1 /* gmarkup.c - Simple XML-like parser
3 * Copyright 2000 Red Hat, Inc.
5 * GLib is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU Lesser General Public License as
7 * published by the Free Software Foundation; either version 2 of the
8 * License, or (at your option) any later version.
10 * GLib is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * Lesser General Public License for more details.
15 * You should have received a copy of the GNU Lesser General Public
16 * License along with GLib; see the file COPYING.LIB. If not,
17 * write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
18 * Boston, MA 02111-1307, USA.
31 g_markup_error_quark ()
33 static GQuark error_quark
= 0;
36 error_quark
= g_quark_from_static_string ("g-markup-error-quark");
44 STATE_AFTER_OPEN_ANGLE
,
45 STATE_AFTER_CLOSE_ANGLE
,
46 STATE_AFTER_ELISION_SLASH
, /* the slash that obviates need for end element */
47 STATE_INSIDE_OPEN_TAG_NAME
,
48 STATE_INSIDE_ATTRIBUTE_NAME
,
49 STATE_BETWEEN_ATTRIBUTES
,
50 STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
,
51 STATE_INSIDE_ATTRIBUTE_VALUE
,
53 STATE_AFTER_CLOSE_TAG_SLASH
,
54 STATE_INSIDE_CLOSE_TAG_NAME
,
55 STATE_INSIDE_PASSTHROUGH
,
59 struct _GMarkupParseContext
61 const GMarkupParser
*parser
;
63 GMarkupParseFlags flags
;
69 GDestroyNotify dnotify
;
71 /* A piece of character data or an element that
72 * hasn't "ended" yet so we haven't yet called
73 * the callback for it.
75 GString
*partial_chunk
;
77 GMarkupParseState state
;
84 const gchar
*current_text
;
85 gint current_text_len
;
86 const gchar
*current_text_end
;
88 GString
*leftover_char_portion
;
90 /* used to save the start of the last interesting thingy */
95 guint document_empty
: 1;
100 * g_markup_parse_context_new:
101 * @parser: a #GMarkupParser
102 * @flags: one or more #GMarkupParseFlags
103 * @user_data: user data to pass to #GMarkupParser functions
104 * @user_data_dnotify: user data destroy notifier called when the parse context is freed
106 * Creates a new parse context. A parse context is used to parse
107 * marked-up documents. You can feed any number of documents into
108 * a context, as long as no errors occur; once an error occurs,
109 * the parse context can't continue to parse text (you have to free it
110 * and create a new parse context).
112 * Return value: a new #GMarkupParseContext
114 GMarkupParseContext
*
115 g_markup_parse_context_new (const GMarkupParser
*parser
,
116 GMarkupParseFlags flags
,
118 GDestroyNotify user_data_dnotify
)
120 GMarkupParseContext
*context
;
122 g_return_val_if_fail (parser
!= NULL
, NULL
);
124 context
= g_new (GMarkupParseContext
, 1);
126 context
->parser
= parser
;
127 context
->flags
= flags
;
128 context
->user_data
= user_data
;
129 context
->dnotify
= user_data_dnotify
;
131 context
->line_number
= 1;
132 context
->char_number
= 1;
134 context
->partial_chunk
= NULL
;
136 context
->state
= STATE_START
;
137 context
->tag_stack
= NULL
;
138 context
->attr_names
= NULL
;
139 context
->attr_values
= NULL
;
140 context
->cur_attr
= -1;
141 context
->alloc_attrs
= 0;
143 context
->current_text
= NULL
;
144 context
->current_text_len
= -1;
145 context
->current_text_end
= NULL
;
146 context
->leftover_char_portion
= NULL
;
148 context
->start
= NULL
;
149 context
->iter
= NULL
;
151 context
->document_empty
= TRUE
;
152 context
->parsing
= FALSE
;
158 * g_markup_parse_context_free:
159 * @context: a #GMarkupParseContext
161 * Frees a #GMarkupParseContext. Can't be called from inside
162 * one of the #GMarkupParser functions.
