (context);
1072 /* We might have just returned from our ignore subparser */
1073 if (context->flags & G_MARKUP_IGNORE_QUALIFIED && strchr (current_element (context), ':'))
1075 g_markup_parse_context_pop (context);
1081 if (context->parser->end_element)
1082 (* context->parser->end_element) (context,
1083 current_element (context),
1087 ensure_no_outstanding_subparser (context);
1091 mark_error (context, tmp_error);
1092 g_propagate_error (error, tmp_error);
1099 * g_markup_parse_context_parse:
1100 * @context: a #GMarkupParseContext
1101 * @text: chunk of text to parse
1102 * @text_len: length of @text in bytes
1103 * @error: return location for a #GError
1105 * Feed some data to the #GMarkupParseContext.
1107 * The data need not be valid UTF-8; an error will be signaled if
1108 * it's invalid. The data need not be an entire document; you can
1109 * feed a document into the parser incrementally, via multiple calls
1110 * to this function. Typically, as you receive data from a network
1111 * connection or file, you feed each received chunk of data into this
1112 * function, aborting the process if an error occurs. Once an error
1113 * is reported, no further data may be fed to the #GMarkupParseContext;
1114 * all errors are fatal.
1116 * Return value: %FALSE if an error occurred, %TRUE on success
1119 g_markup_parse_context_parse (GMarkupParseContext *context,
1124 g_return_val_if_fail (context != NULL, FALSE);
1125 g_return_val_if_fail (text != NULL, FALSE);
1126 g_return_val_if_fail (context->state != STATE_ERROR, FALSE);
1127 g_return_val_if_fail (!context->parsing, FALSE);
1130 text_len = strlen (text);
1135 context->parsing = TRUE;
1138 context->current_text = text;
1139 context->current_text_len = text_len;
1140 context->current_text_end = context->current_text + text_len;
1141 context->iter = context->current_text;
1142 context->start = context->iter;
1144 while (context->iter != context->current_text_end)
1146 switch (context->state)
1149 /* Possible next state: AFTER_OPEN_ANGLE */
1151 g_assert (context->tag_stack == NULL);
1153 /* whitespace is ignored outside of any elements */
1154 skip_spaces (context);
1156 if (context->iter != context->current_text_end)
1158 if (*context->iter == '<')
1160 /* Move after the open angle */
1161 advance_char (context);
1163 context->state = STATE_AFTER_OPEN_ANGLE;
1165 /* this could start a passthrough */
1166 context->start = context->iter;
1168 /* document is now non-empty */
1169 context->document_empty = FALSE;
1173 set_error_literal (context,
1175 G_MARKUP_ERROR_PARSE,
1176 _("Document must begin with an element (e.g. <book>)"));
1181 case STATE_AFTER_OPEN_ANGLE:
1182 /* Possible next states: INSIDE_OPEN_TAG_NAME,
1183 * AFTER_CLOSE_TAG_SLASH, INSIDE_PASSTHROUGH
1185 if (*context->iter == '?' ||
1186 *context->iter == '!')
1188 /* include < in the passthrough */
1189 const gchar *openangle = "<";
1190 add_to_partial (context, openangle, openangle + 1);
1191 context->start = context->iter;
1192 context->balance = 1;
1193 context->state = STATE_INSIDE_PASSTHROUGH;
1195 else if (*context->iter == '/')
1198 advance_char (context);
1200 context->state = STATE_AFTER_CLOSE_TAG_SLASH;
1202 else if (!IS_COMMON_NAME_END_CHAR (*(context->iter)))
1204 context->state = STATE_INSIDE_OPEN_TAG_NAME;
1206 /* start of tag name */
1207 context->start = context->iter;
1215 G_MARKUP_ERROR_PARSE,
1216 _("'%s' is not a valid character following "
1217 "a '<' character; it may not begin an "
1219 utf8_str (context->iter, buf));
1223 /* The AFTER_CLOSE_ANGLE state is actually sort of
1224 * broken, because it doesn't correspond to a range
1225 * of characters in the input stream as the others do,
1226 * and thus makes things harder to conceptualize
1228 case STATE_AFTER_CLOSE_ANGLE:
1229 /* Possible next states: INSIDE_TEXT, STATE_START */
1230 if (context->tag_stack == NULL)
1232 context->start = NULL;
1233 context->state = STATE_START;
1237 context->start = context->iter;
1238 context->state = STATE_INSIDE_TEXT;
1242 case STATE_AFTER_ELISION_SLASH:
1243 /* Possible next state: AFTER_CLOSE_ANGLE */
1244 if (*context->iter == '>')
1246 /* move after the close angle */
1247 advance_char (context);
1248 context->state = STATE_AFTER_CLOSE_ANGLE;
1249 emit_end_element (context, error);
1257 G_MARKUP_ERROR_PARSE,
1258 _("Odd character '%s', expected a '>' character "
1259 "to end the empty-element tag '%s'"),
1260 utf8_str (context->iter, buf),
1261 current_element (context));
1265 case STATE_INSIDE_OPEN_TAG_NAME:
1266 /* Possible next states: BETWEEN_ATTRIBUTES */
1268 /* if there's a partial chunk then it's the first part of the
1269 * tag name. If there's a context->start then it's the start
1270 * of the tag name in current_text, the partial chunk goes
1271 * before that start though.
1273 advance_to_name_end (context);
1275 if (context->iter == context->current_text_end)
1277 /* The name hasn't necessarily ended. Merge with
1278 * partial chunk, leave state unchanged.
1280 add_to_partial (context, context->start, context->iter);
1284 /* The name has ended. Combine it with the partial chunk
1285 * if any; push it on the stack; enter next state.
