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.
34 #include "gstrfuncs.h"
36 #include "gtestutils.h"
42 * @Title: Simple XML Subset Parser
43 * @Short_description: parses a subset of XML
44 * @See_also: <ulink url="http://www.w3.org/TR/REC-xml/">XML
45 * Specification</ulink>
47 * The "GMarkup" parser is intended to parse a simple markup format
48 * that's a subset of XML. This is a small, efficient, easy-to-use
49 * parser. It should not be used if you expect to interoperate with
50 * other applications generating full-scale XML. However, it's very
51 * useful for application data files, config files, etc. where you
52 * know your application will be the only one writing the file.
53 * Full-scale XML parsers should be able to parse the subset used by
54 * GMarkup, so you can easily migrate to full-scale XML at a later
55 * time if the need arises.
57 * GMarkup is not guaranteed to signal an error on all invalid XML;
58 * the parser may accept documents that an XML parser would not.
59 * However, XML documents which are not well-formed<footnote
60 * id="wellformed">Being wellformed is a weaker condition than being
61 * valid. See the <ulink url="http://www.w3.org/TR/REC-xml/">XML
62 * specification</ulink> for definitions of these terms.</footnote>
63 * are not considered valid GMarkup documents.
65 * Simplifications to XML include:
67 * <listitem>Only UTF-8 encoding is allowed</listitem>
68 * <listitem>No user-defined entities</listitem>
69 * <listitem>Processing instructions, comments and the doctype declaration
70 * are "passed through" but are not interpreted in any way</listitem>
71 * <listitem>No DTD or validation.</listitem>
74 * The markup format does support:
76 * <listitem>Elements</listitem>
77 * <listitem>Attributes</listitem>
78 * <listitem>5 standard entities:
79 * <literal>&amp; &lt; &gt; &quot; &apos;</literal>
81 * <listitem>Character references</listitem>
82 * <listitem>Sections marked as CDATA</listitem>
86 G_DEFINE_QUARK (g
-markup
-error
-quark
, g_markup_error
)
91 STATE_AFTER_OPEN_ANGLE
,
92 STATE_AFTER_CLOSE_ANGLE
,
93 STATE_AFTER_ELISION_SLASH
, /* the slash that obviates need for end element */
94 STATE_INSIDE_OPEN_TAG_NAME
,
95 STATE_INSIDE_ATTRIBUTE_NAME
,
96 STATE_AFTER_ATTRIBUTE_NAME
,
97 STATE_BETWEEN_ATTRIBUTES
,
98 STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
,
99 STATE_INSIDE_ATTRIBUTE_VALUE_SQ
,
100 STATE_INSIDE_ATTRIBUTE_VALUE_DQ
,
102 STATE_AFTER_CLOSE_TAG_SLASH
,
103 STATE_INSIDE_CLOSE_TAG_NAME
,
104 STATE_AFTER_CLOSE_TAG_NAME
,
105 STATE_INSIDE_PASSTHROUGH
,
111 const char *prev_element
;
112 const GMarkupParser
*prev_parser
;
113 gpointer prev_user_data
;
114 } GMarkupRecursionTracker
;
116 struct _GMarkupParseContext
118 const GMarkupParser
*parser
;
120 GMarkupParseFlags flags
;
125 GMarkupParseState state
;
128 GDestroyNotify dnotify
;
130 /* A piece of character data or an element that
131 * hasn't "ended" yet so we haven't yet called
132 * the callback for it.
134 GString
*partial_chunk
;
135 GSList
*spare_chunks
;
138 GSList
*tag_stack_gstr
;
139 GSList
*spare_list_nodes
;
141 GString
**attr_names
;
142 GString
**attr_values
;
146 const gchar
*current_text
;
147 gssize current_text_len
;
148 const gchar
*current_text_end
;
150 /* used to save the start of the last interesting thingy */
155 guint document_empty
: 1;
157 guint awaiting_pop
: 1;
160 /* subparser support */
161 GSList
*subparser_stack
; /* (GMarkupRecursionTracker *) */
162 const char *subparser_element
;
163 gpointer held_user_data
;
167 * Helpers to reduce our allocation overhead, we have
168 * a well defined allocation lifecycle.
171 get_list_node (GMarkupParseContext
*context
, gpointer data
)
174 if (context
->spare_list_nodes
!= NULL
)
176 node
= context
->spare_list_nodes
;
177 context
->spare_list_nodes
= g_slist_remove_link (context
->spare_list_nodes
, node
);
180 node
= g_slist_alloc();
186 free_list_node (GMarkupParseContext
*context
, GSList
*node
)
189 context
->spare_list_nodes
= g_slist_concat (node
, context
->spare_list_nodes
);
193 string_blank (GString
*string
)
195 string
->str
[0] = '\0';
200 * g_markup_parse_context_new:
201 * @parser: a #GMarkupParser
202 * @flags: one or more #GMarkupParseFlags
203 * @user_data: user data to pass to #GMarkupParser functions
204 * @user_data_dnotify: user data destroy notifier called when
205 * the parse context is freed
207 * Creates a new parse context. A parse context is used to parse
208 * marked-up documents. You can feed any number of documents into
209 * a context, as long as no errors occur; once an error occurs,
210 * the parse context can't continue to parse text (you have to
211 * free it and create a new parse context).
213 * Return value: a new #GMarkupParseContext
215 GMarkupParseContext
*
216 g_markup_parse_context_new (const GMarkupParser
*parser
,
217 GMarkupParseFlags flags
,
219 GDestroyNotify user_data_dnotify
)
221 GMarkupParseContext
*context
;
223 g_return_val_if_fail (parser
!= NULL
, NULL
);
225 context
= g_new (GMarkupParseContext
, 1);
227 context
->parser
= parser
;
228 context
->flags
= flags
;
229 context
->user_data
= user_data
;
230 context
->dnotify
= user_data_dnotify
;
232 context
->line_number
= 1;
233 context
->char_number
= 1;
235 context
->partial_chunk
= NULL
;
236 context
->spare_chunks
= NULL
;
237 context
->spare_list_nodes
= NULL
;
239 context
->state
= STATE_START
;
240 context
->tag_stack
= NULL
;
241 context
->tag_stack_gstr
= NULL
;
242 context
->attr_names
= NULL
;
243 context
->attr_values
= NULL
;
244 context
->cur_attr
= -1;
245 context
->alloc_attrs
= 0;
247 context
->current_text
= NULL
;
248 context
->current_text_len
= -1;
249 context
->current_text_end
= NULL
;
251 context
->start
= NULL
;
252 context
->iter
= NULL
;
254 context
->document_empty
= TRUE
;
255 context
->parsing
= FALSE
;
257 context
->awaiting_pop
= FALSE
;
258 context
->subparser_stack
= NULL
;
259 context
->subparser_element
= NULL
;
261 /* this is only looked at if awaiting_pop = TRUE. initialise anyway. */
262 context
->held_user_data
= NULL
;
264 context
->balance
= 0;
270 string_full_free (gpointer ptr
)
272 g_string_free (ptr
, TRUE
);
275 static void clear_attributes (GMarkupParseContext
*context
);
278 * g_markup_parse_context_free:
279 * @context: a #GMarkupParseContext
281 * Frees a #GMarkupParseContext.
283 * This function can't be called from inside one of the
284 * #GMarkupParser functions or while a subparser is pushed.
