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 attr_names
= g_newa (const gchar
*, context
->cur_attr
+ 2);
1014 attr_values
= g_newa (const gchar
*, context
->cur_attr
+ 2);
1015 for (i
= 0; i
< context
->cur_attr
+ 1; i
++)
1017 attr_names
[i
] = context
->attr_names
[i
]->str
;
1018 attr_values
[i
] = context
->attr_values
[i
]->str
;
1020 attr_names
[i
] = NULL
;
1021 attr_values
[i
] = NULL
;
1023 /* Call user callback for element start */
1025 start_name
= current_element (context
);
1027 if (context
->parser
->start_element
&&
1028 name_validate (context
, start_name
, error
))
1029 (* context
->parser
->start_element
) (context
,
1031 (const gchar
**)attr_names
,
1032 (const gchar
**)attr_values
,
1035 clear_attributes (context
);
1037 if (tmp_error
!= NULL
)
1038 propagate_error (context
, error
, tmp_error
);
1042 * g_markup_parse_context_parse:
1043 * @context: a #GMarkupParseContext
1044 * @text: chunk of text to parse
1045 * @text_len: length of @text in bytes
1046 * @error: return location for a #GError
1048 * Feed some data to the #GMarkupParseContext.
1050 * The data need not be valid UTF-8; an error will be signaled if
1051 * it's invalid. The data need not be an entire document; you can
1052 * feed a document into the parser incrementally, via multiple calls
1053 * to this function. Typically, as you receive data from a network
1054 * connection or file, you feed each received chunk of data into this
1055 * function, aborting the process if an error occurs. Once an error
1056 * is reported, no further data may be fed to the #GMarkupParseContext;
1057 * all errors are fatal.
1059 * Return value: %FALSE if an error occurred, %TRUE on success
1062 g_markup_parse_context_parse (GMarkupParseContext
*context
,
1067 g_return_val_if_fail (context
!= NULL
, FALSE
);
1068 g_return_val_if_fail (text
!= NULL
, FALSE
);
1069 g_return_val_if_fail (context
->state
!= STATE_ERROR
, FALSE
);
1070 g_return_val_if_fail (!context
->parsing
, FALSE
);
1073 text_len
= strlen (text
);
1078 context
->parsing
= TRUE
;
1081 context
->current_text
= text
;
1082 context
->current_text_len
= text_len
;
1083 context
->current_text_end
= context
->current_text
+ text_len
;
1084 context
->iter
= context
->current_text
;
1085 context
->start
= context
->iter
;
1087 while (context
->iter
!= context
->current_text_end
)
1089 switch (context
->state
)
1092 /* Possible next state: AFTER_OPEN_ANGLE */
1094 g_assert (context
->tag_stack
== NULL
);
1096 /* whitespace is ignored outside of any elements */
1097 skip_spaces (context
);
1099 if (context
->iter
!= context
->current_text_end
)
1101 if (*context
->iter
== '<')
1103 /* Move after the open angle */
1104 advance_char (context
);
1106 context
->state
= STATE_AFTER_OPEN_ANGLE
;
1108 /* this could start a passthrough */
1109 context
->start
= context
->iter
;
1111 /* document is now non-empty */
1112 context
->document_empty
= FALSE
;
1116 set_error_literal (context
,
1118 G_MARKUP_ERROR_PARSE
,
1119 _("Document must begin with an element (e.g. <book>)"));
1124 case STATE_AFTER_OPEN_ANGLE
:
1125 /* Possible next states: INSIDE_OPEN_TAG_NAME,
1126 * AFTER_CLOSE_TAG_SLASH, INSIDE_PASSTHROUGH
1128 if (*context
->iter
== '?' ||
1129 *context
->iter
== '!')
1131 /* include < in the passthrough */
1132 const gchar
*openangle
= "<";
1133 add_to_partial (context
, openangle
, openangle
+ 1);
1134 context
->start
= context
->iter
;
1135 context
->balance
= 1;
1136 context
->state
= STATE_INSIDE_PASSTHROUGH
;
1138 else if (*context
->iter
== '/')
1141 advance_char (context
);
1143 context
->state
= STATE_AFTER_CLOSE_TAG_SLASH
;
1145 else if (!IS_COMMON_NAME_END_CHAR (*(context
->iter
)))
1147 context
->state
= STATE_INSIDE_OPEN_TAG_NAME
;
1149 /* start of tag name */
1150 context
->start
= context
->iter
;
1158 G_MARKUP_ERROR_PARSE
,
1159 _("'%s' is not a valid character following "
1160 "a '<' character; it may not begin an "
1162 utf8_str (context
->iter
, buf
));
1166 /* The AFTER_CLOSE_ANGLE state is actually sort of
1167 * broken, because it doesn't correspond to a range
1168 * of characters in the input stream as the others do,
1169 * and thus makes things harder to conceptualize
1171 case STATE_AFTER_CLOSE_ANGLE
:
1172 /* Possible next states: INSIDE_TEXT, STATE_START */
1173 if (context
->tag_stack
== NULL
)
1175 context
->start
= NULL
;
1176 context
->state
= STATE_START
;
1180 context
->start
= context
->iter
;
1181 context
->state
= STATE_INSIDE_TEXT
;
1185 case STATE_AFTER_ELISION_SLASH
:
1186 /* Possible next state: AFTER_CLOSE_ANGLE */
1189 /* We need to pop the tag stack and call the end_element
1190 * function, since this is the close tag
1192 GError
*tmp_error
= NULL
;
1194 g_assert (context
->tag_stack
!= NULL
);
1196 possibly_finish_subparser (context
);
1199 if (context
->parser
->end_element
)
1200 (* context
->parser
->end_element
) (context
,
1201 current_element (context
),
1205 ensure_no_outstanding_subparser (context
);
1209 mark_error (context
, tmp_error
);
1210 g_propagate_error (error
, tmp_error
);
1214 if (*context
->iter
== '>')
1216 /* move after the close angle */
1217 advance_char (context
);
1218 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1226 G_MARKUP_ERROR_PARSE
,
1227 _("Odd character '%s', expected a '>' character "
1228 "to end the empty-element tag '%s'"),
1229 utf8_str (context
->iter
, buf
),
1230 current_element (context
));
1237 case STATE_INSIDE_OPEN_TAG_NAME
:
1238 /* Possible next states: BETWEEN_ATTRIBUTES */
1240 /* if there's a partial chunk then it's the first part of the
1241 * tag name. If there's a context->start then it's the start
1242 * of the tag name in current_text, the partial chunk goes
1243 * before that start though.
