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 /* Mask of all the possible values for GRegexCompileFlags. */
37 #define G_REGEX_COMPILE_MASK (G_REGEX_CASELESS | \
42 G_REGEX_DOLLAR_ENDONLY | \
45 G_REGEX_NO_AUTO_CAPTURE | \
48 G_REGEX_NEWLINE_CR | \
49 G_REGEX_NEWLINE_LF | \
52 /* Mask of all the possible values for GRegexMatchFlags. */
53 #define G_REGEX_MATCH_MASK (G_REGEX_MATCH_ANCHORED | \
54 G_REGEX_MATCH_NOTBOL | \
55 G_REGEX_MATCH_NOTEOL | \
56 G_REGEX_MATCH_NOTEMPTY | \
57 G_REGEX_MATCH_PARTIAL | \
58 G_REGEX_MATCH_NEWLINE_CR | \
59 G_REGEX_MATCH_NEWLINE_LF | \
60 G_REGEX_MATCH_NEWLINE_CRLF | \
61 G_REGEX_MATCH_NEWLINE_ANY)
63 /* if the string is in UTF-8 use g_utf8_ functions, else use
65 #define NEXT_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
66 g_utf8_next_char (s) : \
68 #define PREV_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
69 g_utf8_prev_char (s) : \
74 GRegex *regex; /* the regex */
75 GRegexMatchFlags match_opts; /* options used at match time on the regex */
76 gint matches; /* number of matching sub patterns */
77 gint pos; /* position in the string where last match left off */
78 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
79 gint n_offsets; /* number of offsets */
80 gint *workspace; /* workspace for pcre_dfa_exec() */
81 gint n_workspace; /* number of workspace elements */
82 const gchar *string; /* string passed to the match function */
83 gssize string_len; /* length of string */
88 volatile gint ref_count; /* the ref count for the immutable part */
89 gchar *pattern; /* the pattern */
90 pcre *pcre_re; /* compiled form of the pattern */
91 GRegexCompileFlags compile_opts; /* options used at compile time on the pattern */
92 GRegexMatchFlags match_opts; /* options used at match time on the regex */
93 pcre_extra *extra; /* data stored when G_REGEX_OPTIMIZE is used */
96 /* TRUE if ret is an error code, FALSE otherwise. */
97 #define IS_PCRE_ERROR(ret) ((ret) < PCRE_ERROR_NOMATCH && (ret) != PCRE_ERROR_PARTIAL)
99 typedef struct _InterpolationData InterpolationData;
100 static gboolean interpolation_list_needs_match (GList *list);
101 static gboolean interpolate_replacement (const GMatchInfo *match_info,
104 static GList *split_replacement (const gchar *replacement,
106 static void free_interpolation_data (InterpolationData *data);
110 match_error (gint errcode)
114 case PCRE_ERROR_NOMATCH:
117 case PCRE_ERROR_NULL:
118 /* NULL argument, this should not happen in GRegex */
119 g_warning ("A NULL argument was passed to PCRE");
121 case PCRE_ERROR_BADOPTION:
122 return "bad options";
123 case PCRE_ERROR_BADMAGIC:
124 return _("corrupted object");
125 case PCRE_ERROR_UNKNOWN_OPCODE:
126 return N_("internal error or corrupted object");
127 case PCRE_ERROR_NOMEMORY:
128 return _("out of memory");
129 case PCRE_ERROR_NOSUBSTRING:
130 /* not used by pcre_exec() */
132 case PCRE_ERROR_MATCHLIMIT:
133 return _("backtracking limit reached");
134 case PCRE_ERROR_CALLOUT:
135 /* callouts are not implemented */
137 case PCRE_ERROR_BADUTF8:
138 case PCRE_ERROR_BADUTF8_OFFSET:
139 /* we do not check if strings are valid */
141 case PCRE_ERROR_PARTIAL:
144 case PCRE_ERROR_BADPARTIAL:
145 return _("the pattern contains items not supported for partial matching");
146 case PCRE_ERROR_INTERNAL:
147 return _("internal error");
148 case PCRE_ERROR_BADCOUNT:
149 /* negative ovecsize, this should not happen in GRegex */
150 g_warning ("A negative ovecsize was passed to PCRE");
152 case PCRE_ERROR_DFA_UITEM:
153 return _("the pattern contains items not supported for partial matching");
154 case PCRE_ERROR_DFA_UCOND:
155 return _("back references as conditions are not supported for partial matching");
156 case PCRE_ERROR_DFA_UMLIMIT:
157 /* the match_field field is not used in GRegex */
159 case PCRE_ERROR_DFA_WSSIZE:
160 /* handled expanding the workspace */
162 case PCRE_ERROR_DFA_RECURSE:
163 case PCRE_ERROR_RECURSIONLIMIT:
164 return _("recursion limit reached");
165 case PCRE_ERROR_NULLWSLIMIT:
166 return _("workspace limit for empty substrings reached");
167 case PCRE_ERROR_BADNEWLINE:
168 return _("invalid combination of newline flags");
172 return _("unknown error");
176 translate_compile_error (gint *errcode, const gchar **errmsg)
178 /* Compile errors are created adding 100 to the error code returned
180 * If errcode is known we put the translatable error message in
181 * erromsg. If errcode is unknown we put the generic
182 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
183 * untranslated error message returned by PCRE.
184 * Note that there can be more PCRE errors with the same GRegexError
185 * and that some PCRE errors are useless for us.
191 case G_REGEX_ERROR_STRAY_BACKSLASH:
192 *errmsg = _("\\ at end of pattern");
194 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
195 *errmsg = _("\\c at end of pattern");
197 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
198 *errmsg = _("unrecognized character follows \\");
201 /* A number of Perl escapes are not handled by PCRE.
202 * Therefore it explicitly raises ERR37.
204 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
205 *errmsg = _("case-changing escapes (\\l, \\L, \\u, \\U) are not allowed here");
207 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
208 *errmsg = _("numbers out of order in {} quantifier");
210 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
211 *errmsg = _("number too big in {} quantifier");
213 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
214 *errmsg = _("missing terminating ] for character class");
216 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
217 *errmsg = _("invalid escape sequence in character class");
219 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
220 *errmsg = _("range out of order in character class");
222 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
223 *errmsg = _("nothing to repeat");
225 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
226 *errmsg = _("unrecognized character after (?");
229 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
230 *errmsg = _("unrecognized character after (?<");
233 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
234 *errmsg = _("unrecognized character after (?P");
236 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
237 *errmsg = _("POSIX named classes are supported only within a class");
239 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
240 *errmsg = _("missing terminating )");
243 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
244 *errmsg = _(") without opening (");
247 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
248 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
249 * sequences here, '(?-54' would be an example for the second group.