166 g_markup_parse_context_free (GMarkupParseContext
*context
)
168 g_return_if_fail (context
!= NULL
);
169 g_return_if_fail (!context
->parsing
);
171 if (context
->dnotify
)
172 (* context
->dnotify
) (context
->user_data
);
174 g_strfreev (context
->attr_names
);
175 g_strfreev (context
->attr_values
);
177 g_slist_foreach (context
->tag_stack
, (GFunc
)g_free
, NULL
);
178 g_slist_free (context
->tag_stack
);
180 if (context
->partial_chunk
)
181 g_string_free (context
->partial_chunk
, TRUE
);
183 if (context
->leftover_char_portion
)
184 g_string_free (context
->leftover_char_portion
, TRUE
);
190 mark_error (GMarkupParseContext
*context
,
193 context
->state
= STATE_ERROR
;
195 if (context
->parser
->error
)
196 (*context
->parser
->error
) (context
, error
, context
->user_data
);
200 set_error (GMarkupParseContext
*context
,
210 va_start (args
, format
);
211 s
= g_strdup_vprintf (format
, args
);
214 tmp_error
= g_error_new (G_MARKUP_ERROR
,
216 _("Error on line %d char %d: %s"),
217 context
->line_number
,
218 context
->char_number
,
223 mark_error (context
, tmp_error
);
225 g_propagate_error (error
, tmp_error
);
229 is_name_start_char (gunichar c
)
231 if (g_unichar_isalpha (c
) ||
240 is_name_char (gunichar c
)
242 if (g_unichar_isalnum (c
) ||
254 char_str (gunichar c
,
258 g_unichar_to_utf8 (c
, buf
);
263 utf8_str (const gchar
*utf8
,
266 char_str (g_utf8_get_char (utf8
), buf
);
271 set_unescape_error (GMarkupParseContext
*context
,
273 const gchar
*remaining_text
,
274 const gchar
*remaining_text_end
,
282 gint remaining_newlines
;
285 remaining_newlines
= 0;
287 while (p
!= remaining_text_end
)
290 ++remaining_newlines
;
294 va_start (args
, format
);
295 s
= g_strdup_vprintf (format
, args
);
298 tmp_error
= g_error_new (G_MARKUP_ERROR
,
300 _("Error on line %d: %s"),
301 context
->line_number
- remaining_newlines
,
306 mark_error (context
, tmp_error
);
308 g_propagate_error (error
, tmp_error
);
314 USTATE_AFTER_AMPERSAND
,
315 USTATE_INSIDE_ENTITY_NAME
,
316 USTATE_AFTER_CHARREF_HASH
320 unescape_text (GMarkupParseContext
*context
,
322 const gchar
*text_end
,
326 #define MAX_ENT_LEN 5
332 str
= g_string_new ("");
334 state
= USTATE_INSIDE_TEXT
;
337 while (p
!= text_end
&& context
->state
!= STATE_ERROR
)
339 g_assert (p
< text_end
);
343 case USTATE_INSIDE_TEXT
:
345 while (p
!= text_end
&& *p
!= '&')
346 p
= g_utf8_next_char (p
);
350 g_string_append_len (str
, start
, p
- start
);
355 if (p
!= text_end
&& *p
== '&')
357 p
= g_utf8_next_char (p
);
358 state
= USTATE_AFTER_AMPERSAND
;
363 case USTATE_AFTER_AMPERSAND
:
367 p
= g_utf8_next_char (p
);
370 state
= USTATE_AFTER_CHARREF_HASH
;
372 else if (!is_name_start_char (g_utf8_get_char (p
)))
376 set_unescape_error (context
, error
,
378 G_MARKUP_ERROR_PARSE
,
379 _("Empty entity '&;' seen; valid "
380 "entities are: & " < > '"));
386 set_unescape_error (context
, error
,
388 G_MARKUP_ERROR_PARSE
,
389 _("Character '%s' is not valid at "
390 "the start of an entity name; "
391 "the & character begins an entity; "
392 "if this ampersand isn't supposed "
393 "to be an entity, escape it as "
401 state
= USTATE_INSIDE_ENTITY_NAME
;
407 case USTATE_INSIDE_ENTITY_NAME
:
409 gchar buf
[MAX_ENT_LEN
+1] = {
410 '\0', '\0', '\0', '\0', '\0', '\0'
414 while (p
!= text_end
)
418 else if (!is_name_char (*p
))
422 set_unescape_error (context
, error
,
424 G_MARKUP_ERROR_PARSE
,
425 _("Character '%s' is not valid "
426 "inside an entity name"),
431 p
= g_utf8_next_char (p
);
434 if (context
->state
!= STATE_ERROR
)
449 /* move to after semicolon */
450 p
= g_utf8_next_char (p
);
452 state
= USTATE_INSIDE_TEXT
;
454 if (strcmp (buf
, "lt") == 0)
455 g_string_append_c (str
, '<');
456 else if (strcmp (buf
, "gt") == 0)
457 g_string_append_c (str
, '>');
458 else if (strcmp (buf
, "amp") == 0)
459 g_string_append_c (str
, '&');
460 else if (strcmp (buf
, "quot") == 0)
461 g_string_append_c (str
, '"');
462 else if (strcmp (buf
, "apos") == 0)
463 g_string_append_c (str
, '\'');
466 set_unescape_error (context
, error
,
468 G_MARKUP_ERROR_PARSE
,
469 _("Entity name '%s' is not known"),
475 set_unescape_error (context
, error
,
476 /* give line number of the & */
478 G_MARKUP_ERROR_PARSE
,
479 _("Entity did not end with a semicolon; "
480 "most likely you used an ampersand "
481 "character without intending to start "
482 "an entity - escape ampersand as &"));
488 case USTATE_AFTER_CHARREF_HASH
:
490 gboolean is_hex
= FALSE
;
494 p
= g_utf8_next_char (p
);
498 while (p
!= text_end
&& *p
!= ';')
499 p
= g_utf8_next_char (p
);
503 g_assert (*p
== ';');
