1 /* GRegex -- regular expression API wrapper around PCRE.
3 * Copyright (C) 1999, 2000 Scott Wimer
4 * Copyright (C) 2004, Matthias Clasen <mclasen@redhat.com>
5 * Copyright (C) 2005 - 2007, Marco Barisione <marco@barisione.org>
7 * This library is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
12 * This library is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with this library; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
30 #ifdef USE_SYSTEM_PCRE
33 #include "pcre/pcre.h"
36 /* PCRE 7.3 does not contain the definition of PCRE_ERROR_NULLWSLIMIT */
37 #ifndef PCRE_ERROR_NULLWSLIMIT
38 #define PCRE_ERROR_NULLWSLIMIT (-22)
43 /* Mask of all the possible values for GRegexCompileFlags. */
44 #define G_REGEX_COMPILE_MASK (G_REGEX_CASELESS | \
49 G_REGEX_DOLLAR_ENDONLY | \
52 G_REGEX_NO_AUTO_CAPTURE | \
55 G_REGEX_NEWLINE_CR | \
56 G_REGEX_NEWLINE_LF | \
59 /* Mask of all the possible values for GRegexMatchFlags. */
60 #define G_REGEX_MATCH_MASK (G_REGEX_MATCH_ANCHORED | \
61 G_REGEX_MATCH_NOTBOL | \
62 G_REGEX_MATCH_NOTEOL | \
63 G_REGEX_MATCH_NOTEMPTY | \
64 G_REGEX_MATCH_PARTIAL | \
65 G_REGEX_MATCH_NEWLINE_CR | \
66 G_REGEX_MATCH_NEWLINE_LF | \
67 G_REGEX_MATCH_NEWLINE_CRLF | \
68 G_REGEX_MATCH_NEWLINE_ANY)
70 /* if the string is in UTF-8 use g_utf8_ functions, else use
72 #define NEXT_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
73 g_utf8_next_char (s) : \
75 #define PREV_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
76 g_utf8_prev_char (s) : \
81 GRegex *regex; /* the regex */
82 GRegexMatchFlags match_opts; /* options used at match time on the regex */
83 gint matches; /* number of matching sub patterns */
84 gint pos; /* position in the string where last match left off */
85 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
86 gint n_offsets; /* number of offsets */
87 gint *workspace; /* workspace for pcre_dfa_exec() */
88 gint n_workspace; /* number of workspace elements */
89 const gchar *string; /* string passed to the match function */
90 gssize string_len; /* length of string */
95 volatile gint ref_count; /* the ref count for the immutable part */
96 gchar *pattern; /* the pattern */
97 pcre *pcre_re; /* compiled form of the pattern */
98 GRegexCompileFlags compile_opts; /* options used at compile time on the pattern */
99 GRegexMatchFlags match_opts; /* options used at match time on the regex */
100 pcre_extra *extra; /* data stored when G_REGEX_OPTIMIZE is used */
103 /* TRUE if ret is an error code, FALSE otherwise. */
104 #define IS_PCRE_ERROR(ret) ((ret) < PCRE_ERROR_NOMATCH && (ret) != PCRE_ERROR_PARTIAL)
106 typedef struct _InterpolationData InterpolationData;
107 static gboolean interpolation_list_needs_match (GList *list);
108 static gboolean interpolate_replacement (const GMatchInfo *match_info,
111 static GList *split_replacement (const gchar *replacement,
113 static void free_interpolation_data (InterpolationData *data);
117 match_error (gint errcode)
121 case PCRE_ERROR_NOMATCH:
124 case PCRE_ERROR_NULL:
125 /* NULL argument, this should not happen in GRegex */
126 g_warning ("A NULL argument was passed to PCRE");
128 case PCRE_ERROR_BADOPTION:
129 return "bad options";
130 case PCRE_ERROR_BADMAGIC:
131 return _("corrupted object");
132 case PCRE_ERROR_UNKNOWN_OPCODE:
133 return N_("internal error or corrupted object");
134 case PCRE_ERROR_NOMEMORY:
135 return _("out of memory");
136 case PCRE_ERROR_NOSUBSTRING:
137 /* not used by pcre_exec() */
139 case PCRE_ERROR_MATCHLIMIT:
140 return _("backtracking limit reached");
141 case PCRE_ERROR_CALLOUT:
142 /* callouts are not implemented */
144 case PCRE_ERROR_BADUTF8:
145 case PCRE_ERROR_BADUTF8_OFFSET:
146 /* we do not check if strings are valid */
148 case PCRE_ERROR_PARTIAL:
151 case PCRE_ERROR_BADPARTIAL:
152 return _("the pattern contains items not supported for partial matching");
153 case PCRE_ERROR_INTERNAL:
154 return _("internal error");
155 case PCRE_ERROR_BADCOUNT:
156 /* negative ovecsize, this should not happen in GRegex */
157 g_warning ("A negative ovecsize was passed to PCRE");
159 case PCRE_ERROR_DFA_UITEM:
160 return _("the pattern contains items not supported for partial matching");
161 case PCRE_ERROR_DFA_UCOND:
162 return _("back references as conditions are not supported for partial matching");
163 case PCRE_ERROR_DFA_UMLIMIT:
164 /* the match_field field is not used in GRegex */
166 case PCRE_ERROR_DFA_WSSIZE:
167 /* handled expanding the workspace */
169 case PCRE_ERROR_DFA_RECURSE:
170 case PCRE_ERROR_RECURSIONLIMIT:
171 return _("recursion limit reached");
172 case PCRE_ERROR_NULLWSLIMIT:
173 return _("workspace limit for empty substrings reached");
174 case PCRE_ERROR_BADNEWLINE:
175 return _("invalid combination of newline flags");
179 return _("unknown error");
183 translate_compile_error (gint *errcode, const gchar **errmsg)
185 /* Compile errors are created adding 100 to the error code returned
187 * If errcode is known we put the translatable error message in
188 * erromsg. If errcode is unknown we put the generic
189 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
190 * untranslated error message returned by PCRE.
191 * Note that there can be more PCRE errors with the same GRegexError
192 * and that some PCRE errors are useless for us.
198 case G_REGEX_ERROR_STRAY_BACKSLASH:
199 *errmsg = _("\\ at end of pattern");
201 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
202 *errmsg = _("\\c at end of pattern");
204 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
205 *errmsg = _("unrecognized character follows \\");
208 /* A number of Perl escapes are not handled by PCRE.
209 * Therefore it explicitly raises ERR37.
211 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
212 *errmsg = _("case-changing escapes (\\l, \\L, \\u, \\U) are not allowed here");
214 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
215 *errmsg = _("numbers out of order in {} quantifier");
217 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
218 *errmsg = _("number too big in {} quantifier");
220 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
221 *errmsg = _("missing terminating ] for character class");
223 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
224 *errmsg = _("invalid escape sequence in character class");
226 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
227 *errmsg = _("range out of order in character class");
229 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
230 *errmsg = _("nothing to repeat");
232 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
233 *errmsg = _("unrecognized character after (?");
236 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
237 *errmsg = _("unrecognized character after (?<");
240 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
241 *errmsg = _("unrecognized character after (?P");
243 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
244 *errmsg = _("POSIX named classes are supported only within a class");
246 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
247 *errmsg = _("missing terminating )");
250 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
251 *errmsg = _(") without opening (");
254 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
255 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
256 * sequences here, '(?-54' would be an example for the second group.