251 *errmsg = _("(?R or (?[+-]digits must be followed by )");
253 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
254 *errmsg = _("reference to non-existent subpattern");
256 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
257 *errmsg = _("missing ) after comment");
259 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
260 *errmsg = _("regular expression too large");
262 case G_REGEX_ERROR_MEMORY_ERROR:
263 *errmsg = _("failed to get memory");
265 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
266 *errmsg = _("lookbehind assertion is not fixed length");
268 case G_REGEX_ERROR_MALFORMED_CONDITION:
269 *errmsg = _("malformed number or name after (?(");
271 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
272 *errmsg = _("conditional group contains more than two branches");
274 case G_REGEX_ERROR_ASSERTION_EXPECTED:
275 *errmsg = _("assertion expected after (?(");
277 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
278 *errmsg = _("unknown POSIX class name");
280 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
281 *errmsg = _("POSIX collating elements are not supported");
283 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
284 *errmsg = _("character value in \\x{...} sequence is too large");
286 case G_REGEX_ERROR_INVALID_CONDITION:
287 *errmsg = _("invalid condition (?(0)");
289 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
290 *errmsg = _("\\C not allowed in lookbehind assertion");
292 case G_REGEX_ERROR_INFINITE_LOOP:
293 *errmsg = _("recursive call could loop indefinitely");
295 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
296 *errmsg = _("missing terminator in subpattern name");
298 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
299 *errmsg = _("two named subpatterns have the same name");
301 case G_REGEX_ERROR_MALFORMED_PROPERTY:
302 *errmsg = _("malformed \\P or \\p sequence");
304 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
305 *errmsg = _("unknown property name after \\P or \\p");
307 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
308 *errmsg = _("subpattern name is too long (maximum 32 characters)");
310 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
311 *errmsg = _("too many named subpatterns (maximum 10,000)");
313 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
314 *errmsg = _("octal value is greater than \\377");
316 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
317 *errmsg = _("DEFINE group contains more than one branch");
319 case G_REGEX_ERROR_DEFINE_REPETION:
320 *errmsg = _("repeating a DEFINE group is not allowed");
322 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
323 *errmsg = _("inconsistent NEWLINE options");
325 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
326 *errmsg = _("\\g is not followed by a braced name or an optionally "
327 "braced non-zero number");
330 *errcode = G_REGEX_ERROR_INTERNAL;
331 *errmsg = _("unexpected repeat");
334 *errcode = G_REGEX_ERROR_INTERNAL;
335 *errmsg = _("code overflow");
338 *errcode = G_REGEX_ERROR_INTERNAL;
339 *errmsg = _("overran compiling workspace");
342 *errcode = G_REGEX_ERROR_INTERNAL;
343 *errmsg = _("previously-checked referenced subpattern not found");
346 /* This should not happen as we never pass a NULL erroffset */
347 g_warning ("erroffset passed as NULL");
348 *errcode = G_REGEX_ERROR_COMPILE;
351 /* This should not happen as we check options before passing them
352 * to pcre_compile2() */
353 g_warning ("unknown option bit(s) set");
354 *errcode = G_REGEX_ERROR_COMPILE;
359 /* These errors should not happen as we are using an UTF8-enabled PCRE
360 * and we do not check if strings are valid */
361 g_warning ("%s", *errmsg);
362 *errcode = G_REGEX_ERROR_COMPILE;
365 *errcode = G_REGEX_ERROR_COMPILE;
372 match_info_new (const GRegex *regex,
379 GMatchInfo *match_info;
382 string_len = strlen (string);
384 match_info = g_new0 (GMatchInfo, 1);
385 match_info->regex = g_regex_ref ((GRegex *)regex);
386 match_info->string = string;
387 match_info->string_len = string_len;
388 match_info->matches = PCRE_ERROR_NOMATCH;
389 match_info->pos = start_position;
390 match_info->match_opts = match_options;
394 /* These values should be enough for most cases, if they are not
395 * enough g_regex_match_all_full() will expand them. */
396 match_info->n_offsets = 24;
397 match_info->n_workspace = 100;
398 match_info->workspace = g_new (gint, match_info->n_workspace);
403 pcre_fullinfo (regex->pcre_re, regex->extra,
404 PCRE_INFO_CAPTURECOUNT, &capture_count);
405 match_info->n_offsets = (capture_count + 1) * 3;
408 match_info->offsets = g_new0 (gint, match_info->n_offsets);
409 /* Set an invalid position for the previous match. */
410 match_info->offsets[0] = -1;
411 match_info->offsets[1] = -1;
417 * g_match_info_get_regex:
418 * @match_info: a #GMatchInfo
420 * Returns #GRegex object used in @match_info. It belongs to Glib
421 * and must not be freed. Use g_regex_ref() if you need to keep it
422 * after you free @match_info object.
424 * Returns: #GRegex object used in @match_info
429 g_match_info_get_regex (const GMatchInfo *match_info)
431 g_return_val_if_fail (match_info != NULL, NULL);
432 return match_info->regex;
436 * g_match_info_get_string:
437 * @match_info: a #GMatchInfo
439 * Returns the string searched with @match_info. This is the
440 * string passed to g_regex_match() or g_regex_replace() so
441 * you may not free it before calling this function.
443 * Returns: the string searched with @match_info
448 g_match_info_get_string (const GMatchInfo *match_info)
450 g_return_val_if_fail (match_info != NULL, NULL);
451 return match_info->string;
456 * @match_info: a #GMatchInfo
458 * Frees all the memory associated with the #GMatchInfo structure.
463 g_match_info_free (GMatchInfo *match_info)
467 g_regex_unref (match_info->regex);
468 g_free (match_info->offsets);
469 g_free (match_info->workspace);
476 * @match_info: a #GMatchInfo structure
477 * @error: location to store the error occuring, or %NULL to ignore errors
479 * Scans for the next match using the same parameters of the previous
480 * call to g_regex_match_full() or g_regex_match() that returned
483 * The match is done on the string passed to the match function, so you
484 * cannot free it before calling this function.
486 * Returns: %TRUE is the string matched, %FALSE otherwise
491 g_match_info_next (GMatchInfo *match_info,
494 gint prev_match_start;
497 g_return_val_if_fail (match_info != NULL, FALSE);
498 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
499 g_return_val_if_fail (match_info->pos >= 0, FALSE);
501 prev_match_start = match_info->offsets[0];
502 prev_match_end = match_info->offsets[1];
504 match_info->matches = pcre_exec (match_info->regex->pcre_re,
505 match_info->regex->extra,
507 match_info->string_len,
509 match_info->regex->match_opts | match_info->match_opts,
511 match_info->n_offsets);
512 if (IS_PCRE_ERROR (match_info->matches))
514 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
515 _("Error while matching regular expression %s: %s"),
516 match_info->regex->pattern, match_error (match_info->matches));
520 /* avoid infinite loops if the pattern is an empty string or something
522 if (match_info->pos == match_info->offsets[1])
524 if (match_info->pos > match_info->string_len)
526 /* we have reached the end of the string */
527 match_info->pos = -1;
528 match_info->matches = PCRE_ERROR_NOMATCH;
532 match_info->pos = NEXT_CHAR (match_info->regex,
533 &match_info->string[match_info->pos]) -
538 match_info->pos = match_info->offsets[1];
541 /* it's possible to get two identical matches when we are matching
542 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
543 * the string is "RegExTest" we have:
544 * - search at position 0: match from 0 to 0
545 * - search at position 1: match from 3 to 3
546 * - search at position 3: match from 3 to 3 (duplicate)
547 * - search at position 4: match from 5 to 5
548 * - search at position 5: match from 5 to 5 (duplicate)
549 * - search at position 6: no match -> stop
550 * so we have to ignore the duplicates.