505 /* digit is between start and p */
509 gchar
*digit
= g_strndup (start
, p
- start
);
512 gchar
*digit_end
= digit
+ (p
- start
);
516 l
= strtoul (digit
, &end
, 16);
518 l
= strtoul (digit
, &end
, 10);
520 if (end
!= digit_end
|| errno
!= 0)
522 set_unescape_error (context
, error
,
524 G_MARKUP_ERROR_PARSE
,
525 _("Failed to parse '%s', which "
526 "should have been a digit "
527 "inside a character reference "
528 "(ê for example) - perhaps "
529 "the digit is too large"),
534 /* characters XML permits */
538 (l
>= 0x20 && l
<= 0xD7FF) ||
539 (l
>= 0xE000 && l
<= 0xFFFD) ||
540 (l
>= 0x10000 && l
<= 0x10FFFF))
543 g_string_append (str
,
548 set_unescape_error (context
, error
,
550 G_MARKUP_ERROR_PARSE
,
551 _("Character reference '%s' does not encode a permitted character"),
558 /* Move to next state */
559 p
= g_utf8_next_char (p
); /* past semicolon */
561 state
= USTATE_INSIDE_TEXT
;
565 set_unescape_error (context
, error
,
567 G_MARKUP_ERROR_PARSE
,
568 _("Empty character reference; "
569 "should include a digit such as "
575 set_unescape_error (context
, error
,
577 G_MARKUP_ERROR_PARSE
,
578 _("Character reference did not end with a "
580 "most likely you used an ampersand "
581 "character without intending to start "
582 "an entity - escape ampersand as &"));
588 g_assert_not_reached ();
593 /* If no errors, we should have returned to USTATE_INSIDE_TEXT */
594 g_assert (context
->state
== STATE_ERROR
||
595 state
== USTATE_INSIDE_TEXT
);
597 if (context
->state
== STATE_ERROR
)
599 g_string_free (str
, TRUE
);
605 *unescaped
= g_string_free (str
, FALSE
);
613 advance_char (GMarkupParseContext
*context
)
616 context
->iter
= g_utf8_next_char (context
->iter
);
617 context
->char_number
+= 1;
618 if (*context
->iter
== '\n')
620 context
->line_number
+= 1;
621 context
->char_number
= 1;
624 return context
->iter
!= context
->current_text_end
;
628 skip_spaces (GMarkupParseContext
*context
)
632 if (!g_unichar_isspace (g_utf8_get_char (context
->iter
)))
635 while (advance_char (context
));
639 advance_to_name_end (GMarkupParseContext
*context
)
643 if (!is_name_char (g_utf8_get_char (context
->iter
)))
646 while (advance_char (context
));
650 add_to_partial (GMarkupParseContext
*context
,
651 const gchar
*text_start
,
652 const gchar
*text_end
)
654 if (context
->partial_chunk
== NULL
)
655 context
->partial_chunk
= g_string_new ("");
657 if (text_start
!= text_end
)
658 g_string_append_len (context
->partial_chunk
, text_start
,
659 text_end
- text_start
);
661 /* Invariant here that partial_chunk exists */
665 truncate_partial (GMarkupParseContext
*context
)
667 if (context
->partial_chunk
!= NULL
)
669 context
->partial_chunk
= g_string_truncate (context
->partial_chunk
, 0);
674 current_element (GMarkupParseContext
*context
)
676 return context
->tag_stack
->data
;
680 current_attribute (GMarkupParseContext
*context
)
682 g_assert (context
->cur_attr
>= 0);
683 return context
->attr_names
[context
->cur_attr
];
687 find_current_text_end (GMarkupParseContext
*context
)
689 /* This function must be safe (non-segfaulting) on invalid UTF8 */
690 const gchar
*end
= context
->current_text
+ context
->current_text_len
;
694 g_assert (context
->current_text_len
> 0);
696 p
= context
->current_text
;
697 next
= g_utf8_find_next_char (p
, end
);
702 next
= g_utf8_find_next_char (p
, end
);
705 /* p is now the start of the last character or character portion. */
707 next
= g_utf8_next_char (p
); /* this only touches *p, nothing beyond */
711 /* whole character */
712 context
->current_text_end
= end
;
717 context
->leftover_char_portion
= g_string_new_len (p
, end
- p
);
718 context
->current_text_len
-= (end
- p
);
719 context
->current_text_end
= p
;
724 add_attribute (GMarkupParseContext
*context
, char *name
)
726 if (context
->cur_attr
+ 2 >= context
->alloc_attrs
)
728 context
->alloc_attrs
+= 5; /* silly magic number */
729 context
->attr_names
= g_realloc (context
->attr_names
, sizeof(char*)*context
->alloc_attrs
);
730 context
->attr_values
= g_realloc (context
->attr_values
, sizeof(char*)*context
->alloc_attrs
);
733 context
->attr_names
[context
->cur_attr
] = name
;
734 context
->attr_values
[context
->cur_attr
] = NULL
;
735 context
->attr_names
[context
->cur_attr
+1] = NULL
;
739 * g_markup_parse_context_parse:
740 * @context: a #GMarkupParseContext
741 * @text: chunk of text to parse
742 * @text_len: length of @text in bytes
743 * @error: return location for a #GError
745 * Feed some data to the #GMarkupParseContext. The data need not
746 * be valid UTF-8; an error will be signaled if it's invalid.