258 *errmsg = _("(?R or (?[+-]digits must be followed by )");
260 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
261 *errmsg = _("reference to non-existent subpattern");
263 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
264 *errmsg = _("missing ) after comment");
266 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
267 *errmsg = _("regular expression too large");
269 case G_REGEX_ERROR_MEMORY_ERROR:
270 *errmsg = _("failed to get memory");
272 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
273 *errmsg = _("lookbehind assertion is not fixed length");
275 case G_REGEX_ERROR_MALFORMED_CONDITION:
276 *errmsg = _("malformed number or name after (?(");
278 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
279 *errmsg = _("conditional group contains more than two branches");
281 case G_REGEX_ERROR_ASSERTION_EXPECTED:
282 *errmsg = _("assertion expected after (?(");
284 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
285 *errmsg = _("unknown POSIX class name");
287 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
288 *errmsg = _("POSIX collating elements are not supported");
290 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
291 *errmsg = _("character value in \\x{...} sequence is too large");
293 case G_REGEX_ERROR_INVALID_CONDITION:
294 *errmsg = _("invalid condition (?(0)");
296 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
297 *errmsg = _("\\C not allowed in lookbehind assertion");
299 case G_REGEX_ERROR_INFINITE_LOOP:
300 *errmsg = _("recursive call could loop indefinitely");
302 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
303 *errmsg = _("missing terminator in subpattern name");
305 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
306 *errmsg = _("two named subpatterns have the same name");
308 case G_REGEX_ERROR_MALFORMED_PROPERTY:
309 *errmsg = _("malformed \\P or \\p sequence");
311 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
312 *errmsg = _("unknown property name after \\P or \\p");
314 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
315 *errmsg = _("subpattern name is too long (maximum 32 characters)");
317 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
318 *errmsg = _("too many named subpatterns (maximum 10,000)");
320 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
321 *errmsg = _("octal value is greater than \\377");
323 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
324 *errmsg = _("DEFINE group contains more than one branch");
326 case G_REGEX_ERROR_DEFINE_REPETION:
327 *errmsg = _("repeating a DEFINE group is not allowed");
329 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
330 *errmsg = _("inconsistent NEWLINE options");
332 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
333 *errmsg = _("\\g is not followed by a braced name or an optionally "
334 "braced non-zero number");
337 *errcode = G_REGEX_ERROR_INTERNAL;
338 *errmsg = _("unexpected repeat");
341 *errcode = G_REGEX_ERROR_INTERNAL;
342 *errmsg = _("code overflow");
345 *errcode = G_REGEX_ERROR_INTERNAL;
346 *errmsg = _("overran compiling workspace");
349 *errcode = G_REGEX_ERROR_INTERNAL;
350 *errmsg = _("previously-checked referenced subpattern not found");
353 /* This should not happen as we never pass a NULL erroffset */
354 g_warning ("erroffset passed as NULL");
355 *errcode = G_REGEX_ERROR_COMPILE;
358 /* This should not happen as we check options before passing them
359 * to pcre_compile2() */
360 g_warning ("unknown option bit(s) set");
361 *errcode = G_REGEX_ERROR_COMPILE;
366 /* These errors should not happen as we are using an UTF8-enabled PCRE
367 * and we do not check if strings are valid */
368 g_warning ("%s", *errmsg);
369 *errcode = G_REGEX_ERROR_COMPILE;
372 *errcode = G_REGEX_ERROR_COMPILE;
379 match_info_new (const GRegex *regex,
386 GMatchInfo *match_info;
389 string_len = strlen (string);
391 match_info = g_new0 (GMatchInfo, 1);
392 match_info->regex = g_regex_ref ((GRegex *)regex);
393 match_info->string = string;
394 match_info->string_len = string_len;
395 match_info->matches = PCRE_ERROR_NOMATCH;
396 match_info->pos = start_position;
397 match_info->match_opts = match_options;
401 /* These values should be enough for most cases, if they are not
402 * enough g_regex_match_all_full() will expand them. */
403 match_info->n_offsets = 24;
404 match_info->n_workspace = 100;
405 match_info->workspace = g_new (gint, match_info->n_workspace);
410 pcre_fullinfo (regex->pcre_re, regex->extra,
411 PCRE_INFO_CAPTURECOUNT, &capture_count);
412 match_info->n_offsets = (capture_count + 1) * 3;
415 match_info->offsets = g_new0 (gint, match_info->n_offsets);
416 /* Set an invalid position for the previous match. */
417 match_info->offsets[0] = -1;
418 match_info->offsets[1] = -1;
424 * g_match_info_get_regex:
425 * @match_info: a #GMatchInfo
427 * Returns #GRegex object used in @match_info. It belongs to Glib
428 * and must not be freed. Use g_regex_ref() if you need to keep it
429 * after you free @match_info object.
431 * Returns: #GRegex object used in @match_info
436 g_match_info_get_regex (const GMatchInfo *match_info)
438 g_return_val_if_fail (match_info != NULL, NULL);
439 return match_info->regex;
443 * g_match_info_get_string:
444 * @match_info: a #GMatchInfo
446 * Returns the string searched with @match_info. This is the
447 * string passed to g_regex_match() or g_regex_replace() so
448 * you may not free it before calling this function.
450 * Returns: the string searched with @match_info
455 g_match_info_get_string (const GMatchInfo *match_info)
457 g_return_val_if_fail (match_info != NULL, NULL);
458 return match_info->string;
463 * @match_info: a #GMatchInfo
465 * Frees all the memory associated with the #GMatchInfo structure.
470 g_match_info_free (GMatchInfo *match_info)
474 g_regex_unref (match_info->regex);
475 g_free (match_info->offsets);
476 g_free (match_info->workspace);
483 * @match_info: a #GMatchInfo structure
484 * @error: location to store the error occuring, or %NULL to ignore errors
486 * Scans for the next match using the same parameters of the previous
487 * call to g_regex_match_full() or g_regex_match() that returned
490 * The match is done on the string passed to the match function, so you
491 * cannot free it before calling this function.
493 * Returns: %TRUE is the string matched, %FALSE otherwise
498 g_match_info_next (GMatchInfo *match_info,
501 gint prev_match_start;
504 g_return_val_if_fail (match_info != NULL, FALSE);
505 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
506 g_return_val_if_fail (match_info->pos >= 0, FALSE);
508 prev_match_start = match_info->offsets[0];
509 prev_match_end = match_info->offsets[1];
511 match_info->matches = pcre_exec (match_info->regex->pcre_re,
512 match_info->regex->extra,
514 match_info->string_len,
516 match_info->regex->match_opts | match_info->match_opts,
518 match_info->n_offsets);
519 if (IS_PCRE_ERROR (match_info->matches))
521 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
522 _("Error while matching regular expression %s: %s"),
523 match_info->regex->pattern, match_error (match_info->matches));
527 /* avoid infinite loops if the pattern is an empty string or something
529 if (match_info->pos == match_info->offsets[1])
531 if (match_info->pos > match_info->string_len)
533 /* we have reached the end of the string */
534 match_info->pos = -1;
535 match_info->matches = PCRE_ERROR_NOMATCH;
539 match_info->pos = NEXT_CHAR (match_info->regex,
540 &match_info->string[match_info->pos]) -
545 match_info->pos = match_info->offsets[1];
548 /* it's possibile to get two identical matches when we are matching
549 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
550 * the string is "RegExTest" we have:
551 * - search at position 0: match from 0 to 0
552 * - search at position 1: match from 3 to 3
553 * - search at position 3: match from 3 to 3 (duplicate)
554 * - search at position 4: match from 5 to 5
555 * - search at position 5: match from 5 to 5 (duplicate)
556 * - search at position 6: no match -> stop
557 * so we have to ignore the duplicates.
558 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
559 if (match_info->matches >= 0 &&
560 prev_match_start == match_info->offsets[0] &&
561 prev_match_end == match_info->offsets[1])
563 /* ignore this match and search the next one */
564 return g_match_info_next (match_info, error);
567 return match_info->matches >= 0;
571 * g_match_info_matches:
572 * @match_info: a #GMatchInfo structure
574 * Returns whether the previous match operation succeeded.
576 * Returns: %TRUE if the previous match operation succeeded,
582 g_match_info_matches (const GMatchInfo *match_info)
584 g_return_val_if_fail (match_info != NULL, FALSE);
586 return match_info->matches >= 0;
590 * g_match_info_get_match_count:
591 * @match_info: a #GMatchInfo structure
593 * Retrieves the number of matched substrings (including substring 0,
594 * that is the whole matched text), so 1 is returned if the pattern
595 * has no substrings in it and 0 is returned if the match failed.
597 * If the last match was obtained using the DFA algorithm, that is
598 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
599 * count is not that of the number of capturing parentheses but that of
600 * the number of matched substrings.