287 g_markup_parse_context_free (GMarkupParseContext
*context
)
289 g_return_if_fail (context
!= NULL
);
290 g_return_if_fail (!context
->parsing
);
291 g_return_if_fail (!context
->subparser_stack
);
292 g_return_if_fail (!context
->awaiting_pop
);
294 if (context
->dnotify
)
295 (* context
->dnotify
) (context
->user_data
);
297 clear_attributes (context
);
298 g_free (context
->attr_names
);
299 g_free (context
->attr_values
);
301 g_slist_free_full (context
->tag_stack_gstr
, string_full_free
);
302 g_slist_free (context
->tag_stack
);
304 g_slist_free_full (context
->spare_chunks
, string_full_free
);
305 g_slist_free (context
->spare_list_nodes
);
307 if (context
->partial_chunk
)
308 g_string_free (context
->partial_chunk
, TRUE
);
313 static void pop_subparser_stack (GMarkupParseContext
*context
);
316 mark_error (GMarkupParseContext
*context
,
319 context
->state
= STATE_ERROR
;
321 if (context
->parser
->error
)
322 (*context
->parser
->error
) (context
, error
, context
->user_data
);
324 /* report the error all the way up to free all the user-data */
325 while (context
->subparser_stack
)
327 pop_subparser_stack (context
);
328 context
->awaiting_pop
= FALSE
; /* already been freed */
330 if (context
->parser
->error
)
331 (*context
->parser
->error
) (context
, error
, context
->user_data
);
336 set_error (GMarkupParseContext
*context
,
340 ...) G_GNUC_PRINTF (4, 5);
343 set_error_literal (GMarkupParseContext
*context
,
346 const gchar
*message
)
350 tmp_error
= g_error_new_literal (G_MARKUP_ERROR
, code
, message
);
352 g_prefix_error (&tmp_error
,
353 _("Error on line %d char %d: "),
354 context
->line_number
,
355 context
->char_number
);
357 mark_error (context
, tmp_error
);
359 g_propagate_error (error
, tmp_error
);
363 set_error (GMarkupParseContext
*context
,
373 va_start (args
, format
);
374 s
= g_strdup_vprintf (format
, args
);
377 /* Make sure that the GError message is valid UTF-8
378 * even if it is complaining about invalid UTF-8 in the markup
380 s_valid
= _g_utf8_make_valid (s
);
381 set_error_literal (context
, error
, code
, s
);
388 propagate_error (GMarkupParseContext
*context
,
392 if (context
->flags
& G_MARKUP_PREFIX_ERROR_POSITION
)
393 g_prefix_error (&src
,
394 _("Error on line %d char %d: "),
395 context
->line_number
,
396 context
->char_number
);
398 mark_error (context
, src
);
400 g_propagate_error (dest
, src
);
403 #define IS_COMMON_NAME_END_CHAR(c) \
404 ((c) == '=' || (c) == '/' || (c) == '>' || (c) == ' ')
407 slow_name_validate (GMarkupParseContext
*context
,
411 const gchar
*p
= name
;
413 if (!g_utf8_validate (name
, strlen (name
), NULL
))
415 set_error (context
, error
, G_MARKUP_ERROR_BAD_UTF8
,
416 _("Invalid UTF-8 encoded text in name - not valid '%s'"), name
);
420 if (!(g_ascii_isalpha (*p
) ||
421 (!IS_COMMON_NAME_END_CHAR (*p
) &&
424 g_unichar_isalpha (g_utf8_get_char (p
))))))
426 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
427 _("'%s' is not a valid name"), name
);
431 for (p
= g_utf8_next_char (name
); *p
!= '\0'; p
= g_utf8_next_char (p
))
434 if (!(g_ascii_isalnum (*p
) ||
435 (!IS_COMMON_NAME_END_CHAR (*p
) &&
440 g_unichar_isalpha (g_utf8_get_char (p
))))))
442 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
443 _("'%s' is not a valid name: '%c'"), name
, *p
);
451 * Use me for elements, attributes etc.
454 name_validate (GMarkupParseContext
*context
,
461 /* name start char */
463 if (G_UNLIKELY (IS_COMMON_NAME_END_CHAR (*p
) ||
464 !(g_ascii_isalpha (*p
) || *p
== '_' || *p
== ':')))
467 for (mask
= *p
++; *p
!= '\0'; p
++)
472 if (G_UNLIKELY (!(g_ascii_isalnum (*p
) ||
473 (!IS_COMMON_NAME_END_CHAR (*p
) &&
481 if (mask
& 0x80) /* un-common / non-ascii */
487 return slow_name_validate (context
, name
, error
);
491 text_validate (GMarkupParseContext
*context
,
496 if (!g_utf8_validate (p
, len
, NULL
))
498 set_error (context
, error
, G_MARKUP_ERROR_BAD_UTF8
,
499 _("Invalid UTF-8 encoded text in name - not valid '%s'"), p
);
507 char_str (gunichar c
,
511 g_unichar_to_utf8 (c
, buf
);
516 utf8_str (const gchar
*utf8
,
519 char_str (g_utf8_get_char (utf8
), buf
);
524 set_unescape_error (GMarkupParseContext
*context
,
526 const gchar
*remaining_text
,
534 gint remaining_newlines
;
537 remaining_newlines
= 0;
542 ++remaining_newlines
;
546 va_start (args
, format
);
547 s
= g_strdup_vprintf (format
, args
);
550 tmp_error
= g_error_new (G_MARKUP_ERROR
,
552 _("Error on line %d: %s"),
553 context
->line_number
- remaining_newlines
,
558 mark_error (context
, tmp_error
);
560 g_propagate_error (error
, tmp_error
);
564 * re-write the GString in-place, unescaping anything that escaped.
565 * most XML does not contain entities, or escaping.
568 unescape_gstring_inplace (GMarkupParseContext
*context
,
576 gboolean normalize_attribute
;
580 /* are we unescaping an attribute or not ? */
581 if (context
->state
== STATE_INSIDE_ATTRIBUTE_VALUE_SQ
||
582 context
->state
== STATE_INSIDE_ATTRIBUTE_VALUE_DQ
)
583 normalize_attribute
= TRUE
;
585 normalize_attribute
= FALSE
;
588 * Meeks' theorum: unescaping can only shrink text.
589 * for < etc. this is obvious, for  more
590 * thought is required, but this is patently so.
593 for (from
= to
= string
->str
; *from
!= '\0'; from
++, to
++)
600 if (normalize_attribute
&& (*to
== '\t' || *to
== '\n'))
604 *to
= normalize_attribute
? ' ' : '\n';
613 gboolean is_hex
= FALSE
;
625 /* digit is between start and p */
628 l
= strtoul (from
, &end
, 16);
630 l
= strtoul (from
, &end
, 10);
632 if (end
== from
|| errno
!= 0)
634 set_unescape_error (context
, error
,
635 from
, G_MARKUP_ERROR_PARSE
,
636 _("Failed to parse '%-.*s', which "
637 "should have been a digit "
638 "inside a character reference "
639 "(ê for example) - perhaps "
640 "the digit is too large"),
644 else if (*end
!= ';')
646 set_unescape_error (context
, error
,
647 from
, G_MARKUP_ERROR_PARSE
,
648 _("Character reference did not end with a "
650 "most likely you used an ampersand "
651 "character without intending to start "
652 "an entity - escape ampersand as &"));
657 /* characters XML 1.1 permits */
658 if ((0 < l
&& l
<= 0xD7FF) ||
659 (0xE000 <= l
&& l
<= 0xFFFD) ||
660 (0x10000 <= l
&& l
<= 0x10FFFF))
665 to
+= strlen (buf
) - 1;
667 if (l
>= 0x80) /* not ascii */
672 set_unescape_error (context
, error
,
673 from
, G_MARKUP_ERROR_PARSE
,
674 _("Character reference '%-.*s' does not "
675 "encode a permitted character"),
682 else if (strncmp (from
, "lt;", 3) == 0)
687 else if (strncmp (from
, "gt;", 3) == 0)
692 else if (strncmp (from
, "amp;", 4) == 0)
697 else if (strncmp (from
, "quot;", 5) == 0)
702 else if (strncmp (from
, "apos;", 5) == 0)
710 set_unescape_error (context
, error
,
711 from
, G_MARKUP_ERROR_PARSE
,
712 _("Empty entity '&;' seen; valid "
713 "entities are: & " < > '"));
716 const char *end
= strchr (from
, ';');
718 set_unescape_error (context
, error
,
719 from
, G_MARKUP_ERROR_PARSE
,
720 _("Entity name '%-.*s' is not known"),
723 set_unescape_error (context
, error
,
724 from
, G_MARKUP_ERROR_PARSE
,
725 _("Entity did not end with a semicolon; "
726 "most likely you used an ampersand "
727 "character without intending to start "
728 "an entity - escape ampersand as &"));
735 g_assert (to
- string
->str
<= string
->len
);
736 if (to
- string
->str
!= string
->len
)
737 g_string_truncate (string
, to
- string
->str
);
739 *is_ascii
= !(mask
& 0x80);
744 static inline gboolean
745 advance_char (GMarkupParseContext
*context
)
748 context
->char_number
++;
750 if (G_UNLIKELY (context
->iter
== context
->current_text_end
))
753 else if (G_UNLIKELY (*context
->iter
== '\n'))
755 context
->line_number
++;
756 context
->char_number
= 1;
762 static inline gboolean
765 return c
== ' ' || c
== '\t' || c
== '\n' || c
== '\r';
769 skip_spaces (GMarkupParseContext
*context
)
773 if (!xml_isspace (*context
->iter
))
776 while (advance_char (context
));
780 advance_to_name_end (GMarkupParseContext
*context
)
784 if (IS_COMMON_NAME_END_CHAR (*(context
->iter
)))
786 if (xml_isspace (*(context
->iter
)))
789 while (advance_char (context
));
793 release_chunk (GMarkupParseContext
*context
, GString
*str
)
798 if (str