747 * The data need not be an entire document; you can feed a document
748 * into the parser incrementally, via multiple calls to this function.
749 * Typically, as you receive data from a network connection or file,
750 * you feed each received chunk of data into this function, aborting
751 * the process if an error occurs. Once an error is reported, no further
752 * data may be fed to the #GMarkupParseContext; all errors are fatal.
754 * Return value: %FALSE if an error occurred, %TRUE on success
757 g_markup_parse_context_parse (GMarkupParseContext
*context
,
762 const gchar
*first_invalid
;
764 g_return_val_if_fail (context
!= NULL
, FALSE
);
765 g_return_val_if_fail (text
!= NULL
, FALSE
);
766 g_return_val_if_fail (context
->state
!= STATE_ERROR
, FALSE
);
767 g_return_val_if_fail (!context
->parsing
, FALSE
);
770 text_len
= strlen (text
);
775 context
->parsing
= TRUE
;
777 if (context
->leftover_char_portion
)
779 const gchar
*first_char
;
781 if ((*text
& 0xc0) != 0x80)
784 first_char
= g_utf8_find_next_char (text
, text
+ text_len
);
788 /* leftover_char_portion was completed. Parse it. */
789 GString
*portion
= context
->leftover_char_portion
;
791 g_string_append_len (context
->leftover_char_portion
,
792 text
, first_char
- text
);
794 /* hacks to allow recursion */
795 context
->parsing
= FALSE
;
796 context
->leftover_char_portion
= NULL
;
798 if (!g_markup_parse_context_parse (context
,
799 portion
->str
, portion
->len
,
802 g_assert (context
->state
== STATE_ERROR
);
805 g_string_free (portion
, TRUE
);
806 context
->parsing
= TRUE
;
808 /* Skip the fraction of char that was in this text */
809 text_len
-= (first_char
- text
);
814 /* another little chunk of the leftover char; geez
815 * someone is inefficient.
817 g_string_append_len (context
->leftover_char_portion
,
820 if (context
->leftover_char_portion
->len
> 7)
822 /* The leftover char portion is too big to be
827 G_MARKUP_ERROR_BAD_UTF8
,
828 _("Invalid UTF-8 encoded text"));
835 context
->current_text
= text
;
836 context
->current_text_len
= text_len
;
837 context
->iter
= context
->current_text
;
838 context
->start
= context
->iter
;
840 /* Nothing left after finishing the leftover char, or nothing
841 * passed in to begin with.
843 if (context
->current_text_len
== 0)
846 /* find_current_text_end () assumes the string starts at
847 * a character start, so we need to validate at least
848 * that much. It doesn't assume any following bytes
851 if ((*context
->current_text
& 0xc0) == 0x80) /* not a char start */
855 G_MARKUP_ERROR_BAD_UTF8
,
856 _("Invalid UTF-8 encoded text"));
860 /* Initialize context->current_text_end, possibly adjusting
861 * current_text_len, and add any leftover char portion
863 find_current_text_end (context
);
865 /* Validate UTF8 (must be done after we find the end, since
866 * we could have a trailing incomplete char)
868 if (!g_utf8_validate (context
->current_text
,
869 context
->current_text_len
,
874 p
= context
->current_text
;
875 while (p
!= context
->current_text_end
)
882 context
->line_number
+= newlines
;
886 G_MARKUP_ERROR_BAD_UTF8
,
887 _("Invalid UTF-8 encoded text"));
891 while (context
->iter
!= context
->current_text_end
)
893 switch (context
->state
)
896 /* Possible next state: AFTER_OPEN_ANGLE */
898 g_assert (context
->tag_stack
== NULL
);
900 /* whitespace is ignored outside of any elements */
901 skip_spaces (context
);
903 if (context
->iter
!= context
->current_text_end
)
905 if (*context
->iter
== '<')
907 /* Move after the open angle */
908 advance_char (context
);
910 context
->state
= STATE_AFTER_OPEN_ANGLE
;
912 /* this could start a passthrough */
913 context
->start
= context
->iter
;
915 /* document is now non-empty */
916 context
->document_empty
= FALSE
;
922 G_MARKUP_ERROR_PARSE
,
923 _("Document must begin with an element (e.g. <book>)"));
928 case STATE_AFTER_OPEN_ANGLE
:
929 /* Possible next states: INSIDE_OPEN_TAG_NAME,
930 * AFTER_CLOSE_TAG_SLASH, INSIDE_PASSTHROUGH
932 if (*context
->iter
== '?' ||
933 *context
->iter
== '!')