602 * Returns: Number of matched substrings, or -1 if an error occurred
607 g_match_info_get_match_count (const GMatchInfo *match_info)
609 g_return_val_if_fail (match_info, -1);
611 if (match_info->matches == PCRE_ERROR_NOMATCH)
614 else if (match_info->matches < PCRE_ERROR_NOMATCH)
619 return match_info->matches;
623 * g_match_info_is_partial_match:
624 * @match_info: a #GMatchInfo structure
626 * Usually if the string passed to g_regex_match*() matches as far as
627 * it goes, but is too short to match the entire pattern, %FALSE is
628 * returned. There are circumstances where it might be helpful to
629 * distinguish this case from other cases in which there is no match.
631 * Consider, for example, an application where a human is required to
632 * type in data for a field with specific formatting requirements. An
633 * example might be a date in the form ddmmmyy, defined by the pattern
634 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
635 * If the application sees the user’s keystrokes one by one, and can
636 * check that what has been typed so far is potentially valid, it is
637 * able to raise an error as soon as a mistake is made.
639 * GRegex supports the concept of partial matching by means of the
640 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
641 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
642 * for a complete match, %FALSE otherwise. But, when these functions
643 * return %FALSE, you can check if the match was partial calling
644 * g_match_info_is_partial_match().
646 * When using partial matching you cannot use g_match_info_fetch*().
648 * Because of the way certain internal optimizations are implemented
649 * the partial matching algorithm cannot be used with all patterns.
650 * So repeated single characters such as "a{2,4}" and repeated single
651 * meta-sequences such as "\d+" are not permitted if the maximum number
652 * of occurrences is greater than one. Optional items such as "\d?"
653 * (where the maximum is one) are permitted. Quantifiers with any values
654 * are permitted after parentheses, so the invalid examples above can be
655 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
656 * for a pattern that does not conform to the restrictions, matching
657 * functions return an error.
659 * Returns: %TRUE if the match was partial, %FALSE otherwise
664 g_match_info_is_partial_match (const GMatchInfo *match_info)
666 g_return_val_if_fail (match_info != NULL, FALSE);
668 return match_info->matches == PCRE_ERROR_PARTIAL;
672 * g_match_info_expand_references:
673 * @match_info: a #GMatchInfo or %NULL
674 * @string_to_expand: the string to expand
675 * @error: location to store the error occuring, or %NULL to ignore errors
677 * Returns a new string containing the text in @string_to_expand with
678 * references and escape sequences expanded. References refer to the last
679 * match done with @string against @regex and have the same syntax used by
682 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
683 * passed to g_regex_new().
685 * The backreferences are extracted from the string passed to the match
686 * function, so you cannot call this function after freeing the string.
688 * @match_info may be %NULL in which case @string_to_expand must not
689 * contain references. For instance "foo\n" does not refer to an actual
690 * pattern and '\n' merely will be replaced with \n character,
691 * while to expand "\0" (whole match) one needs the result of a match.
692 * Use g_regex_check_replacement() to find out whether @string_to_expand
693 * contains references.
695 * Returns: the expanded string, or %NULL if an error occurred
700 g_match_info_expand_references (const GMatchInfo *match_info,
701 const gchar *string_to_expand,
706 GError *tmp_error = NULL;
708 g_return_val_if_fail (string_to_expand != NULL, NULL);
709 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
711 list = split_replacement (string_to_expand, &tmp_error);
712 if (tmp_error != NULL)
714 g_propagate_error (error, tmp_error);
718 if (!match_info && interpolation_list_needs_match (list))
720 g_critical ("String '%s' contains references to the match, can't "
721 "expand references without GMatchInfo object",
726 result = g_string_sized_new (strlen (string_to_expand));
727 interpolate_replacement (match_info, result, list);
729 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
732 return g_string_free (result, FALSE);
736 * g_match_info_fetch:
737 * @match_info: #GMatchInfo structure
738 * @match_num: number of the sub expression
740 * Retrieves the text matching the @match_num<!-- -->'th capturing
741 * parentheses. 0 is the full text of the match, 1 is the first paren
742 * set, 2 the second, and so on.
744 * If @match_num is a valid sub pattern but it didn't match anything
745 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
746 * string is returned.
748 * If the match was obtained using the DFA algorithm, that is using
749 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
750 * string is not that of a set of parentheses but that of a matched
751 * substring. Substrings are matched in reverse order of length, so
752 * 0 is the longest match.
754 * The string is fetched from the string passed to the match function,
755 * so you cannot call this function after freeing the string.
757 * Returns: The matched substring, or %NULL if an error occurred.
758 * You have to free the string yourself
763 g_match_info_fetch (const GMatchInfo *match_info,
766 /* we cannot use pcre_get_substring() because it allocates the
767 * string using pcre_malloc(). */
771 g_return_val_if_fail (match_info != NULL, NULL);
772 g_return_val_if_fail (match_num >= 0, NULL);
774 /* match_num does not exist or it didn't matched, i.e. matching "b"
775 * against "(a)?b" then group 0 is empty. */
776 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
778 else if (start == -1)
779 match = g_strdup ("");
781 match = g_strndup (&match_info->string[start], end - start);
787 * g_match_info_fetch_pos:
788 * @match_info: #GMatchInfo structure
789 * @match_num: number of the sub expression
790 * @start_pos: pointer to location where to store the start position
791 * @end_pos: pointer to location where to store the end position
793 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
794 * parentheses. 0 is the full text of the match, 1 is the first
795 * paren set, 2 the second, and so on.
797 * If @match_num is a valid sub pattern but it didn't match anything
798 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
799 * and @end_pos are set to -1 and %TRUE is returned.
801 * If the match was obtained using the DFA algorithm, that is using
802 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
803 * position is not that of a set of parentheses but that of a matched
804 * substring. Substrings are matched in reverse order of length, so
805 * 0 is the longest match.
807 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
808 * the position cannot be fetched, @start_pos and @end_pos are left
814 g_match_info_fetch_pos (const GMatchInfo *match_info,
819 g_return_val_if_fail (match_info != NULL, FALSE);
820 g_return_val_if_fail (match_num >= 0, FALSE);
822 /* make sure the sub expression number they're requesting is less than
823 * the total number of sub expressions that were matched. */
824 if (match_num >= match_info->matches)
827 if (start_pos != NULL)
828 *start_pos = match_info->offsets[2 * match_num];
831 *end_pos = match_info->offsets[2 * match_num + 1];
837 * Returns number of first matched subpattern with name @name.
838 * There may be more than one in case when DUPNAMES is used,
839 * and not all subpatterns with that name match;
840 * pcre_get_stringnumber() does not work in that case.
843 get_matched_substring_number (const GMatchInfo *match_info,
850 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
851 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
853 /* This code is copied from pcre_get.c: get_first_set() */
854 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
862 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
864 gint n = (entry[0] << 8) + entry[1];
865 if (match_info->offsets[n*2] >= 0)
869 return (first[0] << 8) + first[1];
873 * g_match_info_fetch_named:
874 * @match_info: #GMatchInfo structure
875 * @name: name of the subexpression
877 * Retrieves the text matching the capturing parentheses named @name.
879 * If @name is a valid sub pattern name but it didn't match anything
880 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
881 * then an empty string is returned.
883 * The string is fetched from the string passed to the match function,
884 * so you cannot call this function after freeing the string.
886 * Returns: The matched substring, or %NULL if an error occurred.
887 * You have to free the string yourself
892 g_match_info_fetch_named (const GMatchInfo *match_info,
895 /* we cannot use pcre_get_named_substring() because it allocates the
896 * string using pcre_malloc(). */
899 g_return_val_if_fail (match_info != NULL, NULL);
900 g_return_val_if_fail (name != NULL, NULL);
902 num = get_matched_substring_number (match_info, name);
906 return g_match_info_fetch (match_info, num);
910 * g_match_info_fetch_named_pos:
911 * @match_info: #GMatchInfo structure
912 * @name: name of the subexpression
913 * @start_pos: pointer to location where to store the start position
914 * @end_pos: pointer to location where to store the end position
916 * Retrieves the position in bytes of the capturing parentheses named @name.