1245 advance_to_name_end (context
);
1247 if (context
->iter
== context
->current_text_end
)
1249 /* The name hasn't necessarily ended. Merge with
1250 * partial chunk, leave state unchanged.
1252 add_to_partial (context
, context
->start
, context
->iter
);
1256 /* The name has ended. Combine it with the partial chunk
1257 * if any; push it on the stack; enter next state.
1259 add_to_partial (context
, context
->start
, context
->iter
);
1260 push_partial_as_tag (context
);
1262 context
->state
= STATE_BETWEEN_ATTRIBUTES
;
1263 context
->start
= NULL
;
1267 case STATE_INSIDE_ATTRIBUTE_NAME
:
1268 /* Possible next states: AFTER_ATTRIBUTE_NAME */
1270 advance_to_name_end (context
);
1271 add_to_partial (context
, context
->start
, context
->iter
);
1273 /* read the full name, if we enter the equals sign state
1274 * then add the attribute to the list (without the value),
1275 * otherwise store a partial chunk to be prepended later.
1277 if (context
->iter
!= context
->current_text_end
)
1278 context
->state
= STATE_AFTER_ATTRIBUTE_NAME
;
1281 case STATE_AFTER_ATTRIBUTE_NAME
:
1282 /* Possible next states: AFTER_ATTRIBUTE_EQUALS_SIGN */
1284 skip_spaces (context
);
1286 if (context
->iter
!= context
->current_text_end
)
1288 /* The name has ended. Combine it with the partial chunk
1289 * if any; push it on the stack; enter next state.
1291 if (!name_validate (context
, context
->partial_chunk
->str
, error
))
1294 add_attribute (context
, context
->partial_chunk
);
1296 context
->partial_chunk
= NULL
;
1297 context
->start
= NULL
;
1299 if (*context
->iter
== '=')
1301 advance_char (context
);
1302 context
->state
= STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
;
1310 G_MARKUP_ERROR_PARSE
,
1311 _("Odd character '%s', expected a '=' after "
1312 "attribute name '%s' of element '%s'"),
1313 utf8_str (context
->iter
, buf
),
1314 current_attribute (context
),
1315 current_element (context
));
1321 case STATE_BETWEEN_ATTRIBUTES
:
1322 /* Possible next states: AFTER_CLOSE_ANGLE,
1323 * AFTER_ELISION_SLASH, INSIDE_ATTRIBUTE_NAME
1325 skip_spaces (context
);
1327 if (context
->iter
!= context
->current_text_end
)
1329 if (*context
->iter
== '/')
1331 advance_char (context
);
1332 context
->state
= STATE_AFTER_ELISION_SLASH
;
1334 else if (*context
->iter
== '>')
1336 advance_char (context
);
1337 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1339 else if (!IS_COMMON_NAME_END_CHAR (*(context
->iter
)))
1341 context
->state
= STATE_INSIDE_ATTRIBUTE_NAME
;
1342 /* start of attribute name */
1343 context
->start
= context
->iter
;
1351 G_MARKUP_ERROR_PARSE
,
1352 _("Odd character '%s', expected a '>' or '/' "
1353 "character to end the start tag of "
1354 "element '%s', or optionally an attribute; "
1355 "perhaps you used an invalid character in "
1356 "an attribute name"),
1357 utf8_str (context
->iter
, buf
),
1358 current_element (context
));
1361 /* If we're done with attributes, invoke
1362 * the start_element callback
1364 if (context
->state
== STATE_AFTER_ELISION_SLASH
||
1365 context
->state
== STATE_AFTER_CLOSE_ANGLE
)
1366 emit_start_element (context
, error
);
1370 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
:
1371 /* Possible next state: INSIDE_ATTRIBUTE_VALUE_[SQ/DQ] */
1373 skip_spaces (context
);
1375 if (context
->iter
!= context
->current_text_end
)
1377 if (*context
->iter
== '"')
1379 advance_char (context
);
1380 context
->state
= STATE_INSIDE_ATTRIBUTE_VALUE_DQ
;
1381 context
->start
= context
->iter
;
1383 else if (*context
->iter
== '\'')
1385 advance_char (context
);
1386 context
->state
= STATE_INSIDE_ATTRIBUTE_VALUE_SQ
;
1387 context
->start
= context
->iter
;
1395 G_MARKUP_ERROR_PARSE
,
1396 _("Odd character '%s', expected an open quote mark "
1397 "after the equals sign when giving value for "
1398 "attribute '%s' of element '%s'"),
1399 utf8_str (context
->iter
, buf
),
1400 current_attribute (context
),
1401 current_element (context
));
1406 case STATE_INSIDE_ATTRIBUTE_VALUE_SQ
:
1407 case STATE_INSIDE_ATTRIBUTE_VALUE_DQ
:
1408 /* Possible next states: BETWEEN_ATTRIBUTES */
1412 if (context
->state
== STATE_INSIDE_ATTRIBUTE_VALUE_SQ
)
1423 if (*context
->iter
== delim
)
1426 while (advance_char (context
));
1428 if (context
->iter
== context
->current_text_end
)
1430 /* The value hasn't necessarily ended. Merge with
1431 * partial chunk, leave state unchanged.