935 /* include < in the passthrough */
936 const gchar
*openangle
= "<";
937 add_to_partial (context
, openangle
, openangle
+ 1);
938 context
->start
= context
->iter
;
939 context
->state
= STATE_INSIDE_PASSTHROUGH
;
941 else if (*context
->iter
== '/')
944 advance_char (context
);
946 context
->state
= STATE_AFTER_CLOSE_TAG_SLASH
;
948 else if (is_name_start_char (g_utf8_get_char (context
->iter
)))
950 context
->state
= STATE_INSIDE_OPEN_TAG_NAME
;
952 /* start of tag name */
953 context
->start
= context
->iter
;
960 G_MARKUP_ERROR_PARSE
,
961 _("'%s' is not a valid character following "
962 "a '<' character; it may not begin an "
964 utf8_str (context
->iter
, buf
));
968 /* The AFTER_CLOSE_ANGLE state is actually sort of
969 * broken, because it doesn't correspond to a range
970 * of characters in the input stream as the others do,
971 * and thus makes things harder to conceptualize
973 case STATE_AFTER_CLOSE_ANGLE
:
974 /* Possible next states: INSIDE_TEXT, STATE_START */
975 if (context
->tag_stack
== NULL
)
977 context
->start
= NULL
;
978 context
->state
= STATE_START
;
982 context
->start
= context
->iter
;
983 context
->state
= STATE_INSIDE_TEXT
;
987 case STATE_AFTER_ELISION_SLASH
:
988 /* Possible next state: AFTER_CLOSE_ANGLE */
991 /* We need to pop the tag stack and call the end_element
992 * function, since this is the close tag
994 GError
*tmp_error
= NULL
;
996 g_assert (context
->tag_stack
!= NULL
);
999 if (context
->parser
->end_element
)
1000 (* context
->parser
->end_element
) (context
,
1001 context
->tag_stack
->data
,
1007 mark_error (context
, tmp_error
);
1008 g_propagate_error (error
, tmp_error
);
1012 if (*context
->iter
== '>')
1014 /* move after the close angle */
1015 advance_char (context
);
1016 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1023 G_MARKUP_ERROR_PARSE
,
1024 _("Odd character '%s', expected a '>' character "
1025 "to end the start tag of element '%s'"),
1026 utf8_str (context
->iter
, buf
),
1027 current_element (context
));
1031 g_free (context
->tag_stack
->data
);
1032 context
->tag_stack
= g_slist_delete_link (context
->tag_stack
,
1033 context
->tag_stack
);
1037 case STATE_INSIDE_OPEN_TAG_NAME
:
1038 /* Possible next states: BETWEEN_ATTRIBUTES */
1040 /* if there's a partial chunk then it's the first part of the
1041 * tag name. If there's a context->start then it's the start
1042 * of the tag name in current_text, the partial chunk goes
1043 * before that start though.
1045 advance_to_name_end (context
);
1047 if (context
->iter
== context
->current_text_end
)
1049 /* The name hasn't necessarily ended. Merge with
1050 * partial chunk, leave state unchanged.
1052 add_to_partial (context
, context
->start
, context
->iter
);
1056 /* The name has ended. Combine it with the partial chunk
1057 * if any; push it on the stack; enter next state.
1059 add_to_partial (context
, context
->start
, context
->iter
);
1060 context
->tag_stack
=
1061 g_slist_prepend (context
->tag_stack
,
1062 g_string_free (context
->partial_chunk
,
1065 context
->partial_chunk
= NULL
;
1067 context
->state
= STATE_BETWEEN_ATTRIBUTES
;
1068 context
->start
= NULL
;
1072 case STATE_INSIDE_ATTRIBUTE_NAME
:
1073 /* Possible next states: AFTER_ATTRIBUTE_EQUALS_SIGN */
1075 /* read the full name, if we enter the equals sign state
1076 * then add the attribute to the list (without the value),
1077 * otherwise store a partial chunk to be prepended later.
1079 advance_to_name_end (context
);
1081 if (context
->iter
== context
->current_text_end
)
1083 /* The name hasn't necessarily ended. Merge with
1084 * partial chunk, leave state unchanged.
1086 add_to_partial (context
, context
->start
, context
->iter
);
1090 /* The name has ended. Combine it with the partial chunk
1091 * if any; push it on the stack; enter next state.