918 * If @name is a valid sub pattern name but it didn't match anything
919 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
920 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
922 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
923 * the position cannot be fetched, @start_pos and @end_pos are left
929 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
936 g_return_val_if_fail (match_info != NULL, FALSE);
937 g_return_val_if_fail (name != NULL, FALSE);
939 num = get_matched_substring_number (match_info, name);
943 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
947 * g_match_info_fetch_all:
948 * @match_info: a #GMatchInfo structure
950 * Bundles up pointers to each of the matching substrings from a match
951 * and stores them in an array of gchar pointers. The first element in
952 * the returned array is the match number 0, i.e. the entire matched
955 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
956 * "b" against "(a)?b") then an empty string is inserted.
958 * If the last match was obtained using the DFA algorithm, that is using
959 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
960 * strings are not that matched by sets of parentheses but that of the
961 * matched substring. Substrings are matched in reverse order of length,
962 * so the first one is the longest match.
964 * The strings are fetched from the string passed to the match function,
965 * so you cannot call this function after freeing the string.
967 * Returns: a %NULL-terminated array of gchar * pointers. It must be
968 * freed using g_strfreev(). If the previous match failed %NULL is
974 g_match_info_fetch_all (const GMatchInfo *match_info)
976 /* we cannot use pcre_get_substring_list() because the returned value
977 * isn't suitable for g_strfreev(). */
981 g_return_val_if_fail (match_info != NULL, NULL);
983 if (match_info->matches < 0)
986 result = g_new (gchar *, match_info->matches + 1);
987 for (i = 0; i < match_info->matches; i++)
988 result[i] = g_match_info_fetch (match_info, i);
998 g_regex_error_quark (void)
1000 static GQuark error_quark = 0;
1002 if (error_quark == 0)
1003 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1012 * Increases reference count of @regex by 1.
1019 g_regex_ref (GRegex *regex)
1021 g_return_val_if_fail (regex != NULL, NULL);
1022 g_atomic_int_inc (®ex->ref_count);
1030 * Decreases reference count of @regex by 1. When reference count drops
1031 * to zero, it frees all the memory associated with the regex structure.
1036 g_regex_unref (GRegex *regex)
1038 g_return_if_fail (regex != NULL);
1040 if (g_atomic_int_exchange_and_add (®ex->ref_count, -1) - 1 == 0)
1042 g_free (regex->pattern);
1043 if (regex->pcre_re != NULL)
1044 pcre_free (regex->pcre_re);
1045 if (regex->extra != NULL)
1046 pcre_free (regex->extra);
1053 * @pattern: the regular expression
1054 * @compile_options: compile options for the regular expression, or 0
1055 * @match_options: match options for the regular expression, or 0
1056 * @error: return location for a #GError
1058 * Compiles the regular expression to an internal form, and does
1059 * the initial setup of the #GRegex structure.
1061 * Returns: a #GRegex structure. Call g_regex_unref() when you
1067 g_regex_new (const gchar *pattern,
1068 GRegexCompileFlags compile_options,
1069 GRegexMatchFlags match_options,
1074 const gchar *errmsg;
1077 gboolean optimize = FALSE;
1078 static gboolean initialized = FALSE;
1079 unsigned long int pcre_compile_options;
1081 g_return_val_if_fail (pattern != NULL, NULL);
1082 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1083 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1084 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1091 pcre_config (PCRE_CONFIG_UTF8, &support);
1094 msg = N_("PCRE library is compiled without UTF8 support");
1095 g_critical ("%s", msg);
1096 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1100 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1103 msg = N_("PCRE library is compiled without UTF8 properties support");
1104 g_critical ("%s", msg);
1105 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1112 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1113 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1114 if (compile_options & G_REGEX_OPTIMIZE)
1117 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1118 * instead uses UTF-8 only if required with PCRE_UTF8. */
1119 if (compile_options & G_REGEX_RAW)
1122 compile_options &= ~G_REGEX_RAW;
1127 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1128 match_options |= PCRE_NO_UTF8_CHECK;
1131 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1132 * not for the system one. */
1133 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1134 !(compile_options & G_REGEX_NEWLINE_LF))
1136 compile_options |= PCRE_NEWLINE_ANY;
1139 /* compile the pattern */
1140 re = pcre_compile2 (pattern, compile_options, &errcode,
1141 &errmsg, &erroffset, NULL);
1143 /* if the compilation failed, set the error member and return
1149 /* Translate the PCRE error code to GRegexError and use a translated
1150 * error message if possible */
1151 translate_compile_error (&errcode, &errmsg);
1153 /* PCRE uses byte offsets but we want to show character offsets */
1154 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1156 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1157 _("Error while compiling regular "
1158 "expression %s at char %d: %s"),
1159 pattern, erroffset, errmsg);
1160 g_propagate_error (error, tmp_error);
1165 /* For options set at the beginning of the pattern, pcre puts them into
1166 * compile options, e.g. "(?i)foo" will make the pcre structure store
1167 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1168 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1169 compile_options = pcre_compile_options;
1171 if (!(compile_options & G_REGEX_DUPNAMES))
1173 gboolean jchanged = FALSE;
1174 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1176 compile_options |= G_REGEX_DUPNAMES;
1179 regex = g_new0 (GRegex, 1);
1180 regex->ref_count = 1;
1181 regex->pattern = g_strdup (pattern);
1182 regex->pcre_re = re;
1183 regex->compile_opts = compile_options;
1184 regex->match_opts = match_options;
1188 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1191 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1192 G_REGEX_ERROR_OPTIMIZE,
1193 _("Error while optimizing "
1194 "regular expression %s: %s"),
1197 g_propagate_error (error, tmp_error);
1199 g_regex_unref (regex);
1208 * g_regex_get_pattern:
1209 * @regex: a #GRegex structure
1211 * Gets the pattern string associated with @regex, i.e. a copy of
1212 * the string passed to g_regex_new().
1214 * Returns: the pattern of @regex
1219 g_regex_get_pattern (const GRegex *regex)
1221 g_return_val_if_fail (regex != NULL, NULL);
1223 return regex->pattern;
1227 * g_regex_get_max_backref:
1230 * Returns the number of the highest back reference
1231 * in the pattern, or 0 if the pattern does not contain
1234 * Returns: the number of the highest back reference
1239 g_regex_get_max_backref (const GRegex *regex)
1243 pcre_fullinfo (regex->pcre_re, regex->extra,
1244 PCRE_INFO_BACKREFMAX, &value);
1250 * g_regex_get_capture_count:
1253 * Returns the number of capturing subpatterns in the pattern.
1255 * Returns: the number of capturing subpatterns
1260 g_regex_get_capture_count (const GRegex *regex)
1264 pcre_fullinfo (regex->pcre_re, regex->extra,
1265 PCRE_INFO_CAPTURECOUNT, &value);
1271 * g_regex_match_simple:
1272 * @pattern: the regular expression
1273 * @string: the string to scan for matches
1274 * @compile_options: compile options for the regular expression, or 0
1275 * @match_options: match options, or 0
1277 * Scans for a match in @string for @pattern.
1279 * This function is equivalent to g_regex_match() but it does not
1280 * require to compile the pattern with g_regex_new(), avoiding some
1281 * lines of code when you need just to do a match without extracting
1282 * substrings, capture counts, and so on.
1284 * If this function is to be called on the same @pattern more than
1285 * once, it's more efficient to compile the pattern once with
1286 * g_regex_new() and then use g_regex_match().
1288 * Returns: %TRUE if the string matched, %FALSE otherwise
1293 g_regex_match_simple (const gchar *pattern,
1294 const gchar *string,
1295 GRegexCompileFlags compile_options,
1296 GRegexMatchFlags match_options)
1301 regex = g_regex_new (pattern, compile_options, 0, NULL);
1304 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1305 g_regex_unref (regex);
1311 * @regex: a #GRegex structure from g_regex_new()
1312 * @string: the string to scan for matches
1313 * @match_options: match options
1314 * @match_info: pointer to location where to store the #GMatchInfo,
1315 * or %NULL if you do not need it
1317 * Scans for a match in string for the pattern in @regex.
1318 * The @match_options are combined with the match options specified
1319 * when the @regex structure was created, letting you have more
1320 * flexibility in reusing #GRegex structures.