->allocated_len
> 256)
799 { /* large strings are unusual and worth freeing */
800 g_string_free (str
, TRUE
);
804 node
= get_list_node (context
, str
);
805 context
->spare_chunks
= g_slist_concat (node
, context
->spare_chunks
);
809 add_to_partial (GMarkupParseContext
*context
,
810 const gchar
*text_start
,
811 const gchar
*text_end
)
813 if (context
->partial_chunk
== NULL
)
814 { /* allocate a new chunk to parse into */
816 if (context
->spare_chunks
!= NULL
)
818 GSList
*node
= context
->spare_chunks
;
819 context
->spare_chunks
= g_slist_remove_link (context
->spare_chunks
, node
);
820 context
->partial_chunk
= node
->data
;
821 free_list_node (context
, node
);
824 context
->partial_chunk
= g_string_sized_new (MAX (28, text_end
- text_start
));
827 if (text_start
!= text_end
)
828 g_string_insert_len (context
->partial_chunk
, -1,
829 text_start
, text_end
- text_start
);
833 truncate_partial (GMarkupParseContext
*context
)
835 if (context
->partial_chunk
!= NULL
)
836 string_blank (context
->partial_chunk
);
839 static inline const gchar
*
840 current_element (GMarkupParseContext
*context
)
842 return context
->tag_stack
->data
;
846 pop_subparser_stack (GMarkupParseContext
*context
)
848 GMarkupRecursionTracker
*tracker
;
850 g_assert (context
->subparser_stack
);
852 tracker
= context
->subparser_stack
->data
;
854 context
->awaiting_pop
= TRUE
;
855 context
->held_user_data
= context
->user_data
;
857 context
->user_data
= tracker
->prev_user_data
;
858 context
->parser
= tracker
->prev_parser
;
859 context
->subparser_element
= tracker
->prev_element
;
860 g_slice_free (GMarkupRecursionTracker
, tracker
);
862 context
->subparser_stack
= g_slist_delete_link (context
->subparser_stack
,
863 context
->subparser_stack
);
867 push_partial_as_tag (GMarkupParseContext
*context
)
869 GString
*str
= context
->partial_chunk
;
870 /* sadly, this is exported by gmarkup_get_element_stack as-is */
871 context
->tag_stack
= g_slist_concat (get_list_node (context
, str
->str
), context
->tag_stack
);
872 context
->tag_stack_gstr
= g_slist_concat (get_list_node (context
, str
), context
->tag_stack_gstr
);
873 context
->partial_chunk
= NULL
;
877 pop_tag (GMarkupParseContext
*context
)
879 GSList
*nodea
, *nodeb
;
881 nodea
= context
->tag_stack
;
882 nodeb
= context
->tag_stack_gstr
;
883 release_chunk (context
, nodeb
->data
);
884 context
->tag_stack
= g_slist_remove_link (context
->tag_stack
, nodea
);
885 context
->tag_stack_gstr
= g_slist_remove_link (context
->tag_stack_gstr
, nodeb
);
886 free_list_node (context
, nodea
);
887 free_list_node (context
, nodeb
);
891 possibly_finish_subparser (GMarkupParseContext
*context
)
893 if (current_element (context
) == context
->subparser_element
)
894 pop_subparser_stack (context
);
898 ensure_no_outstanding_subparser (GMarkupParseContext
*context
)
900 if (context
->awaiting_pop
)
901 g_critical ("During the first end_element call after invoking a "
902 "subparser you must pop the subparser stack and handle "
903 "the freeing of the subparser user_data. This can be "
904 "done by calling the end function of the subparser. "
905 "Very probably, your program just leaked memory.");
907 /* let valgrind watch the pointer disappear... */
908 context
->held_user_data
= NULL
;
909 context
->awaiting_pop
= FALSE
;
913 current_attribute (GMarkupParseContext
*context
)
915 g_assert (context
->cur_attr
>= 0);
916 return context
->attr_names
[context
->cur_attr
]->str
;
920 add_attribute (GMarkupParseContext
*context
, GString
*str
)
922 if (context
->cur_attr
+ 2 >= context
->alloc_attrs
)
924 context
->alloc_attrs
+= 5; /* silly magic number */
925 context
->attr_names
= g_realloc (context
->attr_names
, sizeof(GString
*)*context
->alloc_attrs
);
926 context
->attr_values
= g_realloc (context
->attr_values
, sizeof(GString
*)*context
->alloc_attrs
);
929 context
->attr_names
[context
->cur_attr
] = str
;
930 context
->attr_values
[context
->cur_attr
] = NULL
;
931 context
->attr_names
[context
->cur_attr
+1] = NULL
;
932 context
->attr_values
[context
->cur_attr
+1] = NULL
;
936 clear_attributes (GMarkupParseContext
*context
)
938 /* Go ahead and free the attributes. */
939 for (; context
->cur_attr
>= 0; context
->cur_attr
--)
941 int pos
= context
->cur_attr
;
942 release_chunk (context
, context
->attr_names
[pos
]);
943 release_chunk (context
, context
->attr_values
[pos
]);
944 context
->attr_names
[pos
] = context
->attr_values
[pos
] = NULL
;
946 g_assert (context
->cur_attr
== -1);
947 g_assert (context
->attr_names
== NULL
||
948 context
->attr_names
[0] == NULL
);
949 g_assert (context
->attr_values
== NULL
||
950 context
->attr_values
[0] == NULL
);
953 /* This has to be a separate function to ensure the alloca's
954 * are unwound on exit - otherwise we grow & blow the stack
955 * with large documents
958 emit_start_element (GMarkupParseContext
*context
,
962 const gchar
*start_name
;
963 const gchar
**attr_names
;
964 const gchar
**attr_values
;
967 attr_names
= g_newa (const gchar
*, context
->cur_attr
+ 2);
968 attr_values
= g_newa (const gchar
*, context
->cur_attr
+ 2);
969 for (i
= 0; i
< context
->cur_attr
+ 1; i
++)
971 attr_names
[i
] = context
->attr_names
[i
]->str
;
972 attr_values
[i
] = context
->attr_values
[i
]->str
;
974 attr_names
[i
] = NULL
;
975 attr_values
[i
] = NULL
;
977 /* Call user callback for element start */
979 start_name
= current_element (context
);
981 if (context
->parser
->start_element
&&
982 name_validate (context
, start_name
, error
))
983 (* context
->parser
->start_element
) (context
,
985 (const gchar
**)attr_names
,
986 (const gchar
**)attr_values
,
989 clear_attributes (context
);
991 if (tmp_error
!= NULL
)
992 propagate_error (context
, error
, tmp_error
);
996 * g_markup_parse_context_parse:
997 * @context: a #GMarkupParseContext
998 * @text: chunk of text to parse
999 * @text_len: length of @text in bytes
1000 * @error: return location for a #GError
1002 * Feed some data to the #GMarkupParseContext.
1004 * The data need not be valid UTF-8; an error will be signaled if
1005 * it's invalid. The data need not be an entire document; you can
1006 * feed a document into the parser incrementally, via multiple calls
1007 * to this function. Typically, as you receive data from a network
1008 * connection or file, you feed each received chunk of data into this
1009 * function, aborting the process if an error occurs. Once an error
1010 * is reported, no further data may be fed to the #GMarkupParseContext;
1011 * all errors are fatal.
1013 * Return value: %FALSE if an error occurred, %TRUE on success
1016 g_markup_parse_context_parse (GMarkupParseContext
*context
,
1021 g_return_val_if_fail (context
!= NULL
, FALSE
);
1022 g_return_val_if_fail (text
!= NULL
, FALSE
);
1023 g_return_val_if_fail (context
->state
!= STATE_ERROR
, FALSE
);
1024 g_return_val_if_fail (!context
->parsing
, FALSE
);
1027 text_len
= strlen (text
);
1032 context
->parsing
= TRUE
;
1035 context
->current_text
= text
;
1036 context
->current_text_len
= text_len
;
1037 context
->current_text_end
= context
->current_text
+ text_len
;
1038 context
->iter
= context
->current_text
;
1039 context
->start
= context
->iter
;
1041 while (context
->iter
!= context
->current_text_end
)
1043 switch (context
->state
)
1046 /* Possible next state: AFTER_OPEN_ANGLE */
1048 g_assert (context
->tag_stack
== NULL
);
1050 /* whitespace is ignored outside of any elements */
1051 skip_spaces (context
);
1053 if (context
->iter
!= context
->current_text_end
)
1055 if (*context
->iter
== '<')
1057 /* Move after the open angle */
1058 advance_char (context
);
1060 context
->state
= STATE_AFTER_OPEN_ANGLE
;
1062 /* this could start a passthrough */
1063 context
->start
= context
->iter
;
1065 /* document is now non-empty */
1066 context
->document_empty
= FALSE
;
1070 set_error_literal (context
,
1072 G_MARKUP_ERROR_PARSE
,
1073 _("Document must begin with an element (e.g. <book>)"));
1078 case STATE_AFTER_OPEN_ANGLE
:
1079 /* Possible next states: INSIDE_OPEN_TAG_NAME,
1080 * AFTER_CLOSE_TAG_SLASH, INSIDE_PASSTHROUGH
1082 if (*context
->iter
== '?' ||
1083 *context
->iter
== '!')