1433 add_to_partial (context
, context
->start
, context
->iter
);
1438 /* The value has ended at the quote mark. Combine it
1439 * with the partial chunk if any; set it for the current
1442 add_to_partial (context
, context
->start
, context
->iter
);
1444 g_assert (context
->cur_attr
>= 0);
1446 if (unescape_gstring_inplace (context
, context
->partial_chunk
, &is_ascii
, error
) &&
1447 (is_ascii
|| text_validate (context
, context
->partial_chunk
->str
,
1448 context
->partial_chunk
->len
, error
)))
1450 /* success, advance past quote and set state. */
1451 context
->attr_values
[context
->cur_attr
] = context
->partial_chunk
;
1452 context
->partial_chunk
= NULL
;
1453 advance_char (context
);
1454 context
->state
= STATE_BETWEEN_ATTRIBUTES
;
1455 context
->start
= NULL
;
1458 truncate_partial (context
);
1462 case STATE_INSIDE_TEXT
:
1463 /* Possible next states: AFTER_OPEN_ANGLE */
1466 if (*context
->iter
== '<')
1469 while (advance_char (context
));
1471 /* The text hasn't necessarily ended. Merge with
1472 * partial chunk, leave state unchanged.
1475 add_to_partial (context
, context
->start
, context
->iter
);
1477 if (context
->iter
!= context
->current_text_end
)
1481 /* The text has ended at the open angle. Call the text
1484 if (unescape_gstring_inplace (context
, context
->partial_chunk
, &is_ascii
, error
) &&
1485 (is_ascii
|| text_validate (context
, context
->partial_chunk
->str
,
1486 context
->partial_chunk
->len
, error
)))
1488 GError
*tmp_error
= NULL
;
1490 if (context
->parser
->text
)
1491 (*context
->parser
->text
) (context
,
1492 context
->partial_chunk
->str
,
1493 context
->partial_chunk
->len
,
1497 if (tmp_error
== NULL
)
1499 /* advance past open angle and set state. */
1500 advance_char (context
);
1501 context
->state
= STATE_AFTER_OPEN_ANGLE
;
1502 /* could begin a passthrough */
1503 context
->start
= context
->iter
;
1506 propagate_error (context
, error
, tmp_error
);
1509 truncate_partial (context
);
1513 case STATE_AFTER_CLOSE_TAG_SLASH
:
1514 /* Possible next state: INSIDE_CLOSE_TAG_NAME */
1515 if (!IS_COMMON_NAME_END_CHAR (*(context
->iter
)))
1517 context
->state
= STATE_INSIDE_CLOSE_TAG_NAME
;
1519 /* start of tag name */
1520 context
->start
= context
->iter
;
1528 G_MARKUP_ERROR_PARSE
,
1529 _("'%s' is not a valid character following "
1530 "the characters '</'; '%s' may not begin an "
1532 utf8_str (context
->iter
, buf
),
1533 utf8_str (context
->iter
, buf
));
1537 case STATE_INSIDE_CLOSE_TAG_NAME
:
1538 /* Possible next state: AFTER_CLOSE_TAG_NAME */
1539 advance_to_name_end (context
);
1540 add_to_partial (context
, context
->start
, context
->iter
);
1542 if (context
->iter
!= context
->current_text_end
)
1543 context
->state
= STATE_AFTER_CLOSE_TAG_NAME
;
1546 case STATE_AFTER_CLOSE_TAG_NAME
:
1547 /* Possible next state: AFTER_CLOSE_TAG_SLASH */
1549 skip_spaces (context
);
1551 if (context
->iter
!= context
->current_text_end
)
1553 GString
*close_name
;
1555 close_name
= context
->partial_chunk
;
1556 context
->partial_chunk
= NULL
;
1558 if (*context
->iter
!= '>')
1564 G_MARKUP_ERROR_PARSE
,
1565 _("'%s' is not a valid character following "
1566 "the close element name '%s'; the allowed "
1567 "character is '>'"),
1568 utf8_str (context
->iter
, buf
),
1571 else if (context
->tag_stack
== NULL
)
1575 G_MARKUP_ERROR_PARSE
,
1576 _("Element '%s' was closed, no element "
1577 "is currently open"),
1580 else if (strcmp (close_name
->str
, current_element (context
)) != 0)
1584 G_MARKUP_ERROR_PARSE
,
1585 _("Element '%s' was closed, but the currently "
1586 "open element is '%s'"),
1588 current_element (context
));
1593 advance_char (context
);
1594 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1595 context
->start
= NULL
;
1597 possibly_finish_subparser (context
);
1599 /* call the end_element callback */
1601 if (context
->parser
->end_element
)
1602 (* context
->parser
->end_element
) (context
,
1607 ensure_no_outstanding_subparser (context
);
1611 propagate_error (context
, error
, tmp_error
);
1613 context
->partial_chunk
= close_name
;
1614 truncate_partial (context
);
1618 case STATE_INSIDE_PASSTHROUGH
:
1619 /* Possible next state: AFTER_CLOSE_ANGLE */
1622 if (*context
->iter
== '<')
1624 if (*context
->iter
== '>')
1630 add_to_partial (context
, context
->start
, context
->iter
);
1631 context
->start
= context
->iter
;
1633 str
= context
->partial_chunk
->str
;
1634 len
= context
->partial_chunk
->len
;
1636 if (str
[1] == '?' && str
[len
- 1] == '?')
1638 if (strncmp (str
, "<!--", 4) == 0 &&
1639 strcmp (str
+ len
- 2, "--") == 0)
1641 if (strncmp (str
, "<![CDATA[", 9) == 0 &&
1642 strcmp (str
+ len
- 2, "]]") == 0)
1644 if (strncmp (str
, "<!DOCTYPE", 9) == 0 &&
1645 context
->balance
== 0)
1649 while (advance_char (context
));
1651 if (context
->iter
== context
->current_text_end
)
1653 /* The passthrough hasn't necessarily ended. Merge with
1654 * partial chunk, leave state unchanged.
1656 add_to_partial (context
, context
->start
, context
->iter
);
1660 /* The passthrough has ended at the close angle. Combine
1661 * it with the partial chunk if any. Call the passthrough
1662 * callback. Note that the open/close angles are
1663 * included in the text of the passthrough.