1322 * A #GMatchInfo structure, used to get information on the match,
1323 * is stored in @match_info if not %NULL. Note that if @match_info
1324 * is not %NULL then it is created even if the function returns %FALSE,
1325 * i.e. you must free it regardless if regular expression actually matched.
1327 * To retrieve all the non-overlapping matches of the pattern in
1328 * string you can use g_match_info_next().
1332 * print_uppercase_words (const gchar *string)
1334 * /* Print all uppercase-only words. */
1336 * GMatchInfo *match_info;
1338 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1339 * g_regex_match (regex, string, 0, &match_info);
1340 * while (g_match_info_matches (match_info))
1342 * gchar *word = g_match_info_fetch (match_info, 0);
1343 * g_print ("Found: %s\n", word);
1345 * g_match_info_next (match_info, NULL);
1347 * g_match_info_free (match_info);
1348 * g_regex_unref (regex);
1352 * @string is not copied and is used in #GMatchInfo internally. If
1353 * you use any #GMatchInfo method (except g_match_info_free()) after
1354 * freeing or modifying @string then the behaviour is undefined.
1356 * Returns: %TRUE is the string matched, %FALSE otherwise
1361 g_regex_match (const GRegex *regex,
1362 const gchar *string,
1363 GRegexMatchFlags match_options,
1364 GMatchInfo **match_info)
1366 return g_regex_match_full (regex, string, -1, 0, match_options,
1371 * g_regex_match_full:
1372 * @regex: a #GRegex structure from g_regex_new()
1373 * @string: the string to scan for matches
1374 * @string_len: the length of @string, or -1 if @string is nul-terminated
1375 * @start_position: starting index of the string to match
1376 * @match_options: match options
1377 * @match_info: pointer to location where to store the #GMatchInfo,
1378 * or %NULL if you do not need it
1379 * @error: location to store the error occuring, or %NULL to ignore errors
1381 * Scans for a match in string for the pattern in @regex.
1382 * The @match_options are combined with the match options specified
1383 * when the @regex structure was created, letting you have more
1384 * flexibility in reusing #GRegex structures.
1386 * Setting @start_position differs from just passing over a shortened
1387 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1388 * that begins with any kind of lookbehind assertion, such as "\b".
1390 * A #GMatchInfo structure, used to get information on the match, is
1391 * stored in @match_info if not %NULL. Note that if @match_info is
1392 * not %NULL then it is created even if the function returns %FALSE,
1393 * i.e. you must free it regardless if regular expression actually
1396 * @string is not copied and is used in #GMatchInfo internally. If
1397 * you use any #GMatchInfo method (except g_match_info_free()) after
1398 * freeing or modifying @string then the behaviour is undefined.
1400 * To retrieve all the non-overlapping matches of the pattern in
1401 * string you can use g_match_info_next().
1405 * print_uppercase_words (const gchar *string)
1407 * /* Print all uppercase-only words. */
1409 * GMatchInfo *match_info;
1410 * GError *error = NULL;
1412 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1413 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1414 * while (g_match_info_matches (match_info))
1416 * gchar *word = g_match_info_fetch (match_info, 0);
1417 * g_print ("Found: %s\n", word);
1419 * g_match_info_next (match_info, &error);
1421 * g_match_info_free (match_info);
1422 * g_regex_unref (regex);
1423 * if (error != NULL)
1425 * g_printerr ("Error while matching: %s\n", error->message);
1426 * g_error_free (error);
1431 * Returns: %TRUE is the string matched, %FALSE otherwise
1436 g_regex_match_full (const GRegex *regex,
1437 const gchar *string,
1439 gint start_position,
1440 GRegexMatchFlags match_options,
1441 GMatchInfo **match_info,
1447 g_return_val_if_fail (regex != NULL, FALSE);
1448 g_return_val_if_fail (string != NULL, FALSE);
1449 g_return_val_if_fail (start_position >= 0, FALSE);
1450 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1451 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1453 info = match_info_new (regex, string, string_len, start_position,
1454 match_options, FALSE);
1455 match_ok = g_match_info_next (info, error);
1456 if (match_info != NULL)
1459 g_match_info_free (info);
1465 * g_regex_match_all:
1466 * @regex: a #GRegex structure from g_regex_new()
1467 * @string: the string to scan for matches
1468 * @match_options: match options
1469 * @match_info: pointer to location where to store the #GMatchInfo,
1470 * or %NULL if you do not need it
1472 * Using the standard algorithm for regular expression matching only
1473 * the longest match in the string is retrieved. This function uses
1474 * a different algorithm so it can retrieve all the possible matches.
1475 * For more documentation see g_regex_match_all_full().
1477 * A #GMatchInfo structure, used to get information on the match, is
1478 * stored in @match_info if not %NULL. Note that if @match_info is
1479 * not %NULL then it is created even if the function returns %FALSE,
1480 * i.e. you must free it regardless if regular expression actually
1483 * @string is not copied and is used in #GMatchInfo internally. If
1484 * you use any #GMatchInfo method (except g_match_info_free()) after
1485 * freeing or modifying @string then the behaviour is undefined.
1487 * Returns: %TRUE is the string matched, %FALSE otherwise
1492 g_regex_match_all (const GRegex *regex,
1493 const gchar *string,
1494 GRegexMatchFlags match_options,
1495 GMatchInfo **match_info)
1497 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1502 * g_regex_match_all_full:
1503 * @regex: a #GRegex structure from g_regex_new()
1504 * @string: the string to scan for matches
1505 * @string_len: the length of @string, or -1 if @string is nul-terminated
1506 * @start_position: starting index of the string to match
1507 * @match_options: match options
1508 * @match_info: pointer to location where to store the #GMatchInfo,
1509 * or %NULL if you do not need it
1510 * @error: location to store the error occuring, or %NULL to ignore errors
1512 * Using the standard algorithm for regular expression matching only
1513 * the longest match in the string is retrieved, it is not possibile
1514 * to obtain all the available matches. For instance matching
1515 * "<a> <b> <c>" against the pattern "<.*>"
1516 * you get "<a> <b> <c>".
1518 * This function uses a different algorithm (called DFA, i.e. deterministic
1519 * finite automaton), so it can retrieve all the possible matches, all
1520 * starting at the same point in the string. For instance matching
1521 * "<a> <b> <c>" against the pattern "<.*>"
1522 * you would obtain three matches: "<a> <b> <c>",
1523 * "<a> <b>" and "<a>".
1525 * The number of matched strings is retrieved using
1526 * g_match_info_get_match_count(). To obtain the matched strings and
1527 * their position you can use, respectively, g_match_info_fetch() and
1528 * g_match_info_fetch_pos(). Note that the strings are returned in
1529 * reverse order of length; that is, the longest matching string is
1532 * Note that the DFA algorithm is slower than the standard one and it
1533 * is not able to capture substrings, so backreferences do not work.
1535 * Setting @start_position differs from just passing over a shortened
1536 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1537 * that begins with any kind of lookbehind assertion, such as "\b".
1539 * A #GMatchInfo structure, used to get information on the match, is
1540 * stored in @match_info if not %NULL. Note that if @match_info is
1541 * not %NULL then it is created even if the function returns %FALSE,
1542 * i.e. you must free it regardless if regular expression actually
1545 * @string is not copied and is used in #GMatchInfo internally. If
1546 * you use any #GMatchInfo method (except g_match_info_free()) after
1547 * freeing or modifying @string then the behaviour is undefined.