1085 /* include < in the passthrough */
1086 const gchar
*openangle
= "<";
1087 add_to_partial (context
, openangle
, openangle
+ 1);
1088 context
->start
= context
->iter
;
1089 context
->balance
= 1;
1090 context
->state
= STATE_INSIDE_PASSTHROUGH
;
1092 else if (*context
->iter
== '/')
1095 advance_char (context
);
1097 context
->state
= STATE_AFTER_CLOSE_TAG_SLASH
;
1099 else if (!IS_COMMON_NAME_END_CHAR (*(context
->iter
)))
1101 context
->state
= STATE_INSIDE_OPEN_TAG_NAME
;
1103 /* start of tag name */
1104 context
->start
= context
->iter
;
1112 G_MARKUP_ERROR_PARSE
,
1113 _("'%s' is not a valid character following "
1114 "a '<' character; it may not begin an "
1116 utf8_str (context
->iter
, buf
));
1120 /* The AFTER_CLOSE_ANGLE state is actually sort of
1121 * broken, because it doesn't correspond to a range
1122 * of characters in the input stream as the others do,
1123 * and thus makes things harder to conceptualize
1125 case STATE_AFTER_CLOSE_ANGLE
:
1126 /* Possible next states: INSIDE_TEXT, STATE_START */
1127 if (context
->tag_stack
== NULL
)
1129 context
->start
= NULL
;
1130 context
->state
= STATE_START
;
1134 context
->start
= context
->iter
;
1135 context
->state
= STATE_INSIDE_TEXT
;
1139 case STATE_AFTER_ELISION_SLASH
:
1140 /* Possible next state: AFTER_CLOSE_ANGLE */
1143 /* We need to pop the tag stack and call the end_element
1144 * function, since this is the close tag
1146 GError
*tmp_error
= NULL
;
1148 g_assert (context
->tag_stack
!= NULL
);
1150 possibly_finish_subparser (context
);
1153 if (context
->parser
->end_element
)
1154 (* context
->parser
->end_element
) (context
,
1155 current_element (context
),
1159 ensure_no_outstanding_subparser (context
);
1163 mark_error (context
, tmp_error
);
1164 g_propagate_error (error
, tmp_error
);
1168 if (*context
->iter
== '>')
1170 /* move after the close angle */
1171 advance_char (context
);
1172 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1180 G_MARKUP_ERROR_PARSE
,
1181 _("Odd character '%s', expected a '>' character "
1182 "to end the empty-element tag '%s'"),
1183 utf8_str (context
->iter
, buf
),
1184 current_element (context
));
1191 case STATE_INSIDE_OPEN_TAG_NAME
:
1192 /* Possible next states: BETWEEN_ATTRIBUTES */
1194 /* if there's a partial chunk then it's the first part of the
1195 * tag name. If there's a context->start then it's the start
1196 * of the tag name in current_text, the partial chunk goes
1197 * before that start though.
1199 advance_to_name_end (context
);
1201 if (context
->iter
== context
->current_text_end
)
1203 /* The name hasn't necessarily ended. Merge with
1204 * partial chunk, leave state unchanged.
1206 add_to_partial (context
, context
->start
, context
->iter
);
1210 /* The name has ended. Combine it with the partial chunk
1211 * if any; push it on the stack; enter next state.
1213 add_to_partial (context
, context
->start
, context
->iter
);
1214 push_partial_as_tag (context
);
1216 context
->state
= STATE_BETWEEN_ATTRIBUTES
;
1217 context
->start
= NULL
;
1221 case STATE_INSIDE_ATTRIBUTE_NAME
:
1222 /* Possible next states: AFTER_ATTRIBUTE_NAME */
1224 advance_to_name_end (context
);
1225 add_to_partial (context
, context
->start
, context
->iter
);
1227 /* read the full name, if we enter the equals sign state
1228 * then add the attribute to the list (without the value),
1229 * otherwise store a partial chunk to be prepended later.
1231 if (context
->iter
!= context
->current_text_end
)
1232 context
->state
= STATE_AFTER_ATTRIBUTE_NAME
;
1235 case STATE_AFTER_ATTRIBUTE_NAME
:
1236 /* Possible next states: AFTER_ATTRIBUTE_EQUALS_SIGN */
1238 skip_spaces (context
);
1240 if (context
->iter
!= context
->current_text_end
)
1242 /* The name has ended. Combine it with the partial chunk
1243 * if any; push it on the stack; enter next state.
1245 if (!name_validate (context
, context
->partial_chunk
->str
, error
))
1248 add_attribute (context
, context
->partial_chunk
);
1250 context
->partial_chunk
= NULL
;
1251 context
->start
= NULL
;
1253 if (*context
->iter
== '=')
1255 advance_char (context
);
1256 context
->state
= STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
;
1264 G_MARKUP_ERROR_PARSE
,
1265 _("Odd character '%s', expected a '=' after "
1266 "attribute name '%s' of element '%s'"),
1267 utf8_str (context
->iter
, buf
),
1268 current_attribute (context
),
1269 current_element (context
));
1275 case STATE_BETWEEN_ATTRIBUTES
:
1276 /* Possible next states: AFTER_CLOSE_ANGLE,
1277 * AFTER_ELISION_SLASH, INSIDE_ATTRIBUTE_NAME
1279 skip_spaces (context
);
1281 if (context
->iter
!= context
->current_text_end
)
1283 if (*context
->iter
== '/')
1285 advance_char (context
);
1286 context
->state
= STATE_AFTER_ELISION_SLASH
;
1288 else if (*context
->iter
== '>')
1290 advance_char (context
);
1291 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1293 else if (!IS_COMMON_NAME_END_CHAR (*(context
->iter
)))
1295 context
->state
= STATE_INSIDE_ATTRIBUTE_NAME
;
1296 /* start of attribute name */
1297 context
->start
= context
->iter
;
1305 G_MARKUP_ERROR_PARSE
,
1306 _("Odd character '%s', expected a '>' or '/' "
1307 "character to end the start tag of "
1308 "element '%s', or optionally an attribute; "
1309 "perhaps you used an invalid character in "
1310 "an attribute name"),
1311 utf8_str (context
->iter
, buf
),
1312 current_element (context
));
1315 /* If we're done with attributes, invoke
1316 * the start_element callback
1318 if (context
->state
== STATE_AFTER_ELISION_SLASH
||
1319 context
->state
== STATE_AFTER_CLOSE_ANGLE
)
1320 emit_start_element (context
, error
);
1324 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
:
1325 /* Possible next state: INSIDE_ATTRIBUTE_VALUE_[SQ/DQ] */
1327 skip_spaces (context
);
1329 if (context
->iter
!= context
->current_text_end
)
1331 if (*context
->iter
== '"')
1333 advance_char (context
);
1334 context
->state
= STATE_INSIDE_ATTRIBUTE_VALUE_DQ
;
1335 context
->start
= context
->iter
;
1337 else if (*context
->iter
== '\'')
1339 advance_char (context
);
1340 context
->state
= STATE_INSIDE_ATTRIBUTE_VALUE_SQ
;
1341 context
->start
= context
->iter
;
1349 G_MARKUP_ERROR_PARSE
,
1350 _("Odd character '%s', expected an open quote mark "
1351 "after the equals sign when giving value for "
1352 "attribute '%s' of element '%s'"),
1353 utf8_str (context
->iter
, buf
),
1354 current_attribute (context
),
1355 current_element (context
));
1360 case STATE_INSIDE_ATTRIBUTE_VALUE_SQ
:
1361 case STATE_INSIDE_ATTRIBUTE_VALUE_DQ
:
1362 /* Possible next states: BETWEEN_ATTRIBUTES */
1366 if (context
->state
== STATE_INSIDE_ATTRIBUTE_VALUE_SQ
)
1377 if (*context
->iter
== delim
)
1380 while (advance_char (context
));
1382 if (context
->iter
== context
->current_text_end
)
1384 /* The value hasn't necessarily ended. Merge with
1385 * partial chunk, leave state unchanged.
1387 add_to_partial (context
, context
->start
, context
->iter
);
1392 /* The value has ended at the quote mark. Combine it
1393 * with the partial chunk if any; set it for the current
1396 add_to_partial (context
, context
->start
, context
->iter
);
1398 g_assert (context
->cur_attr
>= 0);
1400 if (unescape_gstring_inplace (context
, context
->partial_chunk
, &is_ascii
, error
) &&
1401 (is_ascii
|| text_validate (context
, context
->partial_chunk
->str
,
1402 context
->partial_chunk
->len
, error
)))
1404 /* success, advance past quote and set state. */
1405 context
->attr_values
[context
->cur_attr
] = context
->partial_chunk
;
1406 context
->partial_chunk
= NULL
;
1407 advance_char (context
);
1408 context
->state
= STATE_BETWEEN_ATTRIBUTES
;
1409 context
->start
= NULL
;
1412 truncate_partial (context
);
1416 case STATE_INSIDE_TEXT
:
1417 /* Possible next states: AFTER_OPEN_ANGLE */
1420 if (*context
->iter
== '<')
1423 while (advance_char (context
));
1425 /* The text hasn't necessarily ended. Merge with
1426 * partial chunk, leave state unchanged.