1287 add_to_partial (context, context->start, context->iter);
1288 push_partial_as_tag (context);
1290 context->state = STATE_BETWEEN_ATTRIBUTES;
1291 context->start = NULL;
1295 case STATE_INSIDE_ATTRIBUTE_NAME:
1296 /* Possible next states: AFTER_ATTRIBUTE_NAME */
1298 advance_to_name_end (context);
1299 add_to_partial (context, context->start, context->iter);
1301 /* read the full name, if we enter the equals sign state
1302 * then add the attribute to the list (without the value),
1303 * otherwise store a partial chunk to be prepended later.
1305 if (context->iter != context->current_text_end)
1306 context->state = STATE_AFTER_ATTRIBUTE_NAME;
1309 case STATE_AFTER_ATTRIBUTE_NAME:
1310 /* Possible next states: AFTER_ATTRIBUTE_EQUALS_SIGN */
1312 skip_spaces (context);
1314 if (context->iter != context->current_text_end)
1316 /* The name has ended. Combine it with the partial chunk
1317 * if any; push it on the stack; enter next state.
1319 if (!name_validate (context, context->partial_chunk->str, error))
1322 add_attribute (context, context->partial_chunk);
1324 context->partial_chunk = NULL;
1325 context->start = NULL;
1327 if (*context->iter == '=')
1329 advance_char (context);
1330 context->state = STATE_AFTER_ATTRIBUTE_EQUALS_SIGN;
1338 G_MARKUP_ERROR_PARSE,
1339 _("Odd character '%s', expected a '=' after "
1340 "attribute name '%s' of element '%s'"),
1341 utf8_str (context->iter, buf),
1342 current_attribute (context),
1343 current_element (context));
1349 case STATE_BETWEEN_ATTRIBUTES:
1350 /* Possible next states: AFTER_CLOSE_ANGLE,
1351 * AFTER_ELISION_SLASH, INSIDE_ATTRIBUTE_NAME
1353 skip_spaces (context);
1355 if (context->iter != context->current_text_end)
1357 if (*context->iter == '/')
1359 advance_char (context);
1360 context->state = STATE_AFTER_ELISION_SLASH;
1362 else if (*context->iter == '>')
1364 advance_char (context);
1365 context->state = STATE_AFTER_CLOSE_ANGLE;
1367 else if (!IS_COMMON_NAME_END_CHAR (*(context->iter)))
1369 context->state = STATE_INSIDE_ATTRIBUTE_NAME;
1370 /* start of attribute name */
1371 context->start = context->iter;
1379 G_MARKUP_ERROR_PARSE,
1380 _("Odd character '%s', expected a '>' or '/' "
1381 "character to end the start tag of "
1382 "element '%s', or optionally an attribute; "
1383 "perhaps you used an invalid character in "
1384 "an attribute name"),
1385 utf8_str (context->iter, buf),
1386 current_element (context));
1389 /* If we're done with attributes, invoke
1390 * the start_element callback
1392 if (context->state == STATE_AFTER_ELISION_SLASH ||
1393 context->state == STATE_AFTER_CLOSE_ANGLE)
1394 emit_start_element (context, error);
1398 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN:
1399 /* Possible next state: INSIDE_ATTRIBUTE_VALUE_[SQ/DQ] */
1401 skip_spaces (context);
1403 if (context->iter != context->current_text_end)
1405 if (*context->iter == '"')
1407 advance_char (context);
1408 context->state = STATE_INSIDE_ATTRIBUTE_VALUE_DQ;
1409 context->start = context->iter;
1411 else if (*context->iter == '\'')
1413 advance_char (context);
1414 context->state = STATE_INSIDE_ATTRIBUTE_VALUE_SQ;
1415 context->start = context->iter;
1423 G_MARKUP_ERROR_PARSE,
1424 _("Odd character '%s', expected an open quote mark "
1425 "after the equals sign when giving value for "
1426 "attribute '%s' of element '%s'"),
1427 utf8_str (context->iter, buf),
1428 current_attribute (context),
1429 current_element (context));
1434 case STATE_INSIDE_ATTRIBUTE_VALUE_SQ:
1435 case STATE_INSIDE_ATTRIBUTE_VALUE_DQ:
1436 /* Possible next states: BETWEEN_ATTRIBUTES */
1440 if (context->state == STATE_INSIDE_ATTRIBUTE_VALUE_SQ)
1451 if (*context->iter == delim)
1454 while (advance_char (context));
1456 if (context->iter == context->current_text_end)
1458 /* The value hasn't necessarily ended. Merge with
1459 * partial chunk, leave state unchanged.
1461 add_to_partial (context, context->start, context->iter);
1466 /* The value has ended at the quote mark. Combine it
1467 * with the partial chunk if any; set it for the current
1470 add_to_partial (context, context->start, context->iter);
1472 g_assert (context->cur_attr >= 0);
1474 if (unescape_gstring_inplace (context, context->partial_chunk, &is_ascii, error) &&
1475 (is_ascii || text_validate (context, context->partial_chunk->str,
1476 context->partial_chunk->len, error)))
1478 /* success, advance past quote and set state. */
1479 context->attr_values[context->cur_attr] = context->partial_chunk;
1480 context->partial_chunk = NULL;
1481 advance_char (context);
1482 context->state = STATE_BETWEEN_ATTRIBUTES;
1483 context->start = NULL;
1486 truncate_partial (context);
1490 case STATE_INSIDE_TEXT:
1491 /* Possible next states: AFTER_OPEN_ANGLE */
1494 if (*context->iter == '<')
1497 while (advance_char (context));
1499 /* The text hasn't necessarily ended. Merge with
1500 * partial chunk, leave state unchanged.