1093 add_to_partial (context
, context
->start
, context
->iter
);
1095 add_attribute (context
, g_string_free (context
->partial_chunk
, FALSE
));
1097 context
->partial_chunk
= NULL
;
1098 context
->start
= NULL
;
1100 if (*context
->iter
== '=')
1102 advance_char (context
);
1103 context
->state
= STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
;
1110 G_MARKUP_ERROR_PARSE
,
1111 _("Odd character '%s', expected a '=' after "
1112 "attribute name '%s' of element '%s'"),
1113 utf8_str (context
->iter
, buf
),
1114 current_attribute (context
),
1115 current_element (context
));
1121 case STATE_BETWEEN_ATTRIBUTES
:
1122 /* Possible next states: AFTER_CLOSE_ANGLE,
1123 * AFTER_ELISION_SLASH, INSIDE_ATTRIBUTE_NAME
1125 skip_spaces (context
);
1127 if (context
->iter
!= context
->current_text_end
)
1129 if (*context
->iter
== '/')
1131 advance_char (context
);
1132 context
->state
= STATE_AFTER_ELISION_SLASH
;
1134 else if (*context
->iter
== '>')
1137 advance_char (context
);
1138 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1140 else if (is_name_start_char (g_utf8_get_char (context
->iter
)))
1142 context
->state
= STATE_INSIDE_ATTRIBUTE_NAME
;
1143 /* start of attribute name */
1144 context
->start
= context
->iter
;
1151 G_MARKUP_ERROR_PARSE
,
1152 _("Odd character '%s', expected a '>' or '/' "
1153 "character to end the start tag of "
1154 "element '%s', or optionally an attribute; "
1155 "perhaps you used an invalid character in "
1156 "an attribute name"),
1157 utf8_str (context
->iter
, buf
),
1158 current_element (context
));
1161 /* If we're done with attributes, invoke
1162 * the start_element callback
1164 if (context
->state
== STATE_AFTER_ELISION_SLASH
||
1165 context
->state
== STATE_AFTER_CLOSE_ANGLE
)
1167 const gchar
*start_name
;
1168 /* Ugly, but the current code expects an empty array instead of NULL */
1169 const gchar
*empty
= NULL
;
1170 const gchar
**attr_names
= &empty
;
1171 const gchar
**attr_values
= &empty
;
1174 /* Call user callback for element start */
1175 start_name
= current_element (context
);
1177 if (context
->cur_attr
>= 0)
1179 attr_names
= (const gchar
**)context
->attr_names
;
1180 attr_values
= (const gchar
**)context
->attr_values
;
1184 if (context
->parser
->start_element
)
1185 (* context
->parser
->start_element
) (context
,
1187 (const gchar
**)attr_names
,
1188 (const gchar
**)attr_values
,
1192 /* Go ahead and free the attributes. */
1193 for (; context
->cur_attr
>= 0; context
->cur_attr
--)
1195 int pos
= context
->cur_attr
;
1196 g_free (context
->attr_names
[pos
]);
1197 g_free (context
->attr_values
[pos
]);
1198 context
->attr_names
[pos
] = context
->attr_values
[pos
] = NULL
;
1200 context
->cur_attr
= -1;
1202 if (tmp_error
!= NULL
)
1204 mark_error (context
, tmp_error
);
1205 g_propagate_error (error
, tmp_error
);
1211 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
:
1212 /* Possible next state: INSIDE_ATTRIBUTE_VALUE */
1213 if (*context
->iter
== '"')
1215 advance_char (context
);
1216 context
->state
= STATE_INSIDE_ATTRIBUTE_VALUE
;
1217 context
->start
= context
->iter
;
1224 G_MARKUP_ERROR_PARSE
,
1225 _("Odd character '%s', expected an open quote mark "
1226 "after the equals sign when giving value for "
1227 "attribute '%s' of element '%s'"),
1228 utf8_str (context
->iter
, buf
),
1229 current_attribute (context
),
1230 current_element (context
));
1234 case STATE_INSIDE_ATTRIBUTE_VALUE
:
1235 /* Possible next states: BETWEEN_ATTRIBUTES */
1238 if (*context
->iter
== '"')
1241 while (advance_char (context
));
1243 if (context
->iter
== context
->current_text_end
)
1245 /* The value hasn't necessarily ended. Merge with
1246 * partial chunk, leave state unchanged.
1248 add_to_partial (context
, context
->start
, context
->iter
);
1252 /* The value has ended at the quote mark. Combine it
1253 * with the partial chunk if any; set it for the current
1256 add_to_partial (context
, context
->start
, context
->iter
);
1258 g_assert (context
->cur_attr
>= 0);
1260 if (unescape_text (context
,
1261 context
->partial_chunk
->str
,
1262 context
->partial_chunk
->str
+
1263 context
->partial_chunk
->len
,
1264 &context
->attr_values
[context
->cur_attr
],
1267 /* success, advance past quote and set state. */
1268 advance_char (context
);
1269 context
->state
= STATE_BETWEEN_ATTRIBUTES
;
1270 context
->start
= NULL
;
1273 truncate_partial (context
);
1277 case STATE_INSIDE_TEXT
:
1278 /* Possible next states: AFTER_OPEN_ANGLE */
1281 if (*context
->iter
== '<')
1284 while (advance_char (context
));
1286 /* The text hasn't necessarily ended. Merge with
1287 * partial chunk, leave state unchanged.