1665 GError
*tmp_error
= NULL
;
1667 advance_char (context
); /* advance past close angle */
1668 add_to_partial (context
, context
->start
, context
->iter
);
1670 if (context
->flags
& G_MARKUP_TREAT_CDATA_AS_TEXT
&&
1671 strncmp (context
->partial_chunk
->str
, "<![CDATA[", 9) == 0)
1673 if (context
->parser
->text
&&
1674 text_validate (context
,
1675 context
->partial_chunk
->str
+ 9,
1676 context
->partial_chunk
->len
- 12,
1678 (*context
->parser
->text
) (context
,
1679 context
->partial_chunk
->str
+ 9,
1680 context
->partial_chunk
->len
- 12,
1684 else if (context
->parser
->passthrough
&&
1685 text_validate (context
,
1686 context
->partial_chunk
->str
,
1687 context
->partial_chunk
->len
,
1689 (*context
->parser
->passthrough
) (context
,
1690 context
->partial_chunk
->str
,
1691 context
->partial_chunk
->len
,
1695 truncate_partial (context
);
1697 if (tmp_error
== NULL
)
1699 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1700 context
->start
= context
->iter
; /* could begin text */
1703 propagate_error (context
, error
, tmp_error
);
1712 g_assert_not_reached ();
1718 context
->parsing
= FALSE
;
1720 return context
->state
!= STATE_ERROR
;
1724 * g_markup_parse_context_end_parse:
1725 * @context: a #GMarkupParseContext
1726 * @error: return location for a #GError
1728 * Signals to the #GMarkupParseContext that all data has been
1729 * fed into the parse context with g_markup_parse_context_parse().
1731 * This function reports an error if the document isn't complete,
1732 * for example if elements are still open.
1734 * Return value: %TRUE on success, %FALSE if an error was set
1737 g_markup_parse_context_end_parse (GMarkupParseContext
*context
,
1740 g_return_val_if_fail (context
!= NULL
, FALSE
);
1741 g_return_val_if_fail (!context
->parsing
, FALSE
);
1742 g_return_val_if_fail (context
->state
!= STATE_ERROR
, FALSE
);
1744 if (context
->partial_chunk
!= NULL
)
1746 g_string_free (context
->partial_chunk
, TRUE
);
1747 context
->partial_chunk
= NULL
;
1750 if (context
->document_empty
)
1752 set_error_literal (context
, error
, G_MARKUP_ERROR_EMPTY
,
1753 _("Document was empty or contained only whitespace"));
1757 context
->parsing
= TRUE
;
1759 switch (context
->state
)
1765 case STATE_AFTER_OPEN_ANGLE
:
1766 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1767 _("Document ended unexpectedly just after an open angle bracket '<'"));
1770 case STATE_AFTER_CLOSE_ANGLE
:
1771 if (context
->tag_stack
!= NULL
)
1773 /* Error message the same as for INSIDE_TEXT */
1774 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1775 _("Document ended unexpectedly with elements still open - "
1776 "'%s' was the last element opened"),
1777 current_element (context
));
1781 case STATE_AFTER_ELISION_SLASH
:
1782 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1783 _("Document ended unexpectedly, expected to see a close angle "
1784 "bracket ending the tag <%s/>"), current_element (context
));
1787 case STATE_INSIDE_OPEN_TAG_NAME
:
1788 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1789 _("Document ended unexpectedly inside an element name"));
1792 case STATE_INSIDE_ATTRIBUTE_NAME
:
1793 case STATE_AFTER_ATTRIBUTE_NAME
:
1794 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1795 _("Document ended unexpectedly inside an attribute name"));
1798 case STATE_BETWEEN_ATTRIBUTES
:
1799 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1800 _("Document ended unexpectedly inside an element-opening "
1804 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
:
1805 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1806 _("Document ended unexpectedly after the equals sign "
1807 "following an attribute name; no attribute value"));
1810 case STATE_INSIDE_ATTRIBUTE_VALUE_SQ
:
1811 case STATE_INSIDE_ATTRIBUTE_VALUE_DQ
:
1812 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1813 _("Document ended unexpectedly while inside an attribute "
1817 case STATE_INSIDE_TEXT
:
1818 g_assert (context
->tag_stack
!= NULL
);
1819 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1820 _("Document ended unexpectedly with elements still open - "
1821 "'%s' was the last element opened"),
1822 current_element (context
));
1825 case STATE_AFTER_CLOSE_TAG_SLASH
:
1826 case STATE_INSIDE_CLOSE_TAG_NAME
:
1827 case STATE_AFTER_CLOSE_TAG_NAME
:
1828 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1829 _("Document ended unexpectedly inside the close tag for "
1830 "element '%s'"), current_element (context
));
1833 case STATE_INSIDE_PASSTHROUGH
:
1834 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1835 _("Document ended unexpectedly inside a comment or "
1836 "processing instruction"));
1841 g_assert_not_reached ();
1845 context
->parsing
= FALSE
;
1847 return context
->state
!= STATE_ERROR
;
1851 * g_markup_parse_context_get_element:
1852 * @context: a #GMarkupParseContext
1854 * Retrieves the name of the currently open element.
1856 * If called from the start_element or end_element handlers this will
1857 * give the element_name as passed to those functions. For the parent
1858 * elements, see g_markup_parse_context_get_element_stack().
1860 * Returns: the name of the currently open element, or %NULL
1865 g_markup_parse_context_get_element (GMarkupParseContext
*context
)
1867 g_return_val_if_fail (context
!= NULL
, NULL
);
1869 if (context
->tag_stack
== NULL
)
1872 return current_element (context
);
1876 * g_markup_parse_context_get_element_stack:
1877 * @context: a #GMarkupParseContext
1879 * Retrieves the element stack from the internal state of the parser.