1549 * Returns: %TRUE is the string matched, %FALSE otherwise
1554 g_regex_match_all_full (const GRegex *regex,
1555 const gchar *string,
1557 gint start_position,
1558 GRegexMatchFlags match_options,
1559 GMatchInfo **match_info,
1565 g_return_val_if_fail (regex != NULL, FALSE);
1566 g_return_val_if_fail (string != NULL, FALSE);
1567 g_return_val_if_fail (start_position >= 0, FALSE);
1568 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1569 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1571 info = match_info_new (regex, string, string_len, start_position,
1572 match_options, TRUE);
1578 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1579 info->string, info->string_len,
1581 regex->match_opts | match_options,
1582 info->offsets, info->n_offsets,
1583 info->workspace, info->n_workspace);
1584 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1586 /* info->workspace is too small. */
1587 info->n_workspace *= 2;
1588 info->workspace = g_realloc (info->workspace,
1589 info->n_workspace * sizeof (gint));
1592 else if (info->matches == 0)
1594 /* info->offsets is too small. */
1595 info->n_offsets *= 2;
1596 info->offsets = g_realloc (info->offsets,
1597 info->n_offsets * sizeof (gint));
1600 else if (IS_PCRE_ERROR (info->matches))
1602 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1603 _("Error while matching regular expression %s: %s"),
1604 regex->pattern, match_error (info->matches));
1608 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1611 if (match_info != NULL)
1614 g_match_info_free (info);
1616 return info->matches >= 0;
1620 * g_regex_get_string_number:
1621 * @regex: #GRegex structure
1622 * @name: name of the subexpression
1624 * Retrieves the number of the subexpression named @name.
1626 * Returns: The number of the subexpression or -1 if @name
1632 g_regex_get_string_number (const GRegex *regex,
1637 g_return_val_if_fail (regex != NULL, -1);
1638 g_return_val_if_fail (name != NULL, -1);
1640 num = pcre_get_stringnumber (regex->pcre_re, name);
1641 if (num == PCRE_ERROR_NOSUBSTRING)
1648 * g_regex_split_simple:
1649 * @pattern: the regular expression
1650 * @string: the string to scan for matches
1651 * @compile_options: compile options for the regular expression, or 0
1652 * @match_options: match options, or 0
1654 * Breaks the string on the pattern, and returns an array of
1655 * the tokens. If the pattern contains capturing parentheses,
1656 * then the text for each of the substrings will also be returned.
1657 * If the pattern does not match anywhere in the string, then the
1658 * whole string is returned as the first token.
1660 * This function is equivalent to g_regex_split() but it does
1661 * not require to compile the pattern with g_regex_new(), avoiding
1662 * some lines of code when you need just to do a split without
1663 * extracting substrings, capture counts, and so on.
1665 * If this function is to be called on the same @pattern more than
1666 * once, it's more efficient to compile the pattern once with
1667 * g_regex_new() and then use g_regex_split().
1669 * As a special case, the result of splitting the empty string ""
1670 * is an empty vector, not a vector containing a single string.
1671 * The reason for this special case is that being able to represent
1672 * a empty vector is typically more useful than consistent handling
1673 * of empty elements. If you do need to represent empty elements,
1674 * you'll need to check for the empty string before calling this
1677 * A pattern that can match empty strings splits @string into
1678 * separate characters wherever it matches the empty string between
1679 * characters. For example splitting "ab c" using as a separator
1680 * "\s*", you will get "a", "b" and "c".
1682 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1687 g_regex_split_simple (const gchar *pattern,
1688 const gchar *string,
1689 GRegexCompileFlags compile_options,
1690 GRegexMatchFlags match_options)
1695 regex = g_regex_new (pattern, compile_options, 0, NULL);
1698 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1699 g_regex_unref (regex);
1705 * @regex: a #GRegex structure
1706 * @string: the string to split with the pattern
1707 * @match_options: match time option flags
1709 * Breaks the string on the pattern, and returns an array of the tokens.
1710 * If the pattern contains capturing parentheses, then the text for each
1711 * of the substrings will also be returned. If the pattern does not match
1712 * anywhere in the string, then the whole string is returned as the first
1715 * As a special case, the result of splitting the empty string "" is an
1716 * empty vector, not a vector containing a single string. The reason for
1717 * this special case is that being able to represent a empty vector is
1718 * typically more useful than consistent handling of empty elements. If
1719 * you do need to represent empty elements, you'll need to check for the
1720 * empty string before calling this function.
1722 * A pattern that can match empty strings splits @string into separate
1723 * characters wherever it matches the empty string between characters.
1724 * For example splitting "ab c" using as a separator "\s*", you will get
1727 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1732 g_regex_split (const GRegex *regex,
1733 const gchar *string,
1734 GRegexMatchFlags match_options)
1736 return g_regex_split_full (regex, string, -1, 0,
1737 match_options, 0, NULL);
1741 * g_regex_split_full:
1742 * @regex: a #GRegex structure
1743 * @string: the string to split with the pattern
1744 * @string_len: the length of @string, or -1 if @string is nul-terminated
1745 * @start_position: starting index of the string to match
1746 * @match_options: match time option flags
1747 * @max_tokens: the maximum number of tokens to split @string into.
1748 * If this is less than 1, the string is split completely
1749 * @error: return location for a #GError
1751 * Breaks the string on the pattern, and returns an array of the tokens.
1752 * If the pattern contains capturing parentheses, then the text for each
1753 * of the substrings will also be returned. If the pattern does not match
1754 * anywhere in the string, then the whole string is returned as the first
1757 * As a special case, the result of splitting the empty string "" is an
1758 * empty vector, not a vector containing a single string. The reason for
1759 * this special case is that being able to represent a empty vector is
1760 * typically more useful than consistent handling of empty elements. If
1761 * you do need to represent empty elements, you'll need to check for the
1762 * empty string before calling this function.
1764 * A pattern that can match empty strings splits @string into separate
1765 * characters wherever it matches the empty string between characters.
1766 * For example splitting "ab c" using as a separator "\s*", you will get
1769 * Setting @start_position differs from just passing over a shortened
1770 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1771 * that begins with any kind of lookbehind assertion, such as "\b".
1773 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1778 g_regex_split_full (const GRegex *regex,
1779 const gchar *string,
1781 gint start_position,
1782 GRegexMatchFlags match_options,
1786 GError *tmp_error = NULL;
1787 GMatchInfo *match_info;
1792 /* position of the last separator. */
1793 gint last_separator_end;
1794 /* was the last match 0 bytes long? */
1795 gboolean last_match_is_empty;
1796 /* the returned array of char **s */
1797 gchar **string_list;
1799 g_return_val_if_fail (regex != NULL, NULL);
1800 g_return_val_if_fail (string != NULL, NULL);
1801 g_return_val_if_fail (start_position >= 0, NULL);
1802 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1803 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1805 if (max_tokens <= 0)
1806 max_tokens = G_MAXINT;
1809 string_len = strlen (string);
1811 /* zero-length string */
1812 if (string_len - start_position == 0)
1813 return g_new0 (gchar *, 1);
1815 if (max_tokens == 1)
1817 string_list = g_new0 (gchar *, 2);
1818 string_list[0] = g_strndup (&string[start_position],
1819 string_len - start_position);
1825 last_separator_end = start_position;
1826 last_match_is_empty = FALSE;
1828 match_ok = g_regex_match_full (regex, string, string_len, start_position,
1829 match_options, &match_info, &tmp_error);
1830 while (tmp_error == NULL)
1834 last_match_is_empty =
1835 (match_info->offsets[0] == match_info->offsets[1]);
1837 /* we need to skip empty separators at the same position of the end
1838 * of another separator. e.g. the string is "a b" and the separator
1839 * is " *", so from 1 to 2 we have a match and at position 2 we have
1840 * an empty match. */
1841 if (last_separator_end != match_info->offsets[1])
1846 token = g_strndup (string + last_separator_end,