1429 add_to_partial (context
, context
->start
, context
->iter
);
1431 if (context
->iter
!= context
->current_text_end
)
1435 /* The text has ended at the open angle. Call the text
1438 if (unescape_gstring_inplace (context
, context
->partial_chunk
, &is_ascii
, error
) &&
1439 (is_ascii
|| text_validate (context
, context
->partial_chunk
->str
,
1440 context
->partial_chunk
->len
, error
)))
1442 GError
*tmp_error
= NULL
;
1444 if (context
->parser
->text
)
1445 (*context
->parser
->text
) (context
,
1446 context
->partial_chunk
->str
,
1447 context
->partial_chunk
->len
,
1451 if (tmp_error
== NULL
)
1453 /* advance past open angle and set state. */
1454 advance_char (context
);
1455 context
->state
= STATE_AFTER_OPEN_ANGLE
;
1456 /* could begin a passthrough */
1457 context
->start
= context
->iter
;
1460 propagate_error (context
, error
, tmp_error
);
1463 truncate_partial (context
);
1467 case STATE_AFTER_CLOSE_TAG_SLASH
:
1468 /* Possible next state: INSIDE_CLOSE_TAG_NAME */
1469 if (!IS_COMMON_NAME_END_CHAR (*(context
->iter
)))
1471 context
->state
= STATE_INSIDE_CLOSE_TAG_NAME
;
1473 /* start of tag name */
1474 context
->start
= context
->iter
;
1482 G_MARKUP_ERROR_PARSE
,
1483 _("'%s' is not a valid character following "
1484 "the characters '</'; '%s' may not begin an "
1486 utf8_str (context
->iter
, buf
),
1487 utf8_str (context
->iter
, buf
));
1491 case STATE_INSIDE_CLOSE_TAG_NAME
:
1492 /* Possible next state: AFTER_CLOSE_TAG_NAME */
1493 advance_to_name_end (context
);
1494 add_to_partial (context
, context
->start
, context
->iter
);
1496 if (context
->iter
!= context
->current_text_end
)
1497 context
->state
= STATE_AFTER_CLOSE_TAG_NAME
;
1500 case STATE_AFTER_CLOSE_TAG_NAME
:
1501 /* Possible next state: AFTER_CLOSE_TAG_SLASH */
1503 skip_spaces (context
);
1505 if (context
->iter
!= context
->current_text_end
)
1507 GString
*close_name
;
1509 close_name
= context
->partial_chunk
;
1510 context
->partial_chunk
= NULL
;
1512 if (*context
->iter
!= '>')
1518 G_MARKUP_ERROR_PARSE
,
1519 _("'%s' is not a valid character following "
1520 "the close element name '%s'; the allowed "
1521 "character is '>'"),
1522 utf8_str (context
->iter
, buf
),
1525 else if (context
->tag_stack
== NULL
)
1529 G_MARKUP_ERROR_PARSE
,
1530 _("Element '%s' was closed, no element "
1531 "is currently open"),
1534 else if (strcmp (close_name
->str
, current_element (context
)) != 0)
1538 G_MARKUP_ERROR_PARSE
,
1539 _("Element '%s' was closed, but the currently "
1540 "open element is '%s'"),
1542 current_element (context
));
1547 advance_char (context
);
1548 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1549 context
->start
= NULL
;
1551 possibly_finish_subparser (context
);
1553 /* call the end_element callback */
1555 if (context
->parser
->end_element
)
1556 (* context
->parser
->end_element
) (context
,
1561 ensure_no_outstanding_subparser (context
);
1565 propagate_error (context
, error
, tmp_error
);
1567 context
->partial_chunk
= close_name
;
1568 truncate_partial (context
);
1572 case STATE_INSIDE_PASSTHROUGH
:
1573 /* Possible next state: AFTER_CLOSE_ANGLE */
1576 if (*context
->iter
== '<')
1578 if (*context
->iter
== '>')
1584 add_to_partial (context
, context
->start
, context
->iter
);
1585 context
->start
= context
->iter
;
1587 str
= context
->partial_chunk
->str
;
1588 len
= context
->partial_chunk
->len
;
1590 if (str
[1] == '?' && str
[len
- 1] == '?')
1592 if (strncmp (str
, "<!--", 4) == 0 &&
1593 strcmp (str
+ len
- 2, "--") == 0)
1595 if (strncmp (str
, "<![CDATA[", 9) == 0 &&
1596 strcmp (str
+ len
- 2, "]]") == 0)
1598 if (strncmp (str
, "<!DOCTYPE", 9) == 0 &&
1599 context
->balance
== 0)
1603 while (advance_char (context
));
1605 if (context
->iter
== context
->current_text_end
)
1607 /* The passthrough hasn't necessarily ended. Merge with
1608 * partial chunk, leave state unchanged.
1610 add_to_partial (context
, context
->start
, context
->iter
);
1614 /* The passthrough has ended at the close angle. Combine
1615 * it with the partial chunk if any. Call the passthrough
1616 * callback. Note that the open/close angles are
1617 * included in the text of the passthrough.
1619 GError
*tmp_error
= NULL
;
1621 advance_char (context
); /* advance past close angle */
1622 add_to_partial (context
, context
->start
, context
->iter
);
1624 if (context
->flags
& G_MARKUP_TREAT_CDATA_AS_TEXT
&&
1625 strncmp (context
->partial_chunk
->str
, "<![CDATA[", 9) == 0)
1627 if (context
->parser
->text
&&
1628 text_validate (context
,
1629 context
->partial_chunk
->str
+ 9,
1630 context
->partial_chunk
->len
- 12,
1632 (*context
->parser
->text
) (context
,
1633 context
->partial_chunk
->str
+ 9,
1634 context
->partial_chunk
->len
- 12,
1638 else if (context
->parser
->passthrough
&&
1639 text_validate (context
,
1640 context
->partial_chunk
->str
,
1641 context
->partial_chunk
->len
,
1643 (*context
->parser
->passthrough
) (context
,
1644 context
->partial_chunk
->str
,
1645 context
->partial_chunk
->len
,
1649 truncate_partial (context
);
1651 if (tmp_error
== NULL
)
1653 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1654 context
->start
= context
->iter
; /* could begin text */
1657 propagate_error (context
, error
, tmp_error
);
1666 g_assert_not_reached ();
1672 context
->parsing
= FALSE
;
1674 return context
->state
!= STATE_ERROR
;
1678 * g_markup_parse_context_end_parse:
1679 * @context: a #GMarkupParseContext
1680 * @error: return location for a #GError
1682 * Signals to the #GMarkupParseContext that all data has been
1683 * fed into the parse context with g_markup_parse_context_parse().
1685 * This function reports an error if the document isn't complete,
1686 * for example if elements are still open.