1290 add_to_partial (context
, context
->start
, context
->iter
);
1292 if (context
->iter
!= context
->current_text_end
)
1294 gchar
*unescaped
= NULL
;
1296 /* The text has ended at the open angle. Call the text
1300 if (unescape_text (context
,
1301 context
->partial_chunk
->str
,
1302 context
->partial_chunk
->str
+
1303 context
->partial_chunk
->len
,
1307 GError
*tmp_error
= NULL
;
1309 if (context
->parser
->text
)
1310 (*context
->parser
->text
) (context
,
1318 if (tmp_error
== NULL
)
1320 /* advance past open angle and set state. */
1321 advance_char (context
);
1322 context
->state
= STATE_AFTER_OPEN_ANGLE
;
1323 /* could begin a passthrough */
1324 context
->start
= context
->iter
;
1328 mark_error (context
, tmp_error
);
1329 g_propagate_error (error
, tmp_error
);
1333 truncate_partial (context
);
1337 case STATE_AFTER_CLOSE_TAG_SLASH
:
1338 /* Possible next state: INSIDE_CLOSE_TAG_NAME */
1339 if (is_name_start_char (g_utf8_get_char (context
->iter
)))
1341 context
->state
= STATE_INSIDE_CLOSE_TAG_NAME
;
1343 /* start of tag name */
1344 context
->start
= context
->iter
;
1351 G_MARKUP_ERROR_PARSE
,
1352 _("'%s' is not a valid character following "
1353 "the characters '</'; '%s' may not begin an "
1355 utf8_str (context
->iter
, buf
),
1356 utf8_str (context
->iter
, buf
));
1360 case STATE_INSIDE_CLOSE_TAG_NAME
:
1361 /* Possible next state: AFTER_CLOSE_ANGLE */
1362 advance_to_name_end (context
);
1364 if (context
->iter
== context
->current_text_end
)
1366 /* The name hasn't necessarily ended. Merge with
1367 * partial chunk, leave state unchanged.
1369 add_to_partial (context
, context
->start
, context
->iter
);
1373 /* The name has ended. Combine it with the partial chunk
1374 * if any; check that it matches stack top and pop
1375 * stack; invoke proper callback; enter next state.
1379 add_to_partial (context
, context
->start
, context
->iter
);
1381 close_name
= g_string_free (context
->partial_chunk
, FALSE
);
1382 context
->partial_chunk
= NULL
;
1384 if (context
->tag_stack
== NULL
)
1388 G_MARKUP_ERROR_PARSE
,
1389 _("Element '%s' was closed, no element "
1390 "is currently open"),
1393 else if (strcmp (close_name
, current_element (context
)) != 0)
1397 G_MARKUP_ERROR_PARSE
,
1398 _("Element '%s' was closed, but the currently "
1399 "open element is '%s'"),
1401 current_element (context
));
1403 else if (*context
->iter
!= '>')
1408 G_MARKUP_ERROR_PARSE
,
1409 _("'%s' is not a valid character following "
1410 "the close element name '%s'; the allowed "
1411 "character is '>'"),
1412 utf8_str (context
->iter
, buf
),
1418 advance_char (context
);
1419 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1420 context
->start
= NULL
;
1422 /* call the end_element callback */
1424 if (context
->parser
->end_element
)
1425 (* context
->parser
->end_element
) (context
,
1431 /* Pop the tag stack */
1432 g_free (context
->tag_stack
->data
);
1433 context
->tag_stack
= g_slist_delete_link (context
->tag_stack
,
1434 context
->tag_stack
);
1438 mark_error (context
, tmp_error
);
1439 g_propagate_error (error
, tmp_error
);
1443 g_free (close_name
);
1447 case STATE_INSIDE_PASSTHROUGH
:
1448 /* Possible next state: AFTER_CLOSE_ANGLE */
1451 if (*context
->iter
== '>')
1454 while (advance_char (context
));
1456 if (context
->iter
== context
->current_text_end
)
1458 /* The passthrough hasn't necessarily ended. Merge with
1459 * partial chunk, leave state unchanged.
1461 add_to_partial (context
, context
->start
, context
->iter
);
1465 /* The passthrough has ended at the close angle. Combine
1466 * it with the partial chunk if any. Call the passthrough
1467 * callback. Note that the open/close angles are
1468 * included in the text of the passthrough.
1470 GError
*tmp_error
= NULL
;
1472 advance_char (context
); /* advance past close angle */
1473 add_to_partial (context
, context
->start
, context
->iter
);
1475 if (context
->parser
->passthrough
)
1476 (*context
->parser
->passthrough
) (context
,
1477 context
->partial_chunk
->str
,
1478 context
->partial_chunk
->len
,
1482 truncate_partial (context
);
1484 if (tmp_error
== NULL
)
1486 context
->state
= STATE_AFTER_CLOSE_ANGLE
;
1487 context
->start
= context
->iter
; /* could begin text */
1491 mark_error (context
, tmp_error
);
1492 g_propagate_error (error
, tmp_error
);
1502 g_assert_not_reached ();
1508 context
->parsing
= FALSE
;
1510 return context
->state
!= STATE_ERROR
;
1514 * g_markup_parse_context_end_parse:
1515 * @context: a #GMarkupParseContext
1516 * @error: return location for a #GError
1518 * Signals to the #GMarkupParseContext that all data has been
1519 * fed into the parse context with g_markup_parse_context_parse().