1503 add_to_partial (context, context->start, context->iter);
1505 if (context->iter != context->current_text_end)
1509 /* The text has ended at the open angle. Call the text
1512 if (unescape_gstring_inplace (context, context->partial_chunk, &is_ascii, error) &&
1513 (is_ascii || text_validate (context, context->partial_chunk->str,
1514 context->partial_chunk->len, error)))
1516 GError *tmp_error = NULL;
1518 if (context->parser->text)
1519 (*context->parser->text) (context,
1520 context->partial_chunk->str,
1521 context->partial_chunk->len,
1525 if (tmp_error == NULL)
1527 /* advance past open angle and set state. */
1528 advance_char (context);
1529 context->state = STATE_AFTER_OPEN_ANGLE;
1530 /* could begin a passthrough */
1531 context->start = context->iter;
1534 propagate_error (context, error, tmp_error);
1537 truncate_partial (context);
1541 case STATE_AFTER_CLOSE_TAG_SLASH:
1542 /* Possible next state: INSIDE_CLOSE_TAG_NAME */
1543 if (!IS_COMMON_NAME_END_CHAR (*(context->iter)))
1545 context->state = STATE_INSIDE_CLOSE_TAG_NAME;
1547 /* start of tag name */
1548 context->start = context->iter;
1556 G_MARKUP_ERROR_PARSE,
1557 _("'%s' is not a valid character following "
1558 "the characters '</'; '%s' may not begin an "
1560 utf8_str (context->iter, buf),
1561 utf8_str (context->iter, buf));
1565 case STATE_INSIDE_CLOSE_TAG_NAME:
1566 /* Possible next state: AFTER_CLOSE_TAG_NAME */
1567 advance_to_name_end (context);
1568 add_to_partial (context, context->start, context->iter);
1570 if (context->iter != context->current_text_end)
1571 context->state = STATE_AFTER_CLOSE_TAG_NAME;
1574 case STATE_AFTER_CLOSE_TAG_NAME:
1575 /* Possible next state: AFTER_CLOSE_TAG_SLASH */
1577 skip_spaces (context);
1579 if (context->iter != context->current_text_end)
1581 GString *close_name;
1583 close_name = context->partial_chunk;
1584 context->partial_chunk = NULL;
1586 if (*context->iter != '>')
1592 G_MARKUP_ERROR_PARSE,
1593 _("'%s' is not a valid character following "
1594 "the close element name '%s'; the allowed "
1595 "character is '>'"),
1596 utf8_str (context->iter, buf),
1599 else if (context->tag_stack == NULL)
1603 G_MARKUP_ERROR_PARSE,
1604 _("Element '%s' was closed, no element "
1605 "is currently open"),
1608 else if (strcmp (close_name->str, current_element (context)) != 0)
1612 G_MARKUP_ERROR_PARSE,
1613 _("Element '%s' was closed, but the currently "
1614 "open element is '%s'"),
1616 current_element (context));
1620 advance_char (context);
1621 context->state = STATE_AFTER_CLOSE_ANGLE;
1622 context->start = NULL;
1624 emit_end_element (context, error);
1626 context->partial_chunk = close_name;
1627 truncate_partial (context);
1631 case STATE_INSIDE_PASSTHROUGH:
1632 /* Possible next state: AFTER_CLOSE_ANGLE */
1635 if (*context->iter == '<')
1637 if (*context->iter == '>')
1643 add_to_partial (context, context->start, context->iter);
1644 context->start = context->iter;
1646 str = context->partial_chunk->str;
1647 len = context->partial_chunk->len;
1649 if (str[1] == '?' && str[len - 1] == '?')
1651 if (strncmp (str, "<!--", 4) == 0 &&
1652 strcmp (str + len - 2, "--") == 0)
1654 if (strncmp (str, "<![CDATA[", 9) == 0 &&
1655 strcmp (str + len - 2, "]]") == 0)
1657 if (strncmp (str, "<!DOCTYPE", 9) == 0 &&
1658 context->balance == 0)
1662 while (advance_char (context));
1664 if (context->iter == context->current_text_end)
1666 /* The passthrough hasn't necessarily ended. Merge with
1667 * partial chunk, leave state unchanged.
1669 add_to_partial (context, context->start, context->iter);
1673 /* The passthrough has ended at the close angle. Combine
1674 * it with the partial chunk if any. Call the passthrough
1675 * callback. Note that the open/close angles are
1676 * included in the text of the passthrough.
1678 GError *tmp_error = NULL;
1680 advance_char (context); /* advance past close angle */
1681 add_to_partial (context, context->start, context->iter);
1683 if (context->flags & G_MARKUP_TREAT_CDATA_AS_TEXT &&
1684 strncmp (context->partial_chunk->str, "<![CDATA[", 9) == 0)
1686 if (context->parser->text &&
1687 text_validate (context,
1688 context->partial_chunk->str + 9,
1689 context->partial_chunk->len - 12,
1691 (*context->parser->text) (context,
1692 context->partial_chunk->str + 9,
1693 context->partial_chunk->len - 12,
1697 else if (context->parser->passthrough &&
1698 text_validate (context,
1699 context->partial_chunk->str,
1700 context->partial_chunk->len,
1702 (*context->parser->passthrough) (context,
1703 context->partial_chunk->str,
1704 context->partial_chunk->len,
1708 truncate_partial (context);
1710 if (tmp_error == NULL)
1712 context->state = STATE_AFTER_CLOSE_ANGLE;
1713 context->start = context->iter; /* could begin text */
1716 propagate_error (context, error, tmp_error);
1725 g_assert_not_reached ();
1731 context->parsing = FALSE;
1733 return context->state != STATE_ERROR;
1737 * g_markup_parse_context_end_parse:
1738 * @context: a #GMarkupParseContext
1739 * @error: return location for a #GError
1741 * Signals to the #GMarkupParseContext that all data has been
1742 * fed into the parse context with g_markup_parse_context_parse().