551 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
552 if (match_info->matches >= 0 &&
553 prev_match_start == match_info->offsets[0] &&
554 prev_match_end == match_info->offsets[1])
556 /* ignore this match and search the next one */
557 return g_match_info_next (match_info, error);
560 return match_info->matches >= 0;
564 * g_match_info_matches:
565 * @match_info: a #GMatchInfo structure
567 * Returns whether the previous match operation succeeded.
569 * Returns: %TRUE if the previous match operation succeeded,
575 g_match_info_matches (const GMatchInfo *match_info)
577 g_return_val_if_fail (match_info != NULL, FALSE);
579 return match_info->matches >= 0;
583 * g_match_info_get_match_count:
584 * @match_info: a #GMatchInfo structure
586 * Retrieves the number of matched substrings (including substring 0,
587 * that is the whole matched text), so 1 is returned if the pattern
588 * has no substrings in it and 0 is returned if the match failed.
590 * If the last match was obtained using the DFA algorithm, that is
591 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
592 * count is not that of the number of capturing parentheses but that of
593 * the number of matched substrings.
595 * Returns: Number of matched substrings, or -1 if an error occurred
600 g_match_info_get_match_count (const GMatchInfo *match_info)
602 g_return_val_if_fail (match_info, -1);
604 if (match_info->matches == PCRE_ERROR_NOMATCH)
607 else if (match_info->matches < PCRE_ERROR_NOMATCH)
612 return match_info->matches;
616 * g_match_info_is_partial_match:
617 * @match_info: a #GMatchInfo structure
619 * Usually if the string passed to g_regex_match*() matches as far as
620 * it goes, but is too short to match the entire pattern, %FALSE is
621 * returned. There are circumstances where it might be helpful to
622 * distinguish this case from other cases in which there is no match.
624 * Consider, for example, an application where a human is required to
625 * type in data for a field with specific formatting requirements. An
626 * example might be a date in the form ddmmmyy, defined by the pattern
627 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
628 * If the application sees the user’s keystrokes one by one, and can
629 * check that what has been typed so far is potentially valid, it is
630 * able to raise an error as soon as a mistake is made.
632 * GRegex supports the concept of partial matching by means of the
633 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
634 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
635 * for a complete match, %FALSE otherwise. But, when these functions
636 * return %FALSE, you can check if the match was partial calling
637 * g_match_info_is_partial_match().
639 * When using partial matching you cannot use g_match_info_fetch*().
641 * Because of the way certain internal optimizations are implemented
642 * the partial matching algorithm cannot be used with all patterns.
643 * So repeated single characters such as "a{2,4}" and repeated single
644 * meta-sequences such as "\d+" are not permitted if the maximum number
645 * of occurrences is greater than one. Optional items such as "\d?"
646 * (where the maximum is one) are permitted. Quantifiers with any values
647 * are permitted after parentheses, so the invalid examples above can be
648 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
649 * for a pattern that does not conform to the restrictions, matching
650 * functions return an error.
652 * Returns: %TRUE if the match was partial, %FALSE otherwise
657 g_match_info_is_partial_match (const GMatchInfo *match_info)
659 g_return_val_if_fail (match_info != NULL, FALSE);
661 return match_info->matches == PCRE_ERROR_PARTIAL;
665 * g_match_info_expand_references:
666 * @match_info: a #GMatchInfo or %NULL
667 * @string_to_expand: the string to expand
668 * @error: location to store the error occuring, or %NULL to ignore errors
670 * Returns a new string containing the text in @string_to_expand with
671 * references and escape sequences expanded. References refer to the last
672 * match done with @string against @regex and have the same syntax used by
675 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
676 * passed to g_regex_new().
678 * The backreferences are extracted from the string passed to the match
679 * function, so you cannot call this function after freeing the string.
681 * @match_info may be %NULL in which case @string_to_expand must not
682 * contain references. For instance "foo\n" does not refer to an actual
683 * pattern and '\n' merely will be replaced with \n character,
684 * while to expand "\0" (whole match) one needs the result of a match.
685 * Use g_regex_check_replacement() to find out whether @string_to_expand
686 * contains references.
688 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
693 g_match_info_expand_references (const GMatchInfo *match_info,
694 const gchar *string_to_expand,
699 GError *tmp_error = NULL;
701 g_return_val_if_fail (string_to_expand != NULL, NULL);
702 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
704 list = split_replacement (string_to_expand, &tmp_error);
705 if (tmp_error != NULL)
707 g_propagate_error (error, tmp_error);
711 if (!match_info && interpolation_list_needs_match (list))
713 g_critical ("String '%s' contains references to the match, can't "
714 "expand references without GMatchInfo object",
719 result = g_string_sized_new (strlen (string_to_expand));
720 interpolate_replacement (match_info, result, list);
722 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
725 return g_string_free (result, FALSE);
729 * g_match_info_fetch:
730 * @match_info: #GMatchInfo structure
731 * @match_num: number of the sub expression
733 * Retrieves the text matching the @match_num<!-- -->'th capturing
734 * parentheses. 0 is the full text of the match, 1 is the first paren
735 * set, 2 the second, and so on.
737 * If @match_num is a valid sub pattern but it didn't match anything
738 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
739 * string is returned.
741 * If the match was obtained using the DFA algorithm, that is using
742 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
743 * string is not that of a set of parentheses but that of a matched
744 * substring. Substrings are matched in reverse order of length, so
745 * 0 is the longest match.
747 * The string is fetched from the string passed to the match function,
748 * so you cannot call this function after freeing the string.
750 * Returns: (allow-none): The matched substring, or %NULL if an error
751 * occurred. You have to free the string yourself
756 g_match_info_fetch (const GMatchInfo *match_info,
759 /* we cannot use pcre_get_substring() because it allocates the
760 * string using pcre_malloc(). */
764 g_return_val_if_fail (match_info != NULL, NULL);
765 g_return_val_if_fail (match_num >= 0, NULL);
767 /* match_num does not exist or it didn't matched, i.e. matching "b"
768 * against "(a)?b" then group 0 is empty. */
769 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
771 else if (start == -1)
772 match = g_strdup ("");
774 match = g_strndup (&match_info->string[start], end - start);
780 * g_match_info_fetch_pos:
781 * @match_info: #GMatchInfo structure
782 * @match_num: number of the sub expression
783 * @start_pos: (out) (allow-none): pointer to location where to store
784 * the start position, or %NULL
785 * @end_pos: (out) (allow-none): pointer to location where to store
786 * the end position, or %NULL
788 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
789 * parentheses. 0 is the full text of the match, 1 is the first
790 * paren set, 2 the second, and so on.
792 * If @match_num is a valid sub pattern but it didn't match anything
793 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
794 * and @end_pos are set to -1 and %TRUE is returned.
796 * If the match was obtained using the DFA algorithm, that is using
797 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
798 * position is not that of a set of parentheses but that of a matched
799 * substring. Substrings are matched in reverse order of length, so
800 * 0 is the longest match.