1688 * Return value: %TRUE on success, %FALSE if an error was set
1691 g_markup_parse_context_end_parse (GMarkupParseContext
*context
,
1694 g_return_val_if_fail (context
!= NULL
, FALSE
);
1695 g_return_val_if_fail (!context
->parsing
, FALSE
);
1696 g_return_val_if_fail (context
->state
!= STATE_ERROR
, FALSE
);
1698 if (context
->partial_chunk
!= NULL
)
1700 g_string_free (context
->partial_chunk
, TRUE
);
1701 context
->partial_chunk
= NULL
;
1704 if (context
->document_empty
)
1706 set_error_literal (context
, error
, G_MARKUP_ERROR_EMPTY
,
1707 _("Document was empty or contained only whitespace"));
1711 context
->parsing
= TRUE
;
1713 switch (context
->state
)
1719 case STATE_AFTER_OPEN_ANGLE
:
1720 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1721 _("Document ended unexpectedly just after an open angle bracket '<'"));
1724 case STATE_AFTER_CLOSE_ANGLE
:
1725 if (context
->tag_stack
!= NULL
)
1727 /* Error message the same as for INSIDE_TEXT */
1728 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1729 _("Document ended unexpectedly with elements still open - "
1730 "'%s' was the last element opened"),
1731 current_element (context
));
1735 case STATE_AFTER_ELISION_SLASH
:
1736 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1737 _("Document ended unexpectedly, expected to see a close angle "
1738 "bracket ending the tag <%s/>"), current_element (context
));
1741 case STATE_INSIDE_OPEN_TAG_NAME
:
1742 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1743 _("Document ended unexpectedly inside an element name"));
1746 case STATE_INSIDE_ATTRIBUTE_NAME
:
1747 case STATE_AFTER_ATTRIBUTE_NAME
:
1748 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1749 _("Document ended unexpectedly inside an attribute name"));
1752 case STATE_BETWEEN_ATTRIBUTES
:
1753 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1754 _("Document ended unexpectedly inside an element-opening "
1758 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
:
1759 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1760 _("Document ended unexpectedly after the equals sign "
1761 "following an attribute name; no attribute value"));
1764 case STATE_INSIDE_ATTRIBUTE_VALUE_SQ
:
1765 case STATE_INSIDE_ATTRIBUTE_VALUE_DQ
:
1766 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1767 _("Document ended unexpectedly while inside an attribute "
1771 case STATE_INSIDE_TEXT
:
1772 g_assert (context
->tag_stack
!= NULL
);
1773 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1774 _("Document ended unexpectedly with elements still open - "
1775 "'%s' was the last element opened"),
1776 current_element (context
));
1779 case STATE_AFTER_CLOSE_TAG_SLASH
:
1780 case STATE_INSIDE_CLOSE_TAG_NAME
:
1781 case STATE_AFTER_CLOSE_TAG_NAME
:
1782 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1783 _("Document ended unexpectedly inside the close tag for "
1784 "element '%s'"), current_element (context
));
1787 case STATE_INSIDE_PASSTHROUGH
:
1788 set_error_literal (context
, error
, G_MARKUP_ERROR_PARSE
,
1789 _("Document ended unexpectedly inside a comment or "
1790 "processing instruction"));
1795 g_assert_not_reached ();
1799 context
->parsing
= FALSE
;
1801 return context
->state
!= STATE_ERROR
;
1805 * g_markup_parse_context_get_element:
1806 * @context: a #GMarkupParseContext
1808 * Retrieves the name of the currently open element.
1810 * If called from the start_element or end_element handlers this will
1811 * give the element_name as passed to those functions. For the parent
1812 * elements, see g_markup_parse_context_get_element_stack().
1814 * Returns: the name of the currently open element, or %NULL
1819 g_markup_parse_context_get_element (GMarkupParseContext
*context
)
1821 g_return_val_if_fail (context
!= NULL
, NULL
);
1823 if (context
->tag_stack
== NULL
)
1826 return current_element (context
);
1830 * g_markup_parse_context_get_element_stack:
1831 * @context: a #GMarkupParseContext
1833 * Retrieves the element stack from the internal state of the parser.
1835 * The returned #GSList is a list of strings where the first item is
1836 * the currently open tag (as would be returned by
1837 * g_markup_parse_context_get_element()) and the next item is its
1840 * This function is intended to be used in the start_element and
1841 * end_element handlers where g_markup_parse_context_get_element()
1842 * would merely return the name of the element that is being
1845 * Returns: the element stack, which must not be modified
1850 g_markup_parse_context_get_element_stack (GMarkupParseContext
*context
)
1852 g_return_val_if_fail (context
!= NULL
, NULL
);
1853 return context
->tag_stack
;
1857 * g_markup_parse_context_get_position:
1858 * @context: a #GMarkupParseContext
1859 * @line_number: (allow-none): return location for a line number, or %NULL
1860 * @char_number: (allow-none): return location for a char-on-line number, or %NULL
1862 * Retrieves the current line number and the number of the character on
1863 * that line. Intended for use in error messages; there are no strict
1864 * semantics for what constitutes the "current" line number other than
1865 * "the best number we could come up with for error messages."
1868 g_markup_parse_context_get_position (GMarkupParseContext
*context
,
1872 g_return_if_fail (context
!= NULL
);
1875 *line_number
= context
->line_number
;
1878 *char_number
= context
->char_number
;
1882 * g_markup_parse_context_get_user_data:
1883 * @context: a #GMarkupParseContext
1885 * Returns the user_data associated with @context.
1887 * This will either be the user_data that was provided to
1888 * g_markup_parse_context_new() or to the most recent call
1889 * of g_markup_parse_context_push().
1891 * Returns: the provided user_data. The returned data belongs to
1892 * the markup context and will be freed when
1893 * g_markup_parse_context_free() is called.
1898 g_markup_parse_context_get_user_data (GMarkupParseContext
*context
)
1900 return context
->user_data
;
1904 * g_markup_parse_context_push:
1905 * @context: a #GMarkupParseContext
1906 * @parser: a #GMarkupParser
1907 * @user_data: user data to pass to #GMarkupParser functions
1909 * Temporarily redirects markup data to a sub-parser.
1911 * This function may only be called from the start_element handler of
1912 * a #GMarkupParser. It must be matched with a corresponding call to
1913 * g_markup_parse_context_pop() in the matching end_element handler
1914 * (except in the case that the parser aborts due to an error).
1916 * All tags, text and other data between the matching tags is
1917 * redirected to the subparser given by @parser. @user_data is used
1918 * as the user_data for that parser. @user_data is also passed to the
1919 * error callback in the event that an error occurs. This includes
1920 * errors that occur in subparsers of the subparser.
1922 * The end tag matching the start tag for which this call was made is
1923 * handled by the previous parser (which is given its own user_data)
1924 * which is why g_markup_parse_context_pop() is provided to allow "one
1925 * last access" to the @user_data provided to this function. In the
1926 * case of error, the @user_data provided here is passed directly to
1927 * the error callback of the subparser and g_markup_parse_context_pop()
1928 * should not be called. In either case, if @user_data was allocated
1929 * then it ought to be freed from both of these locations.
1931 * This function is not intended to be directly called by users
1932 * interested in invoking subparsers. Instead, it is intended to be
1933 * used by the subparsers themselves to implement a higher-level
1936 * As an example, see the following implementation of a simple
1937 * parser that counts the number of tags encountered.
1946 * counter_start_element (GMarkupParseContext *context,
1947 * const gchar *element_name,
1948 * const gchar **attribute_names,
1949 * const gchar **attribute_values,
1950 * gpointer user_data,
1953 * CounterData *data = user_data;
1955 * data->tag_count++;
1959 * counter_error (GMarkupParseContext *context,
1961 * gpointer user_data)
1963 * CounterData *data = user_data;
1965 * g_slice_free (CounterData, data);
1968 * static GMarkupParser counter_subparser =
1970 * counter_start_element,
1978 * In order to allow this parser to be easily used as a subparser, the
1979 * following interface is provided:
1983 * start_counting (GMarkupParseContext *context)
1985 * CounterData *data = g_slice_new (CounterData);
1987 * data->tag_count = 0;
1988 * g_markup_parse_context_push (context, &counter_subparser, data);
1992 * end_counting (GMarkupParseContext *context)
1994 * CounterData *data = g_markup_parse_context_pop (context);
1997 * result = data->tag_count;
1998 * g_slice_free (CounterData, data);
2004 * The subparser would then be used as follows:
2007 * static void start_element (context, element_name, ...)
2009 * if (strcmp (element_name, "count-these") == 0)
2010 * start_counting (context);
2012 * /* else, handle other tags... */
2015 * static void end_element (context, element_name, ...)
2017 * if (strcmp (element_name, "count-these") == 0)
2018 * g_print ("Counted %d tags\n", end_counting (context));
2020 * /* else, handle other tags... */
2027 g_markup_parse_context_push (GMarkupParseContext
*context
,
2028 const GMarkupParser
*parser
,
2031 GMarkupRecursionTracker
*tracker
;
2033 tracker
= g_slice_new (GMarkupRecursionTracker
);
2034 tracker
->prev_element
= context
->subparser_element
;
2035 tracker
->prev_parser
= context
->parser
;
2036 tracker
->prev_user_data
= context
->user_data
;
2038 context
->subparser_element
= current_element (context
);
2039 context
->parser
= parser
;
2040 context
->user_data
= user_data
;
2042 context
->subparser_stack
= g_slist_prepend (context
->subparser_stack
,
2047 * g_markup_parse_context_pop:
2048 * @context: a #GMarkupParseContext
2050 * Completes the process of a temporary sub-parser redirection.
2052 * This function exists to collect the user_data allocated by a
2053 * matching call to g_markup_parse_context_push(). It must be called
2054 * in the end_element handler corresponding to the start_element
2055 * handler during which g_markup_parse_context_push() was called.
2056 * You must not call this function from the error callback -- the
2057 * @user_data is provided directly to the callback in that case.