1744 * This function reports an error if the document isn't complete,
1745 * for example if elements are still open.
1747 * Return value: %TRUE on success, %FALSE if an error was set
1750 g_markup_parse_context_end_parse (GMarkupParseContext *context,
1753 g_return_val_if_fail (context != NULL, FALSE);
1754 g_return_val_if_fail (!context->parsing, FALSE);
1755 g_return_val_if_fail (context->state != STATE_ERROR, FALSE);
1757 if (context->partial_chunk != NULL)
1759 g_string_free (context->partial_chunk, TRUE);
1760 context->partial_chunk = NULL;
1763 if (context->document_empty)
1765 set_error_literal (context, error, G_MARKUP_ERROR_EMPTY,
1766 _("Document was empty or contained only whitespace"));
1770 context->parsing = TRUE;
1772 switch (context->state)
1778 case STATE_AFTER_OPEN_ANGLE:
1779 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1780 _("Document ended unexpectedly just after an open angle bracket '<'"));
1783 case STATE_AFTER_CLOSE_ANGLE:
1784 if (context->tag_stack != NULL)
1786 /* Error message the same as for INSIDE_TEXT */
1787 set_error (context, error, G_MARKUP_ERROR_PARSE,
1788 _("Document ended unexpectedly with elements still open - "
1789 "'%s' was the last element opened"),
1790 current_element (context));
1794 case STATE_AFTER_ELISION_SLASH:
1795 set_error (context, error, G_MARKUP_ERROR_PARSE,
1796 _("Document ended unexpectedly, expected to see a close angle "
1797 "bracket ending the tag <%s/>"), current_element (context));
1800 case STATE_INSIDE_OPEN_TAG_NAME:
1801 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1802 _("Document ended unexpectedly inside an element name"));
1805 case STATE_INSIDE_ATTRIBUTE_NAME:
1806 case STATE_AFTER_ATTRIBUTE_NAME:
1807 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1808 _("Document ended unexpectedly inside an attribute name"));
1811 case STATE_BETWEEN_ATTRIBUTES:
1812 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1813 _("Document ended unexpectedly inside an element-opening "
1817 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN:
1818 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1819 _("Document ended unexpectedly after the equals sign "
1820 "following an attribute name; no attribute value"));
1823 case STATE_INSIDE_ATTRIBUTE_VALUE_SQ:
1824 case STATE_INSIDE_ATTRIBUTE_VALUE_DQ:
1825 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1826 _("Document ended unexpectedly while inside an attribute "
1830 case STATE_INSIDE_TEXT:
1831 g_assert (context->tag_stack != NULL);
1832 set_error (context, error, G_MARKUP_ERROR_PARSE,
1833 _("Document ended unexpectedly with elements still open - "
1834 "'%s' was the last element opened"),
1835 current_element (context));
1838 case STATE_AFTER_CLOSE_TAG_SLASH:
1839 case STATE_INSIDE_CLOSE_TAG_NAME:
1840 case STATE_AFTER_CLOSE_TAG_NAME:
1841 set_error (context, error, G_MARKUP_ERROR_PARSE,
1842 _("Document ended unexpectedly inside the close tag for "
1843 "element '%s'"), current_element (context));
1846 case STATE_INSIDE_PASSTHROUGH:
1847 set_error_literal (context, error, G_MARKUP_ERROR_PARSE,
1848 _("Document ended unexpectedly inside a comment or "
1849 "processing instruction"));
1854 g_assert_not_reached ();
1858 context->parsing = FALSE;
1860 return context->state != STATE_ERROR;
1864 * g_markup_parse_context_get_element:
1865 * @context: a #GMarkupParseContext
1867 * Retrieves the name of the currently open element.
1869 * If called from the start_element or end_element handlers this will
1870 * give the element_name as passed to those functions. For the parent
1871 * elements, see g_markup_parse_context_get_element_stack().
1873 * Returns: the name of the currently open element, or %NULL
1878 g_markup_parse_context_get_element (GMarkupParseContext *context)
1880 g_return_val_if_fail (context != NULL, NULL);
1882 if (context->tag_stack == NULL)
1885 return current_element (context);
1889 * g_markup_parse_context_get_element_stack:
1890 * @context: a #GMarkupParseContext
1892 * Retrieves the element stack from the internal state of the parser.
1894 * The returned #GSList is a list of strings where the first item is
1895 * the currently open tag (as would be returned by
1896 * g_markup_parse_context_get_element()) and the next item is its
1899 * This function is intended to be used in the start_element and
1900 * end_element handlers where g_markup_parse_context_get_element()
1901 * would merely return the name of the element that is being
1904 * Returns: the element stack, which must not be modified
1909 g_markup_parse_context_get_element_stack (GMarkupParseContext *context)
1911 g_return_val_if_fail (context != NULL, NULL);
1912 return context->tag_stack;
1916 * g_markup_parse_context_get_position:
1917 * @context: a #GMarkupParseContext
1918 * @line_number: (allow-none): return location for a line number, or %NULL
1919 * @char_number: (allow-none): return location for a char-on-line number, or %NULL
1921 * Retrieves the current line number and the number of the character on
1922 * that line. Intended for use in error messages; there are no strict
1923 * semantics for what constitutes the "current" line number other than
1924 * "the best number we could come up with for error messages."