1520 * This function reports an error if the document isn't complete,
1521 * for example if elements are still open.
1523 * Return value: %TRUE on success, %FALSE if an error was set
1526 g_markup_parse_context_end_parse (GMarkupParseContext
*context
,
1529 g_return_val_if_fail (context
!= NULL
, FALSE
);
1530 g_return_val_if_fail (!context
->parsing
, FALSE
);
1531 g_return_val_if_fail (context
->state
!= STATE_ERROR
, FALSE
);
1533 if (context
->partial_chunk
!= NULL
)
1535 g_string_free (context
->partial_chunk
, TRUE
);
1536 context
->partial_chunk
= NULL
;
1539 if (context
->document_empty
)
1541 set_error (context
, error
, G_MARKUP_ERROR_EMPTY
,
1542 _("Document was empty or contained only whitespace"));
1546 context
->parsing
= TRUE
;
1548 switch (context
->state
)
1554 case STATE_AFTER_OPEN_ANGLE
:
1555 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1556 _("Document ended unexpectedly just after an open angle bracket '<'"));
1559 case STATE_AFTER_CLOSE_ANGLE
:
1560 if (context
->tag_stack
!= NULL
)
1562 /* Error message the same as for INSIDE_TEXT */
1563 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1564 _("Document ended unexpectedly with elements still open - "
1565 "'%s' was the last element opened"),
1566 current_element (context
));
1570 case STATE_AFTER_ELISION_SLASH
:
1571 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1572 _("Document ended unexpectedly, expected to see a close angle "
1573 "bracket ending the tag <%s/>"), current_element (context
));
1576 case STATE_INSIDE_OPEN_TAG_NAME
:
1577 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1578 _("Document ended unexpectedly inside an element name"));
1581 case STATE_INSIDE_ATTRIBUTE_NAME
:
1582 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1583 _("Document ended unexpectedly inside an attribute name"));
1586 case STATE_BETWEEN_ATTRIBUTES
:
1587 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1588 _("Document ended unexpectedly inside an element-opening "
1592 case STATE_AFTER_ATTRIBUTE_EQUALS_SIGN
:
1593 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1594 _("Document ended unexpectedly after the equals sign "
1595 "following an attribute name; no attribute value"));
1598 case STATE_INSIDE_ATTRIBUTE_VALUE
:
1599 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1600 _("Document ended unexpectedly while inside an attribute "
1604 case STATE_INSIDE_TEXT
:
1605 g_assert (context
->tag_stack
!= NULL
);
1606 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1607 _("Document ended unexpectedly with elements still open - "
1608 "'%s' was the last element opened"),
1609 current_element (context
));
1612 case STATE_AFTER_CLOSE_TAG_SLASH
:
1613 case STATE_INSIDE_CLOSE_TAG_NAME
:
1614 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1615 _("Document ended unexpectedly inside the close tag for"
1616 "element '%s'"), current_element
);
1619 case STATE_INSIDE_PASSTHROUGH
:
1620 set_error (context
, error
, G_MARKUP_ERROR_PARSE
,
1621 _("Document ended unexpectedly inside a comment or "
1622 "processing instruction"));
1627 g_assert_not_reached ();
1631 context
->parsing
= FALSE
;
1633 return context
->state
!= STATE_ERROR
;
1637 * g_markup_parse_context_get_position:
1638 * @context: a #GMarkupParseContext
1639 * @line_number: return location for a line number, or %NULL
1640 * @char_number: return location for a char-on-line number, or %NULL
1642 * Retrieves the current line number and the number of the character on
1643 * that line. Intended for use in error messages; there are no strict
1644 * semantics for what constitutes the "current" line number other than
1645 * "the best number we could come up with for error messages."
1649 g_markup_parse_context_get_position (GMarkupParseContext
*context
,
1653 g_return_if_fail (context
!= NULL
);
1656 *line_number
= context
->line_number
;
1659 *char_number
= context
->char_number
;
1663 append_escaped_text (GString
*str
,
1671 end
= text
+ length
;
1676 next
= g_utf8_next_char (p
);
1681 g_string_append (str
, "&");
1685 g_string_append (str
, "<");
1689 g_string_append (str
, ">");
1693 g_string_append (str
, "'");
1697 g_string_append (str
, """);
1701 g_string_append_len (str
, p
, next
- p
);
1710 * g_markup_escape_text:
1711 * @text: some valid UTF-8 text
1712 * @length: length of @text in bytes
1714 * Escapes text so that the markup parser will parse it verbatim.
1715 * Less than, greater than, ampersand, etc. are replaced with the
1716 * corresponding entities. This function would typically be used
1717 * when writing out a file to be parsed with the markup parser.
1719 * Return value: escaped text
1722 g_markup_escape_text (const gchar
*text
,
1727 g_return_val_if_fail (text
!= NULL
, NULL
);
1730 length
= strlen (text
);
1732 str
= g_string_new ("");
1733 append_escaped_text (str
, text
, length
);
1735 return g_string_free (str
, FALSE
);