1881 * The returned #GSList is a list of strings where the first item is
1882 * the currently open tag (as would be returned by
1883 * g_markup_parse_context_get_element()) and the next item is its
1886 * This function is intended to be used in the start_element and
1887 * end_element handlers where g_markup_parse_context_get_element()
1888 * would merely return the name of the element that is being
1891 * Returns: the element stack, which must not be modified
1896 g_markup_parse_context_get_element_stack (GMarkupParseContext
*context
)
1898 g_return_val_if_fail (context
!= NULL
, NULL
);
1899 return context
->tag_stack
;
1903 * g_markup_parse_context_get_position:
1904 * @context: a #GMarkupParseContext
1905 * @line_number: (allow-none): return location for a line number, or %NULL
1906 * @char_number: (allow-none): return location for a char-on-line number, or %NULL
1908 * Retrieves the current line number and the number of the character on
1909 * that line. Intended for use in error messages; there are no strict
1910 * semantics for what constitutes the "current" line number other than
1911 * "the best number we could come up with for error messages."
1914 g_markup_parse_context_get_position (GMarkupParseContext
*context
,
1918 g_return_if_fail (context
!= NULL
);
1921 *line_number
= context
->line_number
;
1924 *char_number
= context
->char_number
;
1928 * g_markup_parse_context_get_user_data:
1929 * @context: a #GMarkupParseContext
1931 * Returns the user_data associated with @context.
1933 * This will either be the user_data that was provided to
1934 * g_markup_parse_context_new() or to the most recent call
1935 * of g_markup_parse_context_push().
1937 * Returns: the provided user_data. The returned data belongs to
1938 * the markup context and will be freed when
1939 * g_markup_parse_context_free() is called.
1944 g_markup_parse_context_get_user_data (GMarkupParseContext
*context
)
1946 return context
->user_data
;
1950 * g_markup_parse_context_push:
1951 * @context: a #GMarkupParseContext
1952 * @parser: a #GMarkupParser
1953 * @user_data: user data to pass to #GMarkupParser functions
1955 * Temporarily redirects markup data to a sub-parser.
1957 * This function may only be called from the start_element handler of
1958 * a #GMarkupParser. It must be matched with a corresponding call to
1959 * g_markup_parse_context_pop() in the matching end_element handler
1960 * (except in the case that the parser aborts due to an error).
1962 * All tags, text and other data between the matching tags is
1963 * redirected to the subparser given by @parser. @user_data is used
1964 * as the user_data for that parser. @user_data is also passed to the
1965 * error callback in the event that an error occurs. This includes
1966 * errors that occur in subparsers of the subparser.
1968 * The end tag matching the start tag for which this call was made is
1969 * handled by the previous parser (which is given its own user_data)
1970 * which is why g_markup_parse_context_pop() is provided to allow "one
1971 * last access" to the @user_data provided to this function. In the
1972 * case of error, the @user_data provided here is passed directly to
1973 * the error callback of the subparser and g_markup_parse_context_pop()
1974 * should not be called. In either case, if @user_data was allocated
1975 * then it ought to be freed from both of these locations.
1977 * This function is not intended to be directly called by users
1978 * interested in invoking subparsers. Instead, it is intended to be
1979 * used by the subparsers themselves to implement a higher-level
1982 * As an example, see the following implementation of a simple
1983 * parser that counts the number of tags encountered.
1992 * counter_start_element (GMarkupParseContext *context,
1993 * const gchar *element_name,
1994 * const gchar **attribute_names,
1995 * const gchar **attribute_values,
1996 * gpointer user_data,
1999 * CounterData *data = user_data;
2001 * data->tag_count++;
2005 * counter_error (GMarkupParseContext *context,
2007 * gpointer user_data)
2009 * CounterData *data = user_data;
2011 * g_slice_free (CounterData, data);
2014 * static GMarkupParser counter_subparser =
2016 * counter_start_element,
2024 * In order to allow this parser to be easily used as a subparser, the
2025 * following interface is provided:
2029 * start_counting (GMarkupParseContext *context)
2031 * CounterData *data = g_slice_new (CounterData);
2033 * data->tag_count = 0;
2034 * g_markup_parse_context_push (context, &counter_subparser, data);
2038 * end_counting (GMarkupParseContext *context)
2040 * CounterData *data = g_markup_parse_context_pop (context);
2043 * result = data->tag_count;
2044 * g_slice_free (CounterData, data);
2050 * The subparser would then be used as follows:
2053 * static void start_element (context, element_name, ...)
2055 * if (strcmp (element_name, "count-these") == 0)
2056 * start_counting (context);
2058 * /* else, handle other tags... */
2061 * static void end_element (context, element_name, ...)
2063 * if (strcmp (element_name, "count-these") == 0)
2064 * g_print ("Counted %d tags\n", end_counting (context));
2066 * /* else, handle other tags... */
2073 g_markup_parse_context_push (GMarkupParseContext
*context
,
2074 const GMarkupParser
*parser
,
2077 GMarkupRecursionTracker
*tracker
;
2079 tracker
= g_slice_new (GMarkupRecursionTracker
);
2080 tracker
->prev_element
= context
->subparser_element
;
2081 tracker
->prev_parser
= context
->parser
;
2082 tracker
->prev_user_data
= context
->user_data
;
2084 context
->subparser_element
= current_element (context
);
2085 context
->parser
= parser
;
2086 context
->user_data
= user_data
;
2088 context
->subparser_stack
= g_slist_prepend (context
->subparser_stack
,
2093 * g_markup_parse_context_pop:
2094 * @context: a #GMarkupParseContext
2096 * Completes the process of a temporary sub-parser redirection.
2098 * This function exists to collect the user_data allocated by a
2099 * matching call to g_markup_parse_context_push(). It must be called
2100 * in the end_element handler corresponding to the start_element
2101 * handler during which g_markup_parse_context_push() was called.
2102 * You must not call this function from the error callback -- the
2103 * @user_data is provided directly to the callback in that case.