1927 g_markup_parse_context_get_position (GMarkupParseContext *context,
1931 g_return_if_fail (context != NULL);
1934 *line_number = context->line_number;
1937 *char_number = context->char_number;
1941 * g_markup_parse_context_get_user_data:
1942 * @context: a #GMarkupParseContext
1944 * Returns the user_data associated with @context.
1946 * This will either be the user_data that was provided to
1947 * g_markup_parse_context_new() or to the most recent call
1948 * of g_markup_parse_context_push().
1950 * Returns: the provided user_data. The returned data belongs to
1951 * the markup context and will be freed when
1952 * g_markup_parse_context_free() is called.
1957 g_markup_parse_context_get_user_data (GMarkupParseContext *context)
1959 return context->user_data;
1963 * g_markup_parse_context_push:
1964 * @context: a #GMarkupParseContext
1965 * @parser: a #GMarkupParser
1966 * @user_data: user data to pass to #GMarkupParser functions
1968 * Temporarily redirects markup data to a sub-parser.
1970 * This function may only be called from the start_element handler of
1971 * a #GMarkupParser. It must be matched with a corresponding call to
1972 * g_markup_parse_context_pop() in the matching end_element handler
1973 * (except in the case that the parser aborts due to an error).
1975 * All tags, text and other data between the matching tags is
1976 * redirected to the subparser given by @parser. @user_data is used
1977 * as the user_data for that parser. @user_data is also passed to the
1978 * error callback in the event that an error occurs. This includes
1979 * errors that occur in subparsers of the subparser.
1981 * The end tag matching the start tag for which this call was made is
1982 * handled by the previous parser (which is given its own user_data)
1983 * which is why g_markup_parse_context_pop() is provided to allow "one
1984 * last access" to the @user_data provided to this function. In the
1985 * case of error, the @user_data provided here is passed directly to
1986 * the error callback of the subparser and g_markup_parse_context_pop()
1987 * should not be called. In either case, if @user_data was allocated
1988 * then it ought to be freed from both of these locations.
1990 * This function is not intended to be directly called by users
1991 * interested in invoking subparsers. Instead, it is intended to be
1992 * used by the subparsers themselves to implement a higher-level
1995 * As an example, see the following implementation of a simple
1996 * parser that counts the number of tags encountered.
2005 * counter_start_element (GMarkupParseContext *context,
2006 * const gchar *element_name,
2007 * const gchar **attribute_names,
2008 * const gchar **attribute_values,
2009 * gpointer user_data,
2012 * CounterData *data = user_data;
2014 * data->tag_count++;
2018 * counter_error (GMarkupParseContext *context,
2020 * gpointer user_data)
2022 * CounterData *data = user_data;
2024 * g_slice_free (CounterData, data);
2027 * static GMarkupParser counter_subparser =
2029 * counter_start_element,
2037 * In order to allow this parser to be easily used as a subparser, the
2038 * following interface is provided:
2042 * start_counting (GMarkupParseContext *context)
2044 * CounterData *data = g_slice_new (CounterData);
2046 * data->tag_count = 0;
2047 * g_markup_parse_context_push (context, &counter_subparser, data);
2051 * end_counting (GMarkupParseContext *context)
2053 * CounterData *data = g_markup_parse_context_pop (context);
2056 * result = data->tag_count;
2057 * g_slice_free (CounterData, data);
2063 * The subparser would then be used as follows:
2066 * static void start_element (context, element_name, ...)
2068 * if (strcmp (element_name, "count-these") == 0)
2069 * start_counting (context);
2071 * /* else, handle other tags... */
2074 * static void end_element (context, element_name, ...)
2076 * if (strcmp (element_name, "count-these") == 0)
2077 * g_print ("Counted %d tags\n", end_counting (context));
2079 * /* else, handle other tags... */
2086 g_markup_parse_context_push (GMarkupParseContext *context,
2087 const GMarkupParser *parser,
2090 GMarkupRecursionTracker *tracker;
2092 tracker = g_slice_new (GMarkupRecursionTracker);
2093 tracker->prev_element = context->subparser_element;
2094 tracker->prev_parser = context->parser;
2095 tracker->prev_user_data = context->user_data;
2097 context->subparser_element = current_element (context);
2098 context->parser = parser;
2099 context->user_data = user_data;
2101 context->subparser_stack = g_slist_prepend (context->subparser_stack,
2106 * g_markup_parse_context_pop:
2107 * @context: a #GMarkupParseContext
2109 * Completes the process of a temporary sub-parser redirection.
2111 * This function exists to collect the user_data allocated by a
2112 * matching call to g_markup_parse_context_push(). It must be called
2113 * in the end_element handler corresponding to the start_element
2114 * handler during which g_markup_parse_context_push() was called.
2115 * You must not call this function from the error callback -- the
2116 * @user_data is provided directly to the callback in that case.