802 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
803 * the position cannot be fetched, @start_pos and @end_pos are left
809 g_match_info_fetch_pos (const GMatchInfo *match_info,
814 g_return_val_if_fail (match_info != NULL, FALSE);
815 g_return_val_if_fail (match_num >= 0, FALSE);
817 /* make sure the sub expression number they're requesting is less than
818 * the total number of sub expressions that were matched. */
819 if (match_num >= match_info->matches)
822 if (start_pos != NULL)
823 *start_pos = match_info->offsets[2 * match_num];
826 *end_pos = match_info->offsets[2 * match_num + 1];
832 * Returns number of first matched subpattern with name @name.
833 * There may be more than one in case when DUPNAMES is used,
834 * and not all subpatterns with that name match;
835 * pcre_get_stringnumber() does not work in that case.
838 get_matched_substring_number (const GMatchInfo *match_info,
845 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
846 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
848 /* This code is copied from pcre_get.c: get_first_set() */
849 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
857 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
859 gint n = (entry[0] << 8) + entry[1];
860 if (match_info->offsets[n*2] >= 0)
864 return (first[0] << 8) + first[1];
868 * g_match_info_fetch_named:
869 * @match_info: #GMatchInfo structure
870 * @name: name of the subexpression
872 * Retrieves the text matching the capturing parentheses named @name.
874 * If @name is a valid sub pattern name but it didn't match anything
875 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
876 * then an empty string is returned.
878 * The string is fetched from the string passed to the match function,
879 * so you cannot call this function after freeing the string.
881 * Returns: (allow-none): The matched substring, or %NULL if an error
882 * occurred. You have to free the string yourself
887 g_match_info_fetch_named (const GMatchInfo *match_info,
890 /* we cannot use pcre_get_named_substring() because it allocates the
891 * string using pcre_malloc(). */
894 g_return_val_if_fail (match_info != NULL, NULL);
895 g_return_val_if_fail (name != NULL, NULL);
897 num = get_matched_substring_number (match_info, name);
901 return g_match_info_fetch (match_info, num);
905 * g_match_info_fetch_named_pos:
906 * @match_info: #GMatchInfo structure
907 * @name: name of the subexpression
908 * @start_pos: (out) (allow-none): pointer to location where to store
909 * the start position, or %NULL
910 * @end_pos: (out) (allow-none): pointer to location where to store
911 * the end position, or %NULL
913 * Retrieves the position in bytes of the capturing parentheses named @name.
915 * If @name is a valid sub pattern name but it didn't match anything
916 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
917 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
919 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
920 * If the position cannot be fetched, @start_pos and @end_pos
921 * are left unchanged.
926 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
933 g_return_val_if_fail (match_info != NULL, FALSE);
934 g_return_val_if_fail (name != NULL, FALSE);
936 num = get_matched_substring_number (match_info, name);
940 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
944 * g_match_info_fetch_all:
945 * @match_info: a #GMatchInfo structure
947 * Bundles up pointers to each of the matching substrings from a match
948 * and stores them in an array of gchar pointers. The first element in
949 * the returned array is the match number 0, i.e. the entire matched
952 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
953 * "b" against "(a)?b") then an empty string is inserted.
955 * If the last match was obtained using the DFA algorithm, that is using
956 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
957 * strings are not that matched by sets of parentheses but that of the
958 * matched substring. Substrings are matched in reverse order of length,
959 * so the first one is the longest match.
961 * The strings are fetched from the string passed to the match function,
962 * so you cannot call this function after freeing the string.
964 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
965 * It must be freed using g_strfreev(). If the previous match failed
971 g_match_info_fetch_all (const GMatchInfo *match_info)
973 /* we cannot use pcre_get_substring_list() because the returned value
974 * isn't suitable for g_strfreev(). */
978 g_return_val_if_fail (match_info != NULL, NULL);
980 if (match_info->matches < 0)
983 result = g_new (gchar *, match_info->matches + 1);
984 for (i = 0; i < match_info->matches; i++)
985 result[i] = g_match_info_fetch (match_info, i);
995 g_regex_error_quark (void)
997 static GQuark error_quark = 0;
999 if (error_quark == 0)
1000 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1009 * Increases reference count of @regex by 1.
1016 g_regex_ref (GRegex *regex)
1018 g_return_val_if_fail (regex != NULL, NULL);
1019 g_atomic_int_inc (®ex->ref_count);
1027 * Decreases reference count of @regex by 1. When reference count drops
1028 * to zero, it frees all the memory associated with the regex structure.
1033 g_regex_unref (GRegex *regex)
1035 g_return_if_fail (regex != NULL);
1037 if (g_atomic_int_exchange_and_add (®ex->ref_count, -1) - 1 == 0)
1039 g_free (regex->pattern);
1040 if (regex->pcre_re != NULL)
1041 pcre_free (regex->pcre_re);
1042 if (regex->extra != NULL)
1043 pcre_free (regex->extra);
1050 * @pattern: the regular expression
1051 * @compile_options: compile options for the regular expression, or 0
1052 * @match_options: match options for the regular expression, or 0
1053 * @error: return location for a #GError
1055 * Compiles the regular expression to an internal form, and does
1056 * the initial setup of the #GRegex structure.
1058 * Returns: a #GRegex structure. Call g_regex_unref() when you
1064 g_regex_new (const gchar *pattern,
1065 GRegexCompileFlags compile_options,
1066 GRegexMatchFlags match_options,
1071 const gchar *errmsg;
1074 gboolean optimize = FALSE;
1075 static gboolean initialized = FALSE;
1076 unsigned long int pcre_compile_options;
1078 g_return_val_if_fail (pattern != NULL, NULL);
1079 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1080 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1081 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1088 pcre_config (PCRE_CONFIG_UTF8, &support);
1091 msg = N_("PCRE library is compiled without UTF8 support");
1092 g_critical ("%s", msg);
1093 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1097 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1100 msg = N_("PCRE library is compiled without UTF8 properties support");
1101 g_critical ("%s", msg);
1102 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1109 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1110 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1111 if (compile_options & G_REGEX_OPTIMIZE)
1114 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1115 * instead uses UTF-8 only if required with PCRE_UTF8. */
1116 if (compile_options & G_REGEX_RAW)
1119 compile_options &= ~G_REGEX_RAW;
1124 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1125 match_options |= PCRE_NO_UTF8_CHECK;
1128 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1129 * not for the system one. */
1130 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1131 !(compile_options & G_REGEX_NEWLINE_LF))
1133 compile_options |= PCRE_NEWLINE_ANY;
1136 /* compile the pattern */
1137 re = pcre_compile2 (pattern, compile_options, &errcode,
1138 &errmsg, &erroffset, NULL);
1140 /* if the compilation failed, set the error member and return
1146 /* Translate the PCRE error code to GRegexError and use a translated
1147 * error message if possible */
1148 translate_compile_error (&errcode, &errmsg);
1150 /* PCRE uses byte offsets but we want to show character offsets */
1151 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1153 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1154 _("Error while compiling regular "
1155 "expression %s at char %d: %s"),
1156 pattern, erroffset, errmsg);
1157 g_propagate_error (error, tmp_error);
1162 /* For options set at the beginning of the pattern, pcre puts them into
1163 * compile options, e.g. "(?i)foo" will make the pcre structure store
1164 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1165 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1166 compile_options = pcre_compile_options;
1168 if (!(compile_options & G_REGEX_DUPNAMES))
1170 gboolean jchanged = FALSE;
1171 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1173 compile_options |= G_REGEX_DUPNAMES;
1176 regex = g_new0 (GRegex, 1);
1177 regex->ref_count = 1;
1178 regex->pattern = g_strdup (pattern);
1179 regex->pcre_re = re;
1180 regex->compile_opts = compile_options;
1181 regex->match_opts = match_options;
1185 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1188 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1189 G_REGEX_ERROR_OPTIMIZE,
1190 _("Error while optimizing "
1191 "regular expression %s: %s"),
1194 g_propagate_error (error, tmp_error);
1196 g_regex_unref (regex);
1205 * g_regex_get_pattern:
1206 * @regex: a #GRegex structure
1208 * Gets the pattern string associated with @regex, i.e. a copy of
1209 * the string passed to g_regex_new().