2105 * This function is not intended to be directly called by users
2106 * interested in invoking subparsers. Instead, it is intended to
2107 * be used by the subparsers themselves to implement a higher-level
2110 * Returns: the user data passed to g_markup_parse_context_push()
2115 g_markup_parse_context_pop (GMarkupParseContext
*context
)
2119 if (!context
->awaiting_pop
)
2120 possibly_finish_subparser (context
);
2122 g_assert (context
->awaiting_pop
);
2124 context
->awaiting_pop
= FALSE
;
2126 /* valgrind friendliness */
2127 user_data
= context
->held_user_data
;
2128 context
->held_user_data
= NULL
;
2134 append_escaped_text (GString
*str
,
2143 end
= text
+ length
;
2148 next
= g_utf8_next_char (p
);
2153 g_string_append (str
, "&");
2157 g_string_append (str
, "<");
2161 g_string_append (str
, ">");
2165 g_string_append (str
, "'");
2169 g_string_append (str
, """);
2173 c
= g_utf8_get_char (p
);
2174 if ((0x1 <= c
&& c
<= 0x8) ||
2175 (0xb <= c
&& c
<= 0xc) ||
2176 (0xe <= c
&& c
<= 0x1f) ||
2177 (0x7f <= c
&& c
<= 0x84) ||
2178 (0x86 <= c
&& c
<= 0x9f))
2179 g_string_append_printf (str
, "&#x%x;", c
);
2181 g_string_append_len (str
, p
, next
- p
);
2190 * g_markup_escape_text:
2191 * @text: some valid UTF-8 text
2192 * @length: length of @text in bytes, or -1 if the text is nul-terminated
2194 * Escapes text so that the markup parser will parse it verbatim.
2195 * Less than, greater than, ampersand, etc. are replaced with the
2196 * corresponding entities. This function would typically be used
2197 * when writing out a file to be parsed with the markup parser.
2199 * Note that this function doesn't protect whitespace and line endings
2200 * from being processed according to the XML rules for normalization
2201 * of line endings and attribute values.
2203 * Note also that this function will produce character references in
2204 * the range of &#x1; ... &#x1f; for all control sequences
2205 * except for tabstop, newline and carriage return. The character
2206 * references in this range are not valid XML 1.0, but they are
2207 * valid XML 1.1 and will be accepted by the GMarkup parser.
2209 * Return value: a newly allocated string with the escaped text
2212 g_markup_escape_text (const gchar
*text
,
2217 g_return_val_if_fail (text
!= NULL
, NULL
);
2220 length
= strlen (text
);
2222 /* prealloc at least as long as original text */
2223 str
= g_string_sized_new (length
);
2224 append_escaped_text (str
, text
, length
);
2226 return g_string_free (str
, FALSE
);
2231 * @format: a printf-style format string
2232 * @after: location to store a pointer to the character after
2233 * the returned conversion. On a %NULL return, returns the
2234 * pointer to the trailing NUL in the string
2236 * Find the next conversion in a printf-style format string.
2237 * Partially based on code from printf-parser.c,
2238 * Copyright (C) 1999-2000, 2002-2003 Free Software Foundation, Inc.
2240 * Return value: pointer to the next conversion in @format,
2241 * or %NULL, if none.
2244 find_conversion (const char *format
,
2247 const char *start
= format
;
2250 while (*start
!= '\0' && *start
!= '%')
2267 /* Test for positional argument. */
2268 if (*cp
>= '0' && *cp
<= '9')
2272 for (np
= cp
; *np
>= '0' && *np
<= '9'; np
++)
2278 /* Skip the flags. */
2292 /* Skip the field width. */
2297 /* Test for positional argument. */
2298 if (*cp
>= '0' && *cp
<= '9')
2302 for (np
= cp
; *np
>= '0' && *np
<= '9'; np
++)
2310 for (; *cp
>= '0' && *cp
<= '9'; cp
++)
2314 /* Skip the precision. */
2320 /* Test for positional argument. */
2321 if (*cp
>= '0' && *cp
<= '9')
2325 for (np
= cp
; *np
>= '0' && *np
<= '9'; np
++)
2333 for (; *cp
>= '0' && *cp
<= '9'; cp
++)
2338 /* Skip argument type/size specifiers. */
2339 while (*cp
== 'h' ||
2348 /* Skip the conversion character. */
2356 * g_markup_vprintf_escaped:
2357 * @format: printf() style format string
2358 * @args: variable argument list, similar to vprintf()
2360 * Formats the data in @args according to @format, escaping
2361 * all string and character arguments in the fashion
2362 * of g_markup_escape_text(). See g_markup_printf_escaped().
2364 * Return value: newly allocated result from formatting
2365 * operation. Free with g_free().
2370 g_markup_vprintf_escaped (const gchar
*format
,
2375 GString
*result
= NULL
;
2376 gchar
*output1
= NULL
;
2377 gchar
*output2
= NULL
;
2378 const char *p
, *op1
, *op2
;
2381 /* The technique here, is that we make two format strings that
2382 * have the identical conversions in the identical order to the
2383 * original strings, but differ in the text in-between. We
2384 * then use the normal g_strdup_vprintf() to format the arguments
2385 * with the two new format strings. By comparing the results,
2386 * we can figure out what segments of the output come from
2387 * the original format string, and what from the arguments,
2388 * and thus know what portions of the string to escape.
2390 * For instance, for:
2392 * g_markup_printf_escaped ("%s ate %d apples", "Susan & Fred", 5);
2394 * We form the two format strings "%sX%dX" and %sY%sY". The results
2395 * of formatting with those two strings are
2397 * "%sX%dX" => "Susan & FredX5X"
2398 * "%sY%dY" => "Susan & FredY5Y"
2400 * To find the span of the first argument, we find the first position
2401 * where the two arguments differ, which tells us that the first
2402 * argument formatted to "Susan & Fred". We then escape that
2403 * to "Susan & Fred" and join up with the intermediate portions
2404 * of the format string and the second argument to get
2405 * "Susan & Fred ate 5 apples".