1847 match_info->offsets[0] - last_separator_end);
1848 list = g_list_prepend (list, token);
1851 /* if there were substrings, these need to be added to
1853 match_count = g_match_info_get_match_count (match_info);
1854 if (match_count > 1)
1856 for (i = 1; i < match_count; i++)
1857 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
1863 /* if there was no match, copy to end of string. */
1864 if (!last_match_is_empty)
1866 gchar *token = g_strndup (string + last_separator_end,
1867 match_info->string_len - last_separator_end);
1868 list = g_list_prepend (list, token);
1870 /* no more tokens, end the loop. */
1874 /* -1 to leave room for the last part. */
1875 if (token_count >= max_tokens - 1)
1877 /* we have reached the maximum number of tokens, so we copy
1878 * the remaining part of the string. */
1879 if (last_match_is_empty)
1881 /* the last match was empty, so we have moved one char
1882 * after the real position to avoid empty matches at the
1884 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
1886 /* the if is needed in the case we have terminated the available
1887 * tokens, but we are at the end of the string, so there are no
1888 * characters left to copy. */
1889 if (string_len > match_info->pos)
1891 gchar *token = g_strndup (string + match_info->pos,
1892 string_len - match_info->pos);
1893 list = g_list_prepend (list, token);
1899 last_separator_end = match_info->pos;
1900 if (last_match_is_empty)
1901 /* if the last match was empty, g_match_info_next() has moved
1902 * forward to avoid infinite loops, but we still need to copy that
1904 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
1906 match_ok = g_match_info_next (match_info, &tmp_error);
1908 g_match_info_free (match_info);
1909 if (tmp_error != NULL)
1911 g_propagate_error (error, tmp_error);
1912 g_list_foreach (list, (GFunc)g_free, NULL);
1914 match_info->pos = -1;
1918 string_list = g_new (gchar *, g_list_length (list) + 1);
1920 for (last = g_list_last (list); last; last = g_list_previous (last))
1921 string_list[i++] = last->data;
1922 string_list[i] = NULL;
1931 REPL_TYPE_CHARACTER,
1932 REPL_TYPE_SYMBOLIC_REFERENCE,
1933 REPL_TYPE_NUMERIC_REFERENCE,
1934 REPL_TYPE_CHANGE_CASE
1939 CHANGE_CASE_NONE = 1 << 0,
1940 CHANGE_CASE_UPPER = 1 << 1,
1941 CHANGE_CASE_LOWER = 1 << 2,
1942 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
1943 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
1944 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
1945 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
1946 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
1949 struct _InterpolationData
1955 ChangeCase change_case;
1959 free_interpolation_data (InterpolationData *data)
1961 g_free (data->text);
1965 static const gchar *
1966 expand_escape (const gchar *replacement,
1968 InterpolationData *data,
1973 const gchar *error_detail;
1975 GError *tmp_error = NULL;
1983 data->type = REPL_TYPE_CHARACTER;
1988 data->type = REPL_TYPE_CHARACTER;
1993 data->type = REPL_TYPE_CHARACTER;
1998 data->type = REPL_TYPE_CHARACTER;
2003 data->type = REPL_TYPE_CHARACTER;
2008 data->type = REPL_TYPE_CHARACTER;
2013 data->type = REPL_TYPE_CHARACTER;
2018 data->type = REPL_TYPE_CHARACTER;
2028 h = g_ascii_xdigit_value (*p);
2031 error_detail = _("hexadecimal digit or '}' expected");
2042 for (i = 0; i < 2; i++)
2044 h = g_ascii_xdigit_value (*p);
2047 error_detail = _("hexadecimal digit expected");
2054 data->type = REPL_TYPE_STRING;
2055 data->text = g_new0 (gchar, 8);
2056 g_unichar_to_utf8 (x, data->text);
2060 data->type = REPL_TYPE_CHANGE_CASE;
2061 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2065 data->type = REPL_TYPE_CHANGE_CASE;
2066 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2070 data->type = REPL_TYPE_CHANGE_CASE;
2071 data->change_case = CHANGE_CASE_LOWER;
2075 data->type = REPL_TYPE_CHANGE_CASE;
2076 data->change_case = CHANGE_CASE_UPPER;
2080 data->type = REPL_TYPE_CHANGE_CASE;
2081 data->change_case = CHANGE_CASE_NONE;
2087 error_detail = _("missing '<' in symbolic reference");
2096 error_detail = _("unfinished symbolic reference");
2103 error_detail = _("zero-length symbolic reference");
2106 if (g_ascii_isdigit (*q))
2111 h = g_ascii_digit_value (*q);
2114 error_detail = _("digit expected");
2123 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2130 if (!g_ascii_isalnum (*r))
2132 error_detail = _("illegal symbolic reference");
2139 data->text = g_strndup (q, p - q);
2140 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2145 /* if \0 is followed by a number is an octal number representing a
2146 * character, else it is a numeric reference. */
2147 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2150 p = g_utf8_next_char (p);
2163 for (i = 0; i < 3; i++)
2165 h = g_ascii_digit_value (*p);
2175 if (i == 2 && base == 10)
2181 if (base == 8 || i == 3)
2183 data->type = REPL_TYPE_STRING;
2184 data->text = g_new0 (gchar, 8);
2185 g_unichar_to_utf8 (x, data->text);
2189 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2194 error_detail = _("stray final '\\'");
2198 error_detail = _("unknown escape sequence");
2205 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2206 tmp_error = g_error_new (G_REGEX_ERROR,
2207 G_REGEX_ERROR_REPLACE,
2208 _("Error while parsing replacement "
2209 "text \"%s\" at char %lu: %s"),
2211 (gulong)(p - replacement),
2213 g_propagate_error (error, tmp_error);
2219 split_replacement (const gchar *replacement,
2223 InterpolationData *data;
2224 const gchar *p, *start;
2226 start = p = replacement;
2231 data = g_new0 (InterpolationData, 1);
2232 start = p = expand_escape (replacement, p, data, error);
2235 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2237 free_interpolation_data (data);
2241 list = g_list_prepend (list, data);
2246 if (*p == '\\' || *p == '\0')
2250 data = g_new0 (InterpolationData, 1);
2251 data->text = g_strndup (start, p - start);
2252 data->type = REPL_TYPE_STRING;
2253 list = g_list_prepend (list, data);
2259 return g_list_reverse (list);
2262 /* Change the case of c based on change_case. */
2263 #define CHANGE_CASE(c, change_case) \
2264 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2265 g_unichar_tolower (c) : \
2266 g_unichar_toupper (c))
2269 string_append (GString *string,
2271 ChangeCase *change_case)
2275 if (text[0] == '\0')
2278 if (*change_case == CHANGE_CASE_NONE)
2280 g_string_append (string, text);
2282 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2284 c = g_utf8_get_char (text);
2285 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2286 g_string_append (string, g_utf8_next_char (text));
2287 *change_case = CHANGE_CASE_NONE;
2291 while (*text != '\0')
2293 c = g_utf8_get_char (text);
2294 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2295 text = g_utf8_next_char (text);
2301 interpolate_replacement (const GMatchInfo *match_info,
2306 InterpolationData *idata;
2308 ChangeCase change_case = CHANGE_CASE_NONE;
2310 for (list = data; list; list = list->next)
2313 switch (idata->type)
2315 case REPL_TYPE_STRING:
2316 string_append (result, idata->text, &change_case);
2318 case REPL_TYPE_CHARACTER:
2319 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2320 if (change_case & CHANGE_CASE_SINGLE_MASK)
2321 change_case = CHANGE_CASE_NONE;
2323 case REPL_TYPE_NUMERIC_REFERENCE:
2324 match = g_match_info_fetch (match_info, idata->num);
2327 string_append (result, match, &change_case);
2331 case REPL_TYPE_SYMBOLIC_REFERENCE:
2332 match = g_match_info_fetch_named (match_info, idata->text);
2335 string_append (result, match, &change_case);
2339 case REPL_TYPE_CHANGE_CASE:
2340 change_case = idata->change_case;
2348 /* whether actual match_info is needed for replacement, i.e.
2349 * whether there are references
2352 interpolation_list_needs_match (GList *list)
2354 while (list != NULL)
2356 InterpolationData *data = list->data;
2358 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2359 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2372 * @regex: a #GRegex structure
2373 * @string: the string to perform matches against
2374 * @string_len: the length of @string, or -1 if @string is nul-terminated
2375 * @start_position: starting index of the string to match
2376 * @replacement: text to replace each match with
2377 * @match_options: options for the match
2378 * @error: location to store the error occuring, or %NULL to ignore errors
2380 * Replaces all occurrences of the pattern in @regex with the
2381 * replacement text. Backreferences of the form '\number' or
2382 * '\g<number>' in the replacement text are interpolated by the
2383 * number-th captured subexpression of the match, '\g<name>' refers
2384 * to the captured subexpression with the given name. '\0' refers to the
2385 * complete match, but '\0' followed by a number is the octal representation
2386 * of a character. To include a literal '\' in the replacement, write '\\'.