1211 * Returns: the pattern of @regex
1216 g_regex_get_pattern (const GRegex *regex)
1218 g_return_val_if_fail (regex != NULL, NULL);
1220 return regex->pattern;
1224 * g_regex_get_max_backref:
1227 * Returns the number of the highest back reference
1228 * in the pattern, or 0 if the pattern does not contain
1231 * Returns: the number of the highest back reference
1236 g_regex_get_max_backref (const GRegex *regex)
1240 pcre_fullinfo (regex->pcre_re, regex->extra,
1241 PCRE_INFO_BACKREFMAX, &value);
1247 * g_regex_get_capture_count:
1250 * Returns the number of capturing subpatterns in the pattern.
1252 * Returns: the number of capturing subpatterns
1257 g_regex_get_capture_count (const GRegex *regex)
1261 pcre_fullinfo (regex->pcre_re, regex->extra,
1262 PCRE_INFO_CAPTURECOUNT, &value);
1268 * g_regex_get_compile_flags:
1271 * Returns the compile options that @regex was created with.
1273 * Returns: flags from #GRegexCompileFlags
1278 g_regex_get_compile_flags (const GRegex *regex)
1280 g_return_val_if_fail (regex != NULL, 0);
1282 return regex->compile_opts;
1286 * g_regex_get_match_flags:
1289 * Returns the match options that @regex was created with.
1291 * Returns: flags from #GRegexMatchFlags
1296 g_regex_get_match_flags (const GRegex *regex)
1298 g_return_val_if_fail (regex != NULL, 0);
1300 return regex->match_opts;
1304 * g_regex_match_simple:
1305 * @pattern: the regular expression
1306 * @string: the string to scan for matches
1307 * @compile_options: compile options for the regular expression, or 0
1308 * @match_options: match options, or 0
1310 * Scans for a match in @string for @pattern.
1312 * This function is equivalent to g_regex_match() but it does not
1313 * require to compile the pattern with g_regex_new(), avoiding some
1314 * lines of code when you need just to do a match without extracting
1315 * substrings, capture counts, and so on.
1317 * If this function is to be called on the same @pattern more than
1318 * once, it's more efficient to compile the pattern once with
1319 * g_regex_new() and then use g_regex_match().
1321 * Returns: %TRUE if the string matched, %FALSE otherwise
1326 g_regex_match_simple (const gchar *pattern,
1327 const gchar *string,
1328 GRegexCompileFlags compile_options,
1329 GRegexMatchFlags match_options)
1334 regex = g_regex_new (pattern, compile_options, 0, NULL);
1337 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1338 g_regex_unref (regex);
1344 * @regex: a #GRegex structure from g_regex_new()
1345 * @string: the string to scan for matches
1346 * @match_options: match options
1347 * @match_info: (out) (allow-none): pointer to location where to store
1348 * the #GMatchInfo, or %NULL if you do not need it
1350 * Scans for a match in string for the pattern in @regex.
1351 * The @match_options are combined with the match options specified
1352 * when the @regex structure was created, letting you have more
1353 * flexibility in reusing #GRegex structures.
1355 * A #GMatchInfo structure, used to get information on the match,
1356 * is stored in @match_info if not %NULL. Note that if @match_info
1357 * is not %NULL then it is created even if the function returns %FALSE,
1358 * i.e. you must free it regardless if regular expression actually matched.
1360 * To retrieve all the non-overlapping matches of the pattern in
1361 * string you can use g_match_info_next().
1365 * print_uppercase_words (const gchar *string)
1367 * /* Print all uppercase-only words. */
1369 * GMatchInfo *match_info;
1371 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1372 * g_regex_match (regex, string, 0, &match_info);
1373 * while (g_match_info_matches (match_info))
1375 * gchar *word = g_match_info_fetch (match_info, 0);
1376 * g_print ("Found: %s\n", word);
1378 * g_match_info_next (match_info, NULL);
1380 * g_match_info_free (match_info);
1381 * g_regex_unref (regex);
1385 * @string is not copied and is used in #GMatchInfo internally. If
1386 * you use any #GMatchInfo method (except g_match_info_free()) after
1387 * freeing or modifying @string then the behaviour is undefined.
1389 * Returns: %TRUE is the string matched, %FALSE otherwise
1394 g_regex_match (const GRegex *regex,
1395 const gchar *string,
1396 GRegexMatchFlags match_options,
1397 GMatchInfo **match_info)
1399 return g_regex_match_full (regex, string, -1, 0, match_options,
1404 * g_regex_match_full:
1405 * @regex: a #GRegex structure from g_regex_new()
1406 * @string: (array length=string_len): the string to scan for matches
1407 * @string_len: the length of @string, or -1 if @string is nul-terminated
1408 * @start_position: starting index of the string to match
1409 * @match_options: match options
1410 * @match_info: (out) (allow-none): pointer to location where to store
1411 * the #GMatchInfo, or %NULL if you do not need it
1412 * @error: location to store the error occuring, or %NULL to ignore errors
1414 * Scans for a match in string for the pattern in @regex.
1415 * The @match_options are combined with the match options specified
1416 * when the @regex structure was created, letting you have more
1417 * flexibility in reusing #GRegex structures.
1419 * Setting @start_position differs from just passing over a shortened
1420 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1421 * that begins with any kind of lookbehind assertion, such as "\b".
1423 * A #GMatchInfo structure, used to get information on the match, is
1424 * stored in @match_info if not %NULL. Note that if @match_info is
1425 * not %NULL then it is created even if the function returns %FALSE,
1426 * i.e. you must free it regardless if regular expression actually
1429 * @string is not copied and is used in #GMatchInfo internally. If
1430 * you use any #GMatchInfo method (except g_match_info_free()) after
1431 * freeing or modifying @string then the behaviour is undefined.
1433 * To retrieve all the non-overlapping matches of the pattern in
1434 * string you can use g_match_info_next().