2408 /* Create the two modified format strings
2410 format1
= g_string_new (NULL
);
2411 format2
= g_string_new (NULL
);
2416 const char *conv
= find_conversion (p
, &after
);
2420 g_string_append_len (format1
, conv
, after
- conv
);
2421 g_string_append_c (format1
, 'X');
2422 g_string_append_len (format2
, conv
, after
- conv
);
2423 g_string_append_c (format2
, 'Y');
2428 /* Use them to format the arguments
2430 G_VA_COPY (args2
, args
);
2432 output1
= g_strdup_vprintf (format1
->str
, args
);
2439 output2
= g_strdup_vprintf (format2
->str
, args2
);
2444 result
= g_string_new (NULL
);
2446 /* Iterate through the original format string again,
2447 * copying the non-conversion portions and the escaped
2448 * converted arguments to the output string.
2456 const char *output_start
;
2457 const char *conv
= find_conversion (p
, &after
);
2460 if (!conv
) /* The end, after points to the trailing \0 */
2462 g_string_append_len (result
, p
, after
- p
);
2466 g_string_append_len (result
, p
, conv
- p
);
2468 while (*op1
== *op2
)
2474 escaped
= g_markup_escape_text (output_start
, op1
- output_start
);
2475 g_string_append (result
, escaped
);
2484 g_string_free (format1
, TRUE
);
2485 g_string_free (format2
, TRUE
);
2490 return g_string_free (result
, FALSE
);
2496 * g_markup_printf_escaped:
2497 * @format: printf() style format string
2498 * @...: the arguments to insert in the format string
2500 * Formats arguments according to @format, escaping
2501 * all string and character arguments in the fashion
2502 * of g_markup_escape_text(). This is useful when you
2503 * want to insert literal strings into XML-style markup
2504 * output, without having to worry that the strings
2505 * might themselves contain markup.
2508 * const char *store = "Fortnum & Mason";
2509 * const char *item = "Tea";
2512 * output = g_markup_printf_escaped ("<purchase>"
2513 * "<store>%s</store>"
2514 * "<item>%s</item>"
2515 * "</purchase>",
2519 * Return value: newly allocated result from formatting
2520 * operation. Free with g_free().
2525 g_markup_printf_escaped (const gchar
*format
, ...)
2530 va_start (args
, format
);
2531 result
= g_markup_vprintf_escaped (format
, args
);
2538 g_markup_parse_boolean (const char *string
,
2541 char const * const falses
[] = { "false", "f", "no", "n", "0" };
2542 char const * const trues
[] = { "true", "t", "yes", "y", "1" };
2545 for (i
= 0; i
< G_N_ELEMENTS (falses
); i
++)
2547 if (g_ascii_strcasecmp (string
, falses
[i
]) == 0)
2556 for (i
= 0; i
< G_N_ELEMENTS (trues
); i
++)
2558 if (g_ascii_strcasecmp (string
, trues
[i
]) == 0)
2571 * GMarkupCollectType:
2572 * @G_MARKUP_COLLECT_INVALID: used to terminate the list of attributes
2574 * @G_MARKUP_COLLECT_STRING: collect the string pointer directly from
2575 * the attribute_values[] array. Expects a parameter of type (const
2576 * char **). If %G_MARKUP_COLLECT_OPTIONAL is specified and the
2577 * attribute isn't present then the pointer will be set to %NULL
2578 * @G_MARKUP_COLLECT_STRDUP: as with %G_MARKUP_COLLECT_STRING, but
2579 * expects a parameter of type (char **) and g_strdup()s the
2580 * returned pointer. The pointer must be freed with g_free()
2581 * @G_MARKUP_COLLECT_BOOLEAN: expects a parameter of type (gboolean *)
2582 * and parses the attribute value as a boolean. Sets %FALSE if the
2583 * attribute isn't present. Valid boolean values consist of
2584 * (case-insensitive) "false", "f", "no", "n", "0" and "true", "t",
2586 * @G_MARKUP_COLLECT_TRISTATE: as with %G_MARKUP_COLLECT_BOOLEAN, but
2587 * in the case of a missing attribute a value is set that compares
2588 * equal to neither %FALSE nor %TRUE G_MARKUP_COLLECT_OPTIONAL is
2590 * @G_MARKUP_COLLECT_OPTIONAL: can be bitwise ORed with the other fields.
2591 * If present, allows the attribute not to appear. A default value
2592 * is set depending on what value type is used
2594 * A mixed enumerated type and flags field. You must specify one type
2595 * (string, strdup, boolean, tristate). Additionally, you may optionally
2596 * bitwise OR the type with the flag %G_MARKUP_COLLECT_OPTIONAL.
2598 * It is likely that this enum will be extended in the future to
2599 * support other types.
2603 * g_markup_collect_attributes:
2604 * @element_name: the current tag name
2605 * @attribute_names: the attribute names
2606 * @attribute_values: the attribute values
2607 * @error: a pointer to a #GError or %NULL
2608 * @first_type: the #GMarkupCollectType of the first attribute
2609 * @first_attr: the name of the first attribute
2610 * @...: a pointer to the storage location of the first attribute
2611 * (or %NULL), followed by more types names and pointers, ending
2612 * with %G_MARKUP_COLLECT_INVALID
2614 * Collects the attributes of the element from the data passed to the
2615 * #GMarkupParser start_element function, dealing with common error
2616 * conditions and supporting boolean values.
2618 * This utility function is not required to write a parser but can save
2621 * The @element_name, @attribute_names, @attribute_values and @error
2622 * parameters passed to the start_element callback should be passed
2623 * unmodified to this function.
2625 * Following these arguments is a list of "supported" attributes to collect.
2626 * It is an error to specify multiple attributes with the same name. If any
2627 * attribute not in the list appears in the @attribute_names array then an
2628 * unknown attribute error will result.
2630 * The #GMarkupCollectType field allows specifying the type of collection
2631 * to perform and if a given attribute must appear or is optional.
2633 * The attribute name is simply the name of the attribute to collect.
2635 * The pointer should be of the appropriate type (see the descriptions
2636 * under #GMarkupCollectType) and may be %NULL in case a particular
2637 * attribute is to be allowed but ignored.