2059 * This function is not intended to be directly called by users
2060 * interested in invoking subparsers. Instead, it is intended to
2061 * be used by the subparsers themselves to implement a higher-level
2064 * Returns: the user data passed to g_markup_parse_context_push()
2069 g_markup_parse_context_pop (GMarkupParseContext
*context
)
2073 if (!context
->awaiting_pop
)
2074 possibly_finish_subparser (context
);
2076 g_assert (context
->awaiting_pop
);
2078 context
->awaiting_pop
= FALSE
;
2080 /* valgrind friendliness */
2081 user_data
= context
->held_user_data
;
2082 context
->held_user_data
= NULL
;
2088 append_escaped_text (GString
*str
,
2097 end
= text
+ length
;
2102 next
= g_utf8_next_char (p
);
2107 g_string_append (str
, "&");
2111 g_string_append (str
, "<");
2115 g_string_append (str
, ">");
2119 g_string_append (str
, "'");
2123 g_string_append (str
, """);
2127 c
= g_utf8_get_char (p
);
2128 if ((0x1 <= c
&& c
<= 0x8) ||
2129 (0xb <= c
&& c
<= 0xc) ||
2130 (0xe <= c
&& c
<= 0x1f) ||
2131 (0x7f <= c
&& c
<= 0x84) ||
2132 (0x86 <= c
&& c
<= 0x9f))
2133 g_string_append_printf (str
, "&#x%x;", c
);
2135 g_string_append_len (str
, p
, next
- p
);
2144 * g_markup_escape_text:
2145 * @text: some valid UTF-8 text
2146 * @length: length of @text in bytes, or -1 if the text is nul-terminated
2148 * Escapes text so that the markup parser will parse it verbatim.
2149 * Less than, greater than, ampersand, etc. are replaced with the
2150 * corresponding entities. This function would typically be used
2151 * when writing out a file to be parsed with the markup parser.
2153 * Note that this function doesn't protect whitespace and line endings
2154 * from being processed according to the XML rules for normalization
2155 * of line endings and attribute values.
2157 * Note also that this function will produce character references in
2158 * the range of &#x1; ... &#x1f; for all control sequences
2159 * except for tabstop, newline and carriage return. The character
2160 * references in this range are not valid XML 1.0, but they are
2161 * valid XML 1.1 and will be accepted by the GMarkup parser.
2163 * Return value: a newly allocated string with the escaped text
2166 g_markup_escape_text (const gchar
*text
,
2171 g_return_val_if_fail (text
!= NULL
, NULL
);
2174 length
= strlen (text
);
2176 /* prealloc at least as long as original text */
2177 str
= g_string_sized_new (length
);
2178 append_escaped_text (str
, text
, length
);
2180 return g_string_free (str
, FALSE
);
2185 * @format: a printf-style format string
2186 * @after: location to store a pointer to the character after
2187 * the returned conversion. On a %NULL return, returns the
2188 * pointer to the trailing NUL in the string
2190 * Find the next conversion in a printf-style format string.
2191 * Partially based on code from printf-parser.c,
2192 * Copyright (C) 1999-2000, 2002-2003 Free Software Foundation, Inc.
2194 * Return value: pointer to the next conversion in @format,
2195 * or %NULL, if none.
2198 find_conversion (const char *format
,
2201 const char *start
= format
;
2204 while (*start
!= '\0' && *start
!= '%')
2221 /* Test for positional argument. */
2222 if (*cp
>= '0' && *cp
<= '9')
2226 for (np
= cp
; *np
>= '0' && *np
<= '9'; np
++)
2232 /* Skip the flags. */
2246 /* Skip the field width. */
2251 /* Test for positional argument. */
2252 if (*cp
>= '0' && *cp
<= '9')
2256 for (np
= cp
; *np
>= '0' && *np
<= '9'; np
++)
2264 for (; *cp
>= '0' && *cp
<= '9'; cp
++)
2268 /* Skip the precision. */
2274 /* Test for positional argument. */
2275 if (*cp
>= '0' && *cp
<= '9')
2279 for (np
= cp
; *np
>= '0' && *np
<= '9'; np
++)
2287 for (; *cp
>= '0' && *cp
<= '9'; cp
++)
2292 /* Skip argument type/size specifiers. */
2293 while (*cp
== 'h' ||
2302 /* Skip the conversion character. */
2310 * g_markup_vprintf_escaped:
2311 * @format: printf() style format string
2312 * @args: variable argument list, similar to vprintf()
2314 * Formats the data in @args according to @format, escaping
2315 * all string and character arguments in the fashion
2316 * of g_markup_escape_text(). See g_markup_printf_escaped().
2318 * Return value: newly allocated result from formatting
2319 * operation. Free with g_free().
2324 g_markup_vprintf_escaped (const gchar
*format
,
2329 GString
*result
= NULL
;
2330 gchar
*output1
= NULL
;
2331 gchar
*output2
= NULL
;
2332 const char *p
, *op1
, *op2
;
2335 /* The technique here, is that we make two format strings that
2336 * have the identical conversions in the identical order to the
2337 * original strings, but differ in the text in-between. We
2338 * then use the normal g_strdup_vprintf() to format the arguments
2339 * with the two new format strings. By comparing the results,
2340 * we can figure out what segments of the output come from
2341 * the original format string, and what from the arguments,
2342 * and thus know what portions of the string to escape.
2344 * For instance, for:
2346 * g_markup_printf_escaped ("%s ate %d apples", "Susan & Fred", 5);
2348 * We form the two format strings "%sX%dX" and %sY%sY". The results
2349 * of formatting with those two strings are
2351 * "%sX%dX" => "Susan & FredX5X"
2352 * "%sY%dY" => "Susan & FredY5Y"
2354 * To find the span of the first argument, we find the first position
2355 * where the two arguments differ, which tells us that the first
2356 * argument formatted to "Susan & Fred". We then escape that
2357 * to "Susan & Fred" and join up with the intermediate portions
2358 * of the format string and the second argument to get
2359 * "Susan & Fred ate 5 apples".
2362 /* Create the two modified format strings
2364 format1
= g_string_new (NULL
);
2365 format2
= g_string_new (NULL
);
2370 const char *conv
= find_conversion (p
, &after
);
2374 g_string_append_len (format1
, conv
, after
- conv
);
2375 g_string_append_c (format1
, 'X');
2376 g_string_append_len (format2
, conv
, after
- conv
);
2377 g_string_append_c (format2
, 'Y');
2382 /* Use them to format the arguments
2384 G_VA_COPY (args2
, args
);
2386 output1
= g_strdup_vprintf (format1
->str
, args
);
2393 output2
= g_strdup_vprintf (format2
->str
, args2
);
2398 result
= g_string_new (NULL
);
2400 /* Iterate through the original format string again,
2401 * copying the non-conversion portions and the escaped
2402 * converted arguments to the output string.
2410 const char *output_start
;
2411 const char *conv
= find_conversion (p
, &after
);
2414 if (!conv
) /* The end, after points to the trailing \0 */
2416 g_string_append_len (result
, p
, after
- p
);
2420 g_string_append_len (result
, p
, conv
- p
);
2422 while (*op1
== *op2
)
2428 escaped
= g_markup_escape_text (output_start
, op1
- output_start
);
2429 g_string_append (result
, escaped
);
2438 g_string_free (format1
, TRUE
);
2439 g_string_free (format2
, TRUE
);
2444 return g_string_free (result
, FALSE
);
2450 * g_markup_printf_escaped:
2451 * @format: printf() style format string
2452 * @...: the arguments to insert in the format string
2454 * Formats arguments according to @format, escaping
2455 * all string and character arguments in the fashion
2456 * of g_markup_escape_text(). This is useful when you
2457 * want to insert literal strings into XML-style markup
2458 * output, without having to worry that the strings
2459 * might themselves contain markup.
2462 * const char *store = "Fortnum & Mason";
2463 * const char *item = "Tea";
2466 * output = g_markup_printf_escaped ("<purchase>"
2467 * "<store>%s</store>"
2468 * "<item>%s</item>"
2469 * "</purchase>",
2473 * Return value: newly allocated result from formatting
2474 * operation. Free with g_free().
2479 g_markup_printf_escaped (const gchar
*format
, ...)
2484 va_start (args
, format
);
2485 result
= g_markup_vprintf_escaped (format
, args
);
2492 g_markup_parse_boolean (const char *string
,
2495 char const * const falses
[] = { "false", "f", "no", "n", "0" };
2496 char const * const trues
[] = { "true", "t", "yes", "y", "1" };
2499 for (i
= 0; i
< G_N_ELEMENTS (falses
); i
++)
2501 if (g_ascii_strcasecmp (string
, falses
[i
]) == 0)
2510 for (i
= 0; i
< G_N_ELEMENTS (trues
); i
++)
2512 if (g_ascii_strcasecmp (string
, trues
[i
]) == 0)
2525 * GMarkupCollectType:
2526 * @G_MARKUP_COLLECT_INVALID: used to terminate the list of attributes
2528 * @G_MARKUP_COLLECT_STRING: collect the string pointer directly from
2529 * the attribute_values[] array. Expects a parameter of type (const
2530 * char **). If %G_MARKUP_COLLECT_OPTIONAL is specified and the
2531 * attribute isn't present then the pointer will be set to %NULL
2532 * @G_MARKUP_COLLECT_STRDUP: as with %G_MARKUP_COLLECT_STRING, but
2533 * expects a parameter of type (char **) and g_strdup()s the
2534 * returned pointer. The pointer must be freed with g_free()
2535 * @G_MARKUP_COLLECT_BOOLEAN: expects a parameter of type (gboolean *)
2536 * and parses the attribute value as a boolean. Sets %FALSE if the
2537 * attribute isn't present. Valid boolean values consist of
2538 * (case-insensitive) "false", "f", "no", "n", "0" and "true", "t",
2540 * @G_MARKUP_COLLECT_TRISTATE: as with %G_MARKUP_COLLECT_BOOLEAN, but
2541 * in the case of a missing attribute a value is set that compares
2542 * equal to neither %FALSE nor %TRUE G_MARKUP_COLLECT_OPTIONAL is
2544 * @G_MARKUP_COLLECT_OPTIONAL: can be bitwise ORed with the other fields.