1438 * print_uppercase_words (const gchar *string)
1440 * /* Print all uppercase-only words. */
1442 * GMatchInfo *match_info;
1443 * GError *error = NULL;
1445 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1446 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1447 * while (g_match_info_matches (match_info))
1449 * gchar *word = g_match_info_fetch (match_info, 0);
1450 * g_print ("Found: %s\n", word);
1452 * g_match_info_next (match_info, &error);
1454 * g_match_info_free (match_info);
1455 * g_regex_unref (regex);
1456 * if (error != NULL)
1458 * g_printerr ("Error while matching: %s\n", error->message);
1459 * g_error_free (error);
1464 * Returns: %TRUE is the string matched, %FALSE otherwise
1469 g_regex_match_full (const GRegex *regex,
1470 const gchar *string,
1472 gint start_position,
1473 GRegexMatchFlags match_options,
1474 GMatchInfo **match_info,
1480 g_return_val_if_fail (regex != NULL, FALSE);
1481 g_return_val_if_fail (string != NULL, FALSE);
1482 g_return_val_if_fail (start_position >= 0, FALSE);
1483 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1484 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1486 info = match_info_new (regex, string, string_len, start_position,
1487 match_options, FALSE);
1488 match_ok = g_match_info_next (info, error);
1489 if (match_info != NULL)
1492 g_match_info_free (info);
1498 * g_regex_match_all:
1499 * @regex: a #GRegex structure from g_regex_new()
1500 * @string: the string to scan for matches
1501 * @match_options: match options
1502 * @match_info: (out) (allow-none): pointer to location where to store
1503 * the #GMatchInfo, or %NULL if you do not need it
1505 * Using the standard algorithm for regular expression matching only
1506 * the longest match in the string is retrieved. This function uses
1507 * a different algorithm so it can retrieve all the possible matches.
1508 * For more documentation see g_regex_match_all_full().
1510 * A #GMatchInfo structure, used to get information on the match, is
1511 * stored in @match_info if not %NULL. Note that if @match_info is
1512 * not %NULL then it is created even if the function returns %FALSE,
1513 * i.e. you must free it regardless if regular expression actually
1516 * @string is not copied and is used in #GMatchInfo internally. If
1517 * you use any #GMatchInfo method (except g_match_info_free()) after
1518 * freeing or modifying @string then the behaviour is undefined.
1520 * Returns: %TRUE is the string matched, %FALSE otherwise
1525 g_regex_match_all (const GRegex *regex,
1526 const gchar *string,
1527 GRegexMatchFlags match_options,
1528 GMatchInfo **match_info)
1530 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1535 * g_regex_match_all_full:
1536 * @regex: a #GRegex structure from g_regex_new()
1537 * @string: (array length=string_len): the string to scan for matches
1538 * @string_len: the length of @string, or -1 if @string is nul-terminated
1539 * @start_position: starting index of the string to match
1540 * @match_options: match options
1541 * @match_info: (out) (allow-none): pointer to location where to store
1542 * the #GMatchInfo, or %NULL if you do not need it
1543 * @error: location to store the error occuring, or %NULL to ignore errors
1545 * Using the standard algorithm for regular expression matching only
1546 * the longest match in the string is retrieved, it is not possibile
1547 * to obtain all the available matches. For instance matching
1548 * "<a> <b> <c>" against the pattern "<.*>"
1549 * you get "<a> <b> <c>".
1551 * This function uses a different algorithm (called DFA, i.e. deterministic
1552 * finite automaton), so it can retrieve all the possible matches, all
1553 * starting at the same point in the string. For instance matching
1554 * "<a> <b> <c>" against the pattern "<.*>"
1555 * you would obtain three matches: "<a> <b> <c>",
1556 * "<a> <b>" and "<a>".
1558 * The number of matched strings is retrieved using
1559 * g_match_info_get_match_count(). To obtain the matched strings and
1560 * their position you can use, respectively, g_match_info_fetch() and
1561 * g_match_info_fetch_pos(). Note that the strings are returned in
1562 * reverse order of length; that is, the longest matching string is
1565 * Note that the DFA algorithm is slower than the standard one and it
1566 * is not able to capture substrings, so backreferences do not work.
1568 * Setting @start_position differs from just passing over a shortened
1569 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1570 * that begins with any kind of lookbehind assertion, such as "\b".
1572 * A #GMatchInfo structure, used to get information on the match, is
1573 * stored in @match_info if not %NULL. Note that if @match_info is
1574 * not %NULL then it is created even if the function returns %FALSE,
1575 * i.e. you must free it regardless if regular expression actually
1578 * @string is not copied and is used in #GMatchInfo internally. If
1579 * you use any #GMatchInfo method (except g_match_info_free()) after
1580 * freeing or modifying @string then the behaviour is undefined.
1582 * Returns: %TRUE is the string matched, %FALSE otherwise
1587 g_regex_match_all_full (const GRegex *regex,
1588 const gchar *string,
1590 gint start_position,
1591 GRegexMatchFlags match_options,
1592 GMatchInfo **match_info,
1598 g_return_val_if_fail (regex != NULL, FALSE);
1599 g_return_val_if_fail (string != NULL, FALSE);
1600 g_return_val_if_fail (start_position >= 0, FALSE);
1601 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1602 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1604 info = match_info_new (regex, string, string_len, start_position,
1605 match_options, TRUE);
1611 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1612 info->string, info->string_len,
1614 regex->match_opts | match_options,
1615 info->offsets, info->n_offsets,
1616 info->workspace, info->n_workspace);
1617 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1619 /* info->workspace is too small. */
1620 info->n_workspace *= 2;
1621 info->workspace = g_realloc (info->workspace,
1622 info->n_workspace * sizeof (gint));
1625 else if (info->matches == 0)
1627 /* info->offsets is too small. */
1628 info->n_offsets *= 2;
1629 info->offsets = g_realloc (info->offsets,
1630 info->n_offsets * sizeof (gint));
1633 else if (IS_PCRE_ERROR (info->matches))
1635 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1636 _("Error while matching regular expression %s: %s"),
1637 regex->pattern, match_error (info->matches));
1641 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1644 if (match_info != NULL)
1647 g_match_info_free (info);
1649 return info->matches >= 0;
1653 * g_regex_get_string_number:
1654 * @regex: #GRegex structure
1655 * @name: name of the subexpression
1657 * Retrieves the number of the subexpression named @name.
1659 * Returns: The number of the subexpression or -1 if @name
1665 g_regex_get_string_number (const GRegex *regex,
1670 g_return_val_if_fail (regex != NULL, -1);
1671 g_return_val_if_fail (name != NULL, -1);
1673 num = pcre_get_stringnumber (regex->pcre_re, name);
1674 if (num == PCRE_ERROR_NOSUBSTRING)
1681 * g_regex_split_simple:
1682 * @pattern: the regular expression
1683 * @string: the string to scan for matches
1684 * @compile_options: compile options for the regular expression, or 0
1685 * @match_options: match options, or 0
1687 * Breaks the string on the pattern, and returns an array of
1688 * the tokens. If the pattern contains capturing parentheses,
1689 * then the text for each of the substrings will also be returned.
1690 * If the pattern does not match anywhere in the string, then the
1691 * whole string is returned as the first token.
1693 * This function is equivalent to g_regex_split() but it does
1694 * not require to compile the pattern with g_regex_new(), avoiding
1695 * some lines of code when you need just to do a split without
1696 * extracting substrings, capture counts, and so on.
1698 * If this function is to be called on the same @pattern more than
1699 * once, it's more efficient to compile the pattern once with
1700 * g_regex_new() and then use g_regex_split().
1702 * As a special case, the result of splitting the empty string ""
1703 * is an empty vector, not a vector containing a single string.