2118 * This function is not intended to be directly called by users
2119 * interested in invoking subparsers. Instead, it is intended to
2120 * be used by the subparsers themselves to implement a higher-level
2123 * Returns: the user data passed to g_markup_parse_context_push()
2128 g_markup_parse_context_pop (GMarkupParseContext *context)
2132 if (!context->awaiting_pop)
2133 possibly_finish_subparser (context);
2135 g_assert (context->awaiting_pop);
2137 context->awaiting_pop = FALSE;
2139 /* valgrind friendliness */
2140 user_data = context->held_user_data;
2141 context->held_user_data = NULL;
2147 append_escaped_text (GString *str,
2156 end = text + length;
2161 next = g_utf8_next_char (p);
2166 g_string_append (str, "&");
2170 g_string_append (str, "<");
2174 g_string_append (str, ">");
2178 g_string_append (str, "'");
2182 g_string_append (str, """);
2186 c = g_utf8_get_char (p);
2187 if ((0x1 <= c && c <= 0x8) ||
2188 (0xb <= c && c <= 0xc) ||
2189 (0xe <= c && c <= 0x1f) ||
2190 (0x7f <= c && c <= 0x84) ||
2191 (0x86 <= c && c <= 0x9f))
2192 g_string_append_printf (str, "&#x%x;", c);
2194 g_string_append_len (str, p, next - p);
2203 * g_markup_escape_text:
2204 * @text: some valid UTF-8 text
2205 * @length: length of @text in bytes, or -1 if the text is nul-terminated
2207 * Escapes text so that the markup parser will parse it verbatim.
2208 * Less than, greater than, ampersand, etc. are replaced with the
2209 * corresponding entities. This function would typically be used
2210 * when writing out a file to be parsed with the markup parser.
2212 * Note that this function doesn't protect whitespace and line endings
2213 * from being processed according to the XML rules for normalization
2214 * of line endings and attribute values.
2216 * Note also that this function will produce character references in
2217 * the range of &#x1; ... &#x1f; for all control sequences
2218 * except for tabstop, newline and carriage return. The character
2219 * references in this range are not valid XML 1.0, but they are
2220 * valid XML 1.1 and will be accepted by the GMarkup parser.
2222 * Return value: a newly allocated string with the escaped text
2225 g_markup_escape_text (const gchar *text,
2230 g_return_val_if_fail (text != NULL, NULL);
2233 length = strlen (text);
2235 /* prealloc at least as long as original text */
2236 str = g_string_sized_new (length);
2237 append_escaped_text (str, text, length);
2239 return g_string_free (str, FALSE);
2244 * @format: a printf-style format string
2245 * @after: location to store a pointer to the character after
2246 * the returned conversion. On a %NULL return, returns the
2247 * pointer to the trailing NUL in the string
2249 * Find the next conversion in a printf-style format string.
2250 * Partially based on code from printf-parser.c,
2251 * Copyright (C) 1999-2000, 2002-2003 Free Software Foundation, Inc.
2253 * Return value: pointer to the next conversion in @format,
2254 * or %NULL, if none.
2257 find_conversion (const char *format,
2260 const char *start = format;
2263 while (*start != '\0' && *start != '%')
2280 /* Test for positional argument. */
2281 if (*cp >= '0' && *cp <= '9')
2285 for (np = cp; *np >= '0' && *np <= '9'; np++)
2291 /* Skip the flags. */
2305 /* Skip the field width. */
2310 /* Test for positional argument. */
2311 if (*cp >= '0' && *cp <= '9')
2315 for (np = cp; *np >= '0' && *np <= '9'; np++)
2323 for (; *cp >= '0' && *cp <= '9'; cp++)
2327 /* Skip the precision. */
2333 /* Test for positional argument. */
2334 if (*cp >= '0' && *cp <= '9')
2338 for (np = cp; *np >= '0' && *np <= '9'; np++)
2346 for (; *cp >= '0' && *cp <= '9'; cp++)
2351 /* Skip argument type/size specifiers. */
2352 while (*cp == 'h' ||
2361 /* Skip the conversion character. */
2369 * g_markup_vprintf_escaped:
2370 * @format: printf() style format string
2371 * @args: variable argument list, similar to vprintf()
2373 * Formats the data in @args according to @format, escaping
2374 * all string and character arguments in the fashion
2375 * of g_markup_escape_text(). See g_markup_printf_escaped().
2377 * Return value: newly allocated result from formatting
2378 * operation. Free with g_free().
2382 #pragma GCC diagnostic push
2383 #pragma GCC diagnostic ignored "-Wformat-nonliteral"
2386 g_markup_vprintf_escaped (const gchar *format,
2391 GString *result = NULL;
2392 gchar *output1 = NULL;
2393 gchar *output2 = NULL;
2394 const char *p, *op1, *op2;
2397 /* The technique here, is that we make two format strings that
2398 * have the identical conversions in the identical order to the
2399 * original strings, but differ in the text in-between. We
2400 * then use the normal g_strdup_vprintf() to format the arguments
2401 * with the two new format strings. By comparing the results,
2402 * we can figure out what segments of the output come from
2403 * the original format string, and what from the arguments,
2404 * and thus know what portions of the string to escape.
2406 * For instance, for:
2408 * g_markup_printf_escaped ("%s ate %d apples", "Susan & Fred", 5);
2410 * We form the two format strings "%sX%dX" and %sY%sY". The results
2411 * of formatting with those two strings are
2413 * "%sX%dX" => "Susan & FredX5X"
2414 * "%sY%dY" => "Susan & FredY5Y"
2416 * To find the span of the first argument, we find the first position
2417 * where the two arguments differ, which tells us that the first
2418 * argument formatted to "Susan & Fred". We then escape that
2419 * to "Susan & Fred" and join up with the intermediate portions
2420 * of the format string and the second argument to get
2421 * "Susan & Fred ate 5 apples".
2424 /* Create the two modified format strings
2426 format1 = g_string_new (NULL);
2427 format2 = g_string_new (NULL);
2432 const char *conv = find_conversion (p, &after);
2436 g_string_append_len (format1, conv, after - conv);
2437 g_string_append_c (format1, 'X');
2438 g_string_append_len (format2, conv, after - conv);
2439 g_string_append_c (format2, 'Y');
2444 /* Use them to format the arguments
2446 G_VA_COPY (args2, args);
2448 output1 = g_strdup_vprintf (format1->str, args);
2456 output2 = g_strdup_vprintf (format2->str, args2);
2460 result = g_string_new (NULL);
2462 /* Iterate through the original format string again,
2463 * copying the non-conversion portions and the escaped
2464 * converted arguments to the output string.