2387 * There are also escapes that changes the case of the following text:
2390 * <varlistentry><term>\l</term>
2392 * <para>Convert to lower case the next character</para>
2395 * <varlistentry><term>\u</term>
2397 * <para>Convert to upper case the next character</para>
2400 * <varlistentry><term>\L</term>
2402 * <para>Convert to lower case till \E</para>
2405 * <varlistentry><term>\U</term>
2407 * <para>Convert to upper case till \E</para>
2410 * <varlistentry><term>\E</term>
2412 * <para>End case modification</para>
2417 * If you do not need to use backreferences use g_regex_replace_literal().
2419 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2420 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2421 * you can use g_regex_replace_literal().
2423 * Setting @start_position differs from just passing over a shortened
2424 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2425 * begins with any kind of lookbehind assertion, such as "\b".
2427 * Returns: a newly allocated string containing the replacements
2432 g_regex_replace (const GRegex *regex,
2433 const gchar *string,
2435 gint start_position,
2436 const gchar *replacement,
2437 GRegexMatchFlags match_options,
2442 GError *tmp_error = NULL;
2444 g_return_val_if_fail (regex != NULL, NULL);
2445 g_return_val_if_fail (string != NULL, NULL);
2446 g_return_val_if_fail (start_position >= 0, NULL);
2447 g_return_val_if_fail (replacement != NULL, NULL);
2448 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2449 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2451 list = split_replacement (replacement, &tmp_error);
2452 if (tmp_error != NULL)
2454 g_propagate_error (error, tmp_error);
2458 result = g_regex_replace_eval (regex,
2459 string, string_len, start_position,
2461 interpolate_replacement,
2464 if (tmp_error != NULL)
2465 g_propagate_error (error, tmp_error);
2467 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2474 literal_replacement (const GMatchInfo *match_info,
2478 g_string_append (result, data);
2483 * g_regex_replace_literal:
2484 * @regex: a #GRegex structure
2485 * @string: the string to perform matches against
2486 * @string_len: the length of @string, or -1 if @string is nul-terminated
2487 * @start_position: starting index of the string to match
2488 * @replacement: text to replace each match with
2489 * @match_options: options for the match
2490 * @error: location to store the error occuring, or %NULL to ignore errors
2492 * Replaces all occurrences of the pattern in @regex with the
2493 * replacement text. @replacement is replaced literally, to
2494 * include backreferences use g_regex_replace().
2496 * Setting @start_position differs from just passing over a
2497 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2498 * case of a pattern that begins with any kind of lookbehind
2499 * assertion, such as "\b".
2501 * Returns: a newly allocated string containing the replacements
2506 g_regex_replace_literal (const GRegex *regex,
2507 const gchar *string,
2509 gint start_position,
2510 const gchar *replacement,
2511 GRegexMatchFlags match_options,
2514 g_return_val_if_fail (replacement != NULL, NULL);
2515 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2517 return g_regex_replace_eval (regex,
2518 string, string_len, start_position,
2520 literal_replacement,
2521 (gpointer)replacement,
2526 * g_regex_replace_eval:
2527 * @regex: a #GRegex structure from g_regex_new()
2528 * @string: string to perform matches against
2529 * @string_len: the length of @string, or -1 if @string is nul-terminated
2530 * @start_position: starting index of the string to match
2531 * @match_options: options for the match
2532 * @eval: a function to call for each match
2533 * @user_data: user data to pass to the function
2534 * @error: location to store the error occuring, or %NULL to ignore errors
2536 * Replaces occurrences of the pattern in regex with the output of
2537 * @eval for that occurrence.
2539 * Setting @start_position differs from just passing over a shortened
2540 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2541 * that begins with any kind of lookbehind assertion, such as "\b".
2543 * The following example uses g_regex_replace_eval() to replace multiple
2547 * eval_cb (const GMatchInfo *info,
2554 * match = g_match_info_fetch (info, 0);
2555 * r = g_hash_table_lookup ((GHashTable *)data, match);
2556 * g_string_append (res, r);
2568 * h = g_hash_table_new (g_str_hash, g_str_equal);
2570 * g_hash_table_insert (h, "1", "ONE");
2571 * g_hash_table_insert (h, "2", "TWO");
2572 * g_hash_table_insert (h, "3", "THREE");
2573 * g_hash_table_insert (h, "4", "FOUR");
2575 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2576 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2577 * g_hash_table_destroy (h);
2582 * Returns: a newly allocated string containing the replacements
2587 g_regex_replace_eval (const GRegex *regex,
2588 const gchar *string,
2590 gint start_position,
2591 GRegexMatchFlags match_options,
2592 GRegexEvalCallback eval,
2596 GMatchInfo *match_info;
2599 gboolean done = FALSE;
2600 GError *tmp_error = NULL;
2602 g_return_val_if_fail (regex != NULL, NULL);
2603 g_return_val_if_fail (string != NULL, NULL);
2604 g_return_val_if_fail (start_position >= 0, NULL);
2605 g_return_val_if_fail (eval != NULL, NULL);
2606 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2609 string_len = strlen (string);
2611 result = g_string_sized_new (string_len);
2613 /* run down the string making matches. */
2614 g_regex_match_full (regex, string, string_len, start_position,
2615 match_options, &match_info, &tmp_error);
2616 while (!done && g_match_info_matches (match_info))
2618 g_string_append_len (result,
2620 match_info->offsets[0] - str_pos);
2621 done = (*eval) (match_info, result, user_data);
2622 str_pos = match_info->offsets[1];
2623 g_match_info_next (match_info, &tmp_error);
2625 g_match_info_free (match_info);
2626 if (tmp_error != NULL)
2628 g_propagate_error (error, tmp_error);
2629 g_string_free (result, TRUE);
2633 g_string_append_len (result, string + str_pos, string_len - str_pos);
2634 return g_string_free (result, FALSE);
2638 * g_regex_check_replacement:
2639 * @replacement: the replacement string
2640 * @has_references: location to store information about
2641 * references in @replacement or %NULL
2642 * @error: location to store error
2644 * Checks whether @replacement is a valid replacement string
2645 * (see g_regex_replace()), i.e. that all escape sequences in
2648 * If @has_references is not %NULL then @replacement is checked
2649 * for pattern references. For instance, replacement text 'foo\n'
2650 * does not contain references and may be evaluated without information
2651 * about actual match, but '\0\1' (whole match followed by first
2652 * subpattern) requires valid #GMatchInfo object.
2654 * Returns: whether @replacement is a valid replacement string
2659 g_regex_check_replacement (const gchar *replacement,
2660 gboolean *has_references,
2666 list = split_replacement (replacement, &tmp);
2670 g_propagate_error (error, tmp);
2675 *has_references = interpolation_list_needs_match (list);
2677 g_list_foreach (list, (GFunc) free_interpolation_data, NULL);
2684 * g_regex_escape_string:
2685 * @string: the string to escape
2686 * @length: the length of @string, or -1 if @string is nul-terminated
2688 * Escapes the special characters used for regular expressions
2689 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2690 * function is useful to dynamically generate regular expressions.
2692 * @string can contain nul characters that are replaced with "\0",
2693 * in this case remember to specify the correct length of @string
2696 * Returns: a newly-allocated escaped string
2701 g_regex_escape_string (const gchar *string,
2705 const char *p, *piece_start, *end;
2707 g_return_val_if_fail (string != NULL, NULL);
2710 length = strlen (string);
2712 end = string + length;
2713 p = piece_start = string;
2714 escaped = g_string_sized_new (length + 1);
2735 if (p != piece_start)
2736 /* copy the previous piece. */
2737 g_string_append_len (escaped, piece_start, p - piece_start);
2738 g_string_append_c (escaped, '\\');
2740 g_string_append_c (escaped, '0');
2742 g_string_append_c (escaped, *p);
2746 p = g_utf8_next_char (p);
2751 if (piece_start < end)
2752 g_string_append_len (escaped, piece_start, end - piece_start);
2754 return g_string_free (escaped, FALSE);
2757 #define __G_REGEX_C__
2758 #include "galiasdef.c"