1704 * The reason for this special case is that being able to represent
1705 * a empty vector is typically more useful than consistent handling
1706 * of empty elements. If you do need to represent empty elements,
1707 * you'll need to check for the empty string before calling this
1710 * A pattern that can match empty strings splits @string into
1711 * separate characters wherever it matches the empty string between
1712 * characters. For example splitting "ab c" using as a separator
1713 * "\s*", you will get "a", "b" and "c".
1715 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1720 g_regex_split_simple (const gchar *pattern,
1721 const gchar *string,
1722 GRegexCompileFlags compile_options,
1723 GRegexMatchFlags match_options)
1728 regex = g_regex_new (pattern, compile_options, 0, NULL);
1731 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1732 g_regex_unref (regex);
1738 * @regex: a #GRegex structure
1739 * @string: the string to split with the pattern
1740 * @match_options: match time option flags
1742 * Breaks the string on the pattern, and returns an array of the tokens.
1743 * If the pattern contains capturing parentheses, then the text for each
1744 * of the substrings will also be returned. If the pattern does not match
1745 * anywhere in the string, then the whole string is returned as the first
1748 * As a special case, the result of splitting the empty string "" is an
1749 * empty vector, not a vector containing a single string. The reason for
1750 * this special case is that being able to represent a empty vector is
1751 * typically more useful than consistent handling of empty elements. If
1752 * you do need to represent empty elements, you'll need to check for the
1753 * empty string before calling this function.
1755 * A pattern that can match empty strings splits @string into separate
1756 * characters wherever it matches the empty string between characters.
1757 * For example splitting "ab c" using as a separator "\s*", you will get
1760 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1765 g_regex_split (const GRegex *regex,
1766 const gchar *string,
1767 GRegexMatchFlags match_options)
1769 return g_regex_split_full (regex, string, -1, 0,
1770 match_options, 0, NULL);
1774 * g_regex_split_full:
1775 * @regex: a #GRegex structure
1776 * @string: (array length=string_len): the string to split with the pattern
1777 * @string_len: the length of @string, or -1 if @string is nul-terminated
1778 * @start_position: starting index of the string to match
1779 * @match_options: match time option flags
1780 * @max_tokens: the maximum number of tokens to split @string into.
1781 * If this is less than 1, the string is split completely
1782 * @error: return location for a #GError
1784 * Breaks the string on the pattern, and returns an array of the tokens.
1785 * If the pattern contains capturing parentheses, then the text for each
1786 * of the substrings will also be returned. If the pattern does not match
1787 * anywhere in the string, then the whole string is returned as the first
1790 * As a special case, the result of splitting the empty string "" is an
1791 * empty vector, not a vector containing a single string. The reason for
1792 * this special case is that being able to represent a empty vector is
1793 * typically more useful than consistent handling of empty elements. If
1794 * you do need to represent empty elements, you'll need to check for the
1795 * empty string before calling this function.
1797 * A pattern that can match empty strings splits @string into separate
1798 * characters wherever it matches the empty string between characters.
1799 * For example splitting "ab c" using as a separator "\s*", you will get
1802 * Setting @start_position differs from just passing over a shortened
1803 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1804 * that begins with any kind of lookbehind assertion, such as "\b".
1806 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1811 g_regex_split_full (const GRegex *regex,
1812 const gchar *string,
1814 gint start_position,
1815 GRegexMatchFlags match_options,
1819 GError *tmp_error = NULL;
1820 GMatchInfo *match_info;
1825 /* position of the last separator. */
1826 gint last_separator_end;
1827 /* was the last match 0 bytes long? */
1828 gboolean last_match_is_empty;
1829 /* the returned array of char **s */
1830 gchar **string_list;
1832 g_return_val_if_fail (regex != NULL, NULL);
1833 g_return_val_if_fail (string != NULL, NULL);
1834 g_return_val_if_fail (start_position >= 0, NULL);
1835 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1836 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1838 if (max_tokens <= 0)
1839 max_tokens = G_MAXINT;
1842 string_len = strlen (string);
1844 /* zero-length string */
1845 if (string_len - start_position == 0)
1846 return g_new0 (gchar *, 1);
1848 if (max_tokens == 1)
1850 string_list = g_new0 (gchar *, 2);
1851 string_list[0] = g_strndup (&string[start_position],
1852 string_len - start_position);
1858 last_separator_end = start_position;
1859 last_match_is_empty = FALSE;
1861 match_ok = g_regex_match_full (regex, string, string_len, start_position,
1862 match_options, &match_info, &tmp_error);
1863 while (tmp_error == NULL)
1867 last_match_is_empty =
1868 (match_info->offsets[0] == match_info->offsets[1]);
1870 /* we need to skip empty separators at the same position of the end
1871 * of another separator. e.g. the string is "a b" and the separator
1872 * is " *", so from 1 to 2 we have a match and at position 2 we have
1873 * an empty match. */
1874 if (last_separator_end != match_info->offsets[1])
1879 token = g_strndup (string + last_separator_end,
1880 match_info->offsets[0] - last_separator_end);
1881 list = g_list_prepend (list, token);
1884 /* if there were substrings, these need to be added to
1886 match_count = g_match_info_get_match_count (match_info);
1887 if (match_count > 1)
1889 for (i = 1; i < match_count; i++)
1890 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
1896 /* if there was no match, copy to end of string. */
1897 if (!last_match_is_empty)
1899 gchar *token = g_strndup (string + last_separator_end,
1900 match_info->string_len - last_separator_end);
1901 list = g_list_prepend (list, token);
1903 /* no more tokens, end the loop. */
1907 /* -1 to leave room for the last part. */
1908 if (token_count >= max_tokens - 1)
1910 /* we have reached the maximum number of tokens, so we copy
1911 * the remaining part of the string. */
1912 if (last_match_is_empty)
1914 /* the last match was empty, so we have moved one char
1915 * after the real position to avoid empty matches at the
1917 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
1919 /* the if is needed in the case we have terminated the available
1920 * tokens, but we are at the end of the string, so there are no
1921 * characters left to copy. */
1922 if (string_len > match_info->pos)
1924 gchar *token = g_strndup (string + match_info->pos,
1925 string_len - match_info->pos);
1926 list = g_list_prepend (list, token);
1932 last_separator_end = match_info->pos;
1933 if (last_match_is_empty)
1934 /* if the last match was empty, g_match_info_next() has moved
1935 * forward to avoid infinite loops, but we still need to copy that
1937 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
1939 match_ok = g_match_info_next (match_info, &tmp_error);
1941 g_match_info_free (match_info);
1942 if (tmp_error != NULL)