2545 * If present, allows the attribute not to appear. A default value
2546 * is set depending on what value type is used
2548 * A mixed enumerated type and flags field. You must specify one type
2549 * (string, strdup, boolean, tristate). Additionally, you may optionally
2550 * bitwise OR the type with the flag %G_MARKUP_COLLECT_OPTIONAL.
2552 * It is likely that this enum will be extended in the future to
2553 * support other types.
2557 * g_markup_collect_attributes:
2558 * @element_name: the current tag name
2559 * @attribute_names: the attribute names
2560 * @attribute_values: the attribute values
2561 * @error: a pointer to a #GError or %NULL
2562 * @first_type: the #GMarkupCollectType of the first attribute
2563 * @first_attr: the name of the first attribute
2564 * @...: a pointer to the storage location of the first attribute
2565 * (or %NULL), followed by more types names and pointers, ending
2566 * with %G_MARKUP_COLLECT_INVALID
2568 * Collects the attributes of the element from the data passed to the
2569 * #GMarkupParser start_element function, dealing with common error
2570 * conditions and supporting boolean values.
2572 * This utility function is not required to write a parser but can save
2575 * The @element_name, @attribute_names, @attribute_values and @error
2576 * parameters passed to the start_element callback should be passed
2577 * unmodified to this function.
2579 * Following these arguments is a list of "supported" attributes to collect.
2580 * It is an error to specify multiple attributes with the same name. If any
2581 * attribute not in the list appears in the @attribute_names array then an
2582 * unknown attribute error will result.
2584 * The #GMarkupCollectType field allows specifying the type of collection
2585 * to perform and if a given attribute must appear or is optional.
2587 * The attribute name is simply the name of the attribute to collect.
2589 * The pointer should be of the appropriate type (see the descriptions
2590 * under #GMarkupCollectType) and may be %NULL in case a particular
2591 * attribute is to be allowed but ignored.
2593 * This function deals with issuing errors for missing attributes
2594 * (of type %G_MARKUP_ERROR_MISSING_ATTRIBUTE), unknown attributes
2595 * (of type %G_MARKUP_ERROR_UNKNOWN_ATTRIBUTE) and duplicate
2596 * attributes (of type %G_MARKUP_ERROR_INVALID_CONTENT) as well
2597 * as parse errors for boolean-valued attributes (again of type
2598 * %G_MARKUP_ERROR_INVALID_CONTENT). In all of these cases %FALSE
2599 * will be returned and @error will be set as appropriate.
2601 * Return value: %TRUE if successful
2606 g_markup_collect_attributes (const gchar
*element_name
,
2607 const gchar
**attribute_names
,
2608 const gchar
**attribute_values
,
2610 GMarkupCollectType first_type
,
2611 const gchar
*first_attr
,
2614 GMarkupCollectType type
;
2626 va_start (ap
, first_attr
);
2627 while (type
!= G_MARKUP_COLLECT_INVALID
)
2632 mandatory
= !(type
& G_MARKUP_COLLECT_OPTIONAL
);
2633 type
&= (G_MARKUP_COLLECT_OPTIONAL
- 1);
2635 /* tristate records a value != TRUE and != FALSE
2636 * for the case where the attribute is missing
2638 if (type
== G_MARKUP_COLLECT_TRISTATE
)
2641 for (i
= 0; attribute_names
[i
]; i
++)
2642 if (i
>= 40 || !(collected
& (G_GUINT64_CONSTANT(1) << i
)))
2643 if (!strcmp (attribute_names
[i
], attr
))
2646 /* ISO C99 only promises that the user can pass up to 127 arguments.
2647 * Subtracting the first 4 arguments plus the final NULL and dividing
2648 * by 3 arguments per collected attribute, we are left with a maximum
2649 * number of supported attributes of (127 - 5) / 3 = 40.
2651 * In reality, nobody is ever going to call us with anywhere close to
2652 * 40 attributes to collect, so it is safe to assume that if i > 40
2653 * then the user has given some invalid or repeated arguments. These
2654 * problems will be caught and reported at the end of the function.
2656 * We know at this point that we have an error, but we don't know
2657 * what error it is, so just continue...
2660 collected
|= (G_GUINT64_CONSTANT(1) << i
);
2662 value
= attribute_values
[i
];
2664 if (value
== NULL
&& mandatory
)
2666 g_set_error (error
, G_MARKUP_ERROR
,
2667 G_MARKUP_ERROR_MISSING_ATTRIBUTE
,
2668 "element '%s' requires attribute '%s'",
2669 element_name
, attr
);
2677 case G_MARKUP_COLLECT_STRING
:
2679 const char **str_ptr
;
2681 str_ptr
= va_arg (ap
, const char **);
2683 if (str_ptr
!= NULL
)
2688 case G_MARKUP_COLLECT_STRDUP
:
2692 str_ptr
= va_arg (ap
, char **);
2694 if (str_ptr
!= NULL
)
2695 *str_ptr
= g_strdup (value
);
2699 case G_MARKUP_COLLECT_BOOLEAN
:
2700 case G_MARKUP_COLLECT_TRISTATE
:
2705 bool_ptr
= va_arg (ap
, gboolean
*);
2707 if (bool_ptr
!= NULL
)
2709 if (type
== G_MARKUP_COLLECT_TRISTATE
)
2710 /* constructivists rejoice!
2711 * neither false nor true...
2715 else /* G_MARKUP_COLLECT_BOOLEAN */
2721 if (!g_markup_parse_boolean (value
, va_arg (ap
, gboolean
*)))
2723 g_set_error (error
, G_MARKUP_ERROR
,
2724 G_MARKUP_ERROR_INVALID_CONTENT
,
2725 "element '%s', attribute '%s', value '%s' "
2726 "cannot be parsed as a boolean value",
2727 element_name
, attr
, value
);
2737 g_assert_not_reached ();
2740 type
= va_arg (ap
, GMarkupCollectType
);
2741 attr
= va_arg (ap
, const char *);
2746 /* ensure we collected all the arguments */
2747 for (i
= 0; attribute_names
[i
]; i
++)
2748 if ((collected
& (G_GUINT64_CONSTANT(1) << i
)) == 0)
2750 /* attribute not collected: could be caused by two things.
2752 * 1) it doesn't exist in our list of attributes
2753 * 2) it existed but was matched by a duplicate attribute earlier
2759 for (j
= 0; j
< i
; j
++)
2760 if (strcmp (attribute_names
[i
], attribute_names
[j
]) == 0)
2764 /* j is now the first occurrence of attribute_names[i] */
2766 g_set_error (error
, G_MARKUP_ERROR
,
2767 G_MARKUP_ERROR_UNKNOWN_ATTRIBUTE
,
2768 "attribute '%s' invalid for element '%s'",
2769 attribute_names
[i
], element_name
);
2771 g_set_error (error
, G_MARKUP_ERROR
,
2772 G_MARKUP_ERROR_INVALID_CONTENT
,
2773 "attribute '%s' given multiple times for element '%s'",
2774 attribute_names
[i
], element_name
);
2782 /* replay the above to free allocations */
2786 va_start (ap
, first_attr
);
2787 while (type
!= G_MARKUP_COLLECT_INVALID
)
2791 ptr
= va_arg (ap
, gpointer
);
2795 switch (type
& (G_MARKUP_COLLECT_OPTIONAL
- 1))
2797 case G_MARKUP_COLLECT_STRDUP
:
2799 g_free (*(char **) ptr
);
2801 case G_MARKUP_COLLECT_STRING
:
2802 *(char **) ptr
= NULL
;
2805 case G_MARKUP_COLLECT_BOOLEAN
:
2806 *(gboolean
*) ptr
= FALSE
;
2809 case G_MARKUP_COLLECT_TRISTATE
:
2810 *(gboolean
*) ptr
= -1;
2815 type
= va_arg (ap
, GMarkupCollectType
);
2816 attr
= va_arg (ap
, const char *);