2472 const char *output_start;
2473 const char *conv = find_conversion (p, &after);
2476 if (!conv) /* The end, after points to the trailing \0 */
2478 g_string_append_len (result, p, after - p);
2482 g_string_append_len (result, p, conv - p);
2484 while (*op1 == *op2)
2490 escaped = g_markup_escape_text (output_start, op1 - output_start);
2491 g_string_append (result, escaped);
2500 g_string_free (format1, TRUE);
2501 g_string_free (format2, TRUE);
2506 return g_string_free (result, FALSE);
2511 #pragma GCC diagnostic pop
2514 * g_markup_printf_escaped:
2515 * @format: printf() style format string
2516 * @...: the arguments to insert in the format string
2518 * Formats arguments according to @format, escaping
2519 * all string and character arguments in the fashion
2520 * of g_markup_escape_text(). This is useful when you
2521 * want to insert literal strings into XML-style markup
2522 * output, without having to worry that the strings
2523 * might themselves contain markup.
2526 * const char *store = "Fortnum & Mason";
2527 * const char *item = "Tea";
2530 * output = g_markup_printf_escaped ("<purchase>"
2531 * "<store>%s</store>"
2532 * "<item>%s</item>"
2533 * "</purchase>",
2537 * Return value: newly allocated result from formatting
2538 * operation. Free with g_free().
2543 g_markup_printf_escaped (const gchar *format, ...)
2548 va_start (args, format);
2549 result = g_markup_vprintf_escaped (format, args);
2556 g_markup_parse_boolean (const char *string,
2559 char const * const falses[] = { "false", "f", "no", "n", "0" };
2560 char const * const trues[] = { "true", "t", "yes", "y", "1" };
2563 for (i = 0; i < G_N_ELEMENTS (falses); i++)
2565 if (g_ascii_strcasecmp (string, falses[i]) == 0)
2574 for (i = 0; i < G_N_ELEMENTS (trues); i++)
2576 if (g_ascii_strcasecmp (string, trues[i]) == 0)
2589 * GMarkupCollectType:
2590 * @G_MARKUP_COLLECT_INVALID: used to terminate the list of attributes
2592 * @G_MARKUP_COLLECT_STRING: collect the string pointer directly from
2593 * the attribute_values[] array. Expects a parameter of type (const
2594 * char **). If %G_MARKUP_COLLECT_OPTIONAL is specified and the
2595 * attribute isn't present then the pointer will be set to %NULL
2596 * @G_MARKUP_COLLECT_STRDUP: as with %G_MARKUP_COLLECT_STRING, but
2597 * expects a parameter of type (char **) and g_strdup()s the
2598 * returned pointer. The pointer must be freed with g_free()
2599 * @G_MARKUP_COLLECT_BOOLEAN: expects a parameter of type (gboolean *)
2600 * and parses the attribute value as a boolean. Sets %FALSE if the
2601 * attribute isn't present. Valid boolean values consist of
2602 * (case-insensitive) "false", "f", "no", "n", "0" and "true", "t",
2604 * @G_MARKUP_COLLECT_TRISTATE: as with %G_MARKUP_COLLECT_BOOLEAN, but
2605 * in the case of a missing attribute a value is set that compares
2606 * equal to neither %FALSE nor %TRUE G_MARKUP_COLLECT_OPTIONAL is
2608 * @G_MARKUP_COLLECT_OPTIONAL: can be bitwise ORed with the other fields.
2609 * If present, allows the attribute not to appear. A default value
2610 * is set depending on what value type is used
2612 * A mixed enumerated type and flags field. You must specify one type
2613 * (string, strdup, boolean, tristate). Additionally, you may optionally
2614 * bitwise OR the type with the flag %G_MARKUP_COLLECT_OPTIONAL.
2616 * It is likely that this enum will be extended in the future to
2617 * support other types.
2621 * g_markup_collect_attributes:
2622 * @element_name: the current tag name
2623 * @attribute_names: the attribute names
2624 * @attribute_values: the attribute values
2625 * @error: a pointer to a #GError or %NULL
2626 * @first_type: the #GMarkupCollectType of the first attribute
2627 * @first_attr: the name of the first attribute
2628 * @...: a pointer to the storage location of the first attribute
2629 * (or %NULL), followed by more types names and pointers, ending
2630 * with %G_MARKUP_COLLECT_INVALID
2632 * Collects the attributes of the element from the data passed to the
2633 * #GMarkupParser start_element function, dealing with common error
2634 * conditions and supporting boolean values.
2636 * This utility function is not required to write a parser but can save
2639 * The @element_name, @attribute_names, @attribute_values and @error
2640 * parameters passed to the start_element callback should be passed
2641 * unmodified to this function.
2643 * Following these arguments is a list of "supported" attributes to collect.
2644 * It is an error to specify multiple attributes with the same name. If any
2645 * attribute not in the list appears in the @attribute_names array then an
2646 * unknown attribute error will result.
2648 * The #GMarkupCollectType field allows specifying the type of collection
2649 * to perform and if a given attribute must appear or is optional.
2651 * The attribute name is simply the name of the attribute to collect.
2653 * The pointer should be of the appropriate type (see the descriptions
2654 * under #GMarkupCollectType) and may be %NULL in case a particular
2655 * attribute is to be allowed but ignored.