2639 * This function deals with issuing errors for missing attributes
2640 * (of type %G_MARKUP_ERROR_MISSING_ATTRIBUTE), unknown attributes
2641 * (of type %G_MARKUP_ERROR_UNKNOWN_ATTRIBUTE) and duplicate
2642 * attributes (of type %G_MARKUP_ERROR_INVALID_CONTENT) as well
2643 * as parse errors for boolean-valued attributes (again of type
2644 * %G_MARKUP_ERROR_INVALID_CONTENT). In all of these cases %FALSE
2645 * will be returned and @error will be set as appropriate.
2647 * Return value: %TRUE if successful
2652 g_markup_collect_attributes (const gchar
*element_name
,
2653 const gchar
**attribute_names
,
2654 const gchar
**attribute_values
,
2656 GMarkupCollectType first_type
,
2657 const gchar
*first_attr
,
2660 GMarkupCollectType type
;
2672 va_start (ap
, first_attr
);
2673 while (type
!= G_MARKUP_COLLECT_INVALID
)
2678 mandatory
= !(type
& G_MARKUP_COLLECT_OPTIONAL
);
2679 type
&= (G_MARKUP_COLLECT_OPTIONAL
- 1);
2681 /* tristate records a value != TRUE and != FALSE
2682 * for the case where the attribute is missing
2684 if (type
== G_MARKUP_COLLECT_TRISTATE
)
2687 for (i
= 0; attribute_names
[i
]; i
++)
2688 if (i
>= 40 || !(collected
& (G_GUINT64_CONSTANT(1) << i
)))
2689 if (!strcmp (attribute_names
[i
], attr
))
2692 /* ISO C99 only promises that the user can pass up to 127 arguments.
2693 * Subtracting the first 4 arguments plus the final NULL and dividing
2694 * by 3 arguments per collected attribute, we are left with a maximum
2695 * number of supported attributes of (127 - 5) / 3 = 40.
2697 * In reality, nobody is ever going to call us with anywhere close to
2698 * 40 attributes to collect, so it is safe to assume that if i > 40
2699 * then the user has given some invalid or repeated arguments. These
2700 * problems will be caught and reported at the end of the function.
2702 * We know at this point that we have an error, but we don't know
2703 * what error it is, so just continue...
2706 collected
|= (G_GUINT64_CONSTANT(1) << i
);
2708 value
= attribute_values
[i
];
2710 if (value
== NULL
&& mandatory
)
2712 g_set_error (error
, G_MARKUP_ERROR
,
2713 G_MARKUP_ERROR_MISSING_ATTRIBUTE
,
2714 "element '%s' requires attribute '%s'",
2715 element_name
, attr
);
2723 case G_MARKUP_COLLECT_STRING
:
2725 const char **str_ptr
;
2727 str_ptr
= va_arg (ap
, const char **);
2729 if (str_ptr
!= NULL
)
2734 case G_MARKUP_COLLECT_STRDUP
:
2738 str_ptr
= va_arg (ap
, char **);
2740 if (str_ptr
!= NULL
)
2741 *str_ptr
= g_strdup (value
);
2745 case G_MARKUP_COLLECT_BOOLEAN
:
2746 case G_MARKUP_COLLECT_TRISTATE
:
2751 bool_ptr
= va_arg (ap
, gboolean
*);
2753 if (bool_ptr
!= NULL
)
2755 if (type
== G_MARKUP_COLLECT_TRISTATE
)
2756 /* constructivists rejoice!
2757 * neither false nor true...
2761 else /* G_MARKUP_COLLECT_BOOLEAN */
2767 if (!g_markup_parse_boolean (value
, va_arg (ap
, gboolean
*)))
2769 g_set_error (error
, G_MARKUP_ERROR
,
2770 G_MARKUP_ERROR_INVALID_CONTENT
,
2771 "element '%s', attribute '%s', value '%s' "
2772 "cannot be parsed as a boolean value",
2773 element_name
, attr
, value
);
2783 g_assert_not_reached ();
2786 type
= va_arg (ap
, GMarkupCollectType
);
2787 attr
= va_arg (ap
, const char *);
2792 /* ensure we collected all the arguments */
2793 for (i
= 0; attribute_names
[i
]; i
++)
2794 if ((collected
& (G_GUINT64_CONSTANT(1) << i
)) == 0)
2796 /* attribute not collected: could be caused by two things.
2798 * 1) it doesn't exist in our list of attributes
2799 * 2) it existed but was matched by a duplicate attribute earlier
2805 for (j
= 0; j
< i
; j
++)
2806 if (strcmp (attribute_names
[i
], attribute_names
[j
]) == 0)
2810 /* j is now the first occurrence of attribute_names[i] */
2812 g_set_error (error
, G_MARKUP_ERROR
,
2813 G_MARKUP_ERROR_UNKNOWN_ATTRIBUTE
,
2814 "attribute '%s' invalid for element '%s'",
2815 attribute_names
[i
], element_name
);
2817 g_set_error (error
, G_MARKUP_ERROR
,
2818 G_MARKUP_ERROR_INVALID_CONTENT
,
2819 "attribute '%s' given multiple times for element '%s'",
2820 attribute_names
[i
], element_name
);
2828 /* replay the above to free allocations */
2832 va_start (ap
, first_attr
);
2833 while (type
!= G_MARKUP_COLLECT_INVALID
)
2837 ptr
= va_arg (ap
, gpointer
);
2841 switch (type
& (G_MARKUP_COLLECT_OPTIONAL
- 1))
2843 case G_MARKUP_COLLECT_STRDUP
:
2845 g_free (*(char **) ptr
);
2847 case G_MARKUP_COLLECT_STRING
:
2848 *(char **) ptr
= NULL
;
2851 case G_MARKUP_COLLECT_BOOLEAN
:
2852 *(gboolean
*) ptr
= FALSE
;
2855 case G_MARKUP_COLLECT_TRISTATE
:
2856 *(gboolean
*) ptr
= -1;
2861 type
= va_arg (ap
, GMarkupCollectType
);
2862 attr
= va_arg (ap
, const char *);