2657 * This function deals with issuing errors for missing attributes
2658 * (of type %G_MARKUP_ERROR_MISSING_ATTRIBUTE), unknown attributes
2659 * (of type %G_MARKUP_ERROR_UNKNOWN_ATTRIBUTE) and duplicate
2660 * attributes (of type %G_MARKUP_ERROR_INVALID_CONTENT) as well
2661 * as parse errors for boolean-valued attributes (again of type
2662 * %G_MARKUP_ERROR_INVALID_CONTENT). In all of these cases %FALSE
2663 * will be returned and @error will be set as appropriate.
2665 * Return value: %TRUE if successful
2670 g_markup_collect_attributes (const gchar *element_name,
2671 const gchar **attribute_names,
2672 const gchar **attribute_values,
2674 GMarkupCollectType first_type,
2675 const gchar *first_attr,
2678 GMarkupCollectType type;
2690 va_start (ap, first_attr);
2691 while (type != G_MARKUP_COLLECT_INVALID)
2696 mandatory = !(type & G_MARKUP_COLLECT_OPTIONAL);
2697 type &= (G_MARKUP_COLLECT_OPTIONAL - 1);
2699 /* tristate records a value != TRUE and != FALSE
2700 * for the case where the attribute is missing
2702 if (type == G_MARKUP_COLLECT_TRISTATE)
2705 for (i = 0; attribute_names[i]; i++)
2706 if (i >= 40 || !(collected & (G_GUINT64_CONSTANT(1) << i)))
2707 if (!strcmp (attribute_names[i], attr))
2710 /* ISO C99 only promises that the user can pass up to 127 arguments.
2711 * Subtracting the first 4 arguments plus the final NULL and dividing
2712 * by 3 arguments per collected attribute, we are left with a maximum
2713 * number of supported attributes of (127 - 5) / 3 = 40.
2715 * In reality, nobody is ever going to call us with anywhere close to
2716 * 40 attributes to collect, so it is safe to assume that if i > 40
2717 * then the user has given some invalid or repeated arguments. These
2718 * problems will be caught and reported at the end of the function.
2720 * We know at this point that we have an error, but we don't know
2721 * what error it is, so just continue...
2724 collected |= (G_GUINT64_CONSTANT(1) << i);
2726 value = attribute_values[i];
2728 if (value == NULL && mandatory)
2730 g_set_error (error, G_MARKUP_ERROR,
2731 G_MARKUP_ERROR_MISSING_ATTRIBUTE,
2732 "element '%s' requires attribute '%s'",
2733 element_name, attr);
2741 case G_MARKUP_COLLECT_STRING:
2743 const char **str_ptr;
2745 str_ptr = va_arg (ap, const char **);
2747 if (str_ptr != NULL)
2752 case G_MARKUP_COLLECT_STRDUP:
2756 str_ptr = va_arg (ap, char **);
2758 if (str_ptr != NULL)
2759 *str_ptr = g_strdup (value);
2763 case G_MARKUP_COLLECT_BOOLEAN:
2764 case G_MARKUP_COLLECT_TRISTATE:
2769 bool_ptr = va_arg (ap, gboolean *);
2771 if (bool_ptr != NULL)
2773 if (type == G_MARKUP_COLLECT_TRISTATE)
2774 /* constructivists rejoice!
2775 * neither false nor true...
2779 else /* G_MARKUP_COLLECT_BOOLEAN */
2785 if (!g_markup_parse_boolean (value, va_arg (ap, gboolean *)))
2787 g_set_error (error, G_MARKUP_ERROR,
2788 G_MARKUP_ERROR_INVALID_CONTENT,
2789 "element '%s', attribute '%s', value '%s' "
2790 "cannot be parsed as a boolean value",
2791 element_name, attr, value);
2801 g_assert_not_reached ();
2804 type = va_arg (ap, GMarkupCollectType);
2805 attr = va_arg (ap, const char *);
2810 /* ensure we collected all the arguments */
2811 for (i = 0; attribute_names[i]; i++)
2812 if ((collected & (G_GUINT64_CONSTANT(1) << i)) == 0)
2814 /* attribute not collected: could be caused by two things.
2816 * 1) it doesn't exist in our list of attributes
2817 * 2) it existed but was matched by a duplicate attribute earlier
2823 for (j = 0; j < i; j++)
2824 if (strcmp (attribute_names[i], attribute_names[j]) == 0)
2828 /* j is now the first occurrence of attribute_names[i] */
2830 g_set_error (error, G_MARKUP_ERROR,
2831 G_MARKUP_ERROR_UNKNOWN_ATTRIBUTE,
2832 "attribute '%s' invalid for element '%s'",
2833 attribute_names[i], element_name);
2835 g_set_error (error, G_MARKUP_ERROR,
2836 G_MARKUP_ERROR_INVALID_CONTENT,
2837 "attribute '%s' given multiple times for element '%s'",
2838 attribute_names[i], element_name);
2846 /* replay the above to free allocations */
2850 va_start (ap, first_attr);
2851 while (type != G_MARKUP_COLLECT_INVALID)
2855 ptr = va_arg (ap, gpointer);
2859 switch (type & (G_MARKUP_COLLECT_OPTIONAL - 1))
2861 case G_MARKUP_COLLECT_STRDUP:
2863 g_free (*(char **) ptr);
2865 case G_MARKUP_COLLECT_STRING:
2866 *(char **) ptr = NULL;
2869 case G_MARKUP_COLLECT_BOOLEAN:
2870 *(gboolean *) ptr = FALSE;
2873 case G_MARKUP_COLLECT_TRISTATE:
2874 *(gboolean *) ptr = -1;
2879 type = va_arg (ap, GMarkupCollectType);
2880 attr = va_arg (ap, const char *);