1944 g_propagate_error (error, tmp_error);
1945 g_list_foreach (list, (GFunc)g_free, NULL);
1947 match_info->pos = -1;
1951 string_list = g_new (gchar *, g_list_length (list) + 1);
1953 for (last = g_list_last (list); last; last = g_list_previous (last))
1954 string_list[i++] = last->data;
1955 string_list[i] = NULL;
1964 REPL_TYPE_CHARACTER,
1965 REPL_TYPE_SYMBOLIC_REFERENCE,
1966 REPL_TYPE_NUMERIC_REFERENCE,
1967 REPL_TYPE_CHANGE_CASE
1972 CHANGE_CASE_NONE = 1 << 0,
1973 CHANGE_CASE_UPPER = 1 << 1,
1974 CHANGE_CASE_LOWER = 1 << 2,
1975 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
1976 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
1977 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
1978 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
1979 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
1982 struct _InterpolationData
1988 ChangeCase change_case;
1992 free_interpolation_data (InterpolationData *data)
1994 g_free (data->text);
1998 static const gchar *
1999 expand_escape (const gchar *replacement,
2001 InterpolationData *data,
2006 const gchar *error_detail;
2008 GError *tmp_error = NULL;
2016 data->type = REPL_TYPE_CHARACTER;
2021 data->type = REPL_TYPE_CHARACTER;
2026 data->type = REPL_TYPE_CHARACTER;
2031 data->type = REPL_TYPE_CHARACTER;
2036 data->type = REPL_TYPE_CHARACTER;
2041 data->type = REPL_TYPE_CHARACTER;
2046 data->type = REPL_TYPE_CHARACTER;
2051 data->type = REPL_TYPE_CHARACTER;
2061 h = g_ascii_xdigit_value (*p);
2064 error_detail = _("hexadecimal digit or '}' expected");
2075 for (i = 0; i < 2; i++)
2077 h = g_ascii_xdigit_value (*p);
2080 error_detail = _("hexadecimal digit expected");
2087 data->type = REPL_TYPE_STRING;
2088 data->text = g_new0 (gchar, 8);
2089 g_unichar_to_utf8 (x, data->text);
2093 data->type = REPL_TYPE_CHANGE_CASE;
2094 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2098 data->type = REPL_TYPE_CHANGE_CASE;
2099 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2103 data->type = REPL_TYPE_CHANGE_CASE;
2104 data->change_case = CHANGE_CASE_LOWER;
2108 data->type = REPL_TYPE_CHANGE_CASE;
2109 data->change_case = CHANGE_CASE_UPPER;
2113 data->type = REPL_TYPE_CHANGE_CASE;
2114 data->change_case = CHANGE_CASE_NONE;
2120 error_detail = _("missing '<' in symbolic reference");
2129 error_detail = _("unfinished symbolic reference");
2136 error_detail = _("zero-length symbolic reference");
2139 if (g_ascii_isdigit (*q))
2144 h = g_ascii_digit_value (*q);
2147 error_detail = _("digit expected");
2156 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2163 if (!g_ascii_isalnum (*r))
2165 error_detail = _("illegal symbolic reference");
2172 data->text = g_strndup (q, p - q);
2173 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2178 /* if \0 is followed by a number is an octal number representing a
2179 * character, else it is a numeric reference. */
2180 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2183 p = g_utf8_next_char (p);
2196 for (i = 0; i < 3; i++)
2198 h = g_ascii_digit_value (*p);
2208 if (i == 2 && base == 10)
2214 if (base == 8 || i == 3)
2216 data->type = REPL_TYPE_STRING;
2217 data->text = g_new0 (gchar, 8);
2218 g_unichar_to_utf8 (x, data->text);
2222 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2227 error_detail = _("stray final '\\'");
2231 error_detail = _("unknown escape sequence");
2238 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2239 tmp_error = g_error_new (G_REGEX_ERROR,
2240 G_REGEX_ERROR_REPLACE,
2241 _("Error while parsing replacement "
2242 "text \"%s\" at char %lu: %s"),
2244 (gulong)(p - replacement),
2246 g_propagate_error (error, tmp_error);
2252 split_replacement (const gchar *replacement,
2256 InterpolationData *data;
2257 const gchar *p, *start;
2259 start = p = replacement;
2264 data = g_new0 (InterpolationData, 1);
2265 start = p = expand_escape (replacement, p, data, error);
2268 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2270 free_interpolation_data (data);
2274 list = g_list_prepend (list, data);
2279 if (*p == '\\' || *p == '\0')
2283 data = g_new0 (InterpolationData, 1);
2284 data->text = g_strndup (start, p - start);
2285 data->type = REPL_TYPE_STRING;
2286 list = g_list_prepend (list, data);
2292 return g_list_reverse (list);
2295 /* Change the case of c based on change_case. */
2296 #define CHANGE_CASE(c, change_case) \
2297 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2298 g_unichar_tolower (c) : \
2299 g_unichar_toupper (c))
2302 string_append (GString *string,
2304 ChangeCase *change_case)
2308 if (text[0] == '\0')
2311 if (*change_case == CHANGE_CASE_NONE)
2313 g_string_append (string, text);
2315 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2317 c = g_utf8_get_char (text);
2318 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2319 g_string_append (string, g_utf8_next_char (text));
2320 *change_case = CHANGE_CASE_NONE;
2324 while (*text != '\0')
2326 c = g_utf8_get_char (text);
2327 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2328 text = g_utf8_next_char (text);
2334 interpolate_replacement (const GMatchInfo *match_info,
2339 InterpolationData *idata;
2341 ChangeCase change_case = CHANGE_CASE_NONE;
2343 for (list = data; list; list = list->next)
2346 switch (idata->type)
2348 case REPL_TYPE_STRING:
2349 string_append (result, idata->text, &change_case);
2351 case REPL_TYPE_CHARACTER:
2352 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2353 if (change_case & CHANGE_CASE_SINGLE_MASK)
2354 change_case = CHANGE_CASE_NONE;
2356 case REPL_TYPE_NUMERIC_REFERENCE:
2357 match = g_match_info_fetch (match_info, idata->num);
2360 string_append (result, match, &change_case);
2364 case REPL_TYPE_SYMBOLIC_REFERENCE:
2365 match = g_match_info_fetch_named (match_info, idata->text);
2368 string_append (result, match, &change_case);
2372 case REPL_TYPE_CHANGE_CASE:
2373 change_case = idata->change_case;
2381 /* whether actual match_info is needed for replacement, i.e.
2382 * whether there are references
2385 interpolation_list_needs_match (GList *list)
2387 while (list != NULL)
2389 InterpolationData *data = list->data;
2391 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2392 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2405 * @regex: a #GRegex structure
2406 * @string: (array length=string_len): the string to perform matches against
2407 * @string_len: the length of @string, or -1 if @string is nul-terminated
2408 * @start_position: starting index of the string to match
2409 * @replacement: text to replace each match with
2410 * @match_options: options for the match
2411 * @error: location to store the error occuring, or %NULL to ignore errors
2413 * Replaces all occurrences of the pattern in @regex with the
2414 * replacement text. Backreferences of the form '\number' or
2415 * '\g<number>' in the replacement text are interpolated by the
2416 * number-th captured subexpression of the match, '\g<name>' refers
2417 * to the captured subexpression with the given name. '\0' refers to the
2418 * complete match, but '\0' followed by a number is the octal representation
2419 * of a character. To include a literal '\' in the replacement, write '\\'.
2420 * There are also escapes that changes the case of the following text:
2423 * <varlistentry><term>\l</term>
2425 * <para>Convert to lower case the next character</para>
2428 * <varlistentry><term>\u</term>
2430 * <para>Convert to upper case the next character</para>
2433 * <varlistentry><term>\L</term>
2435 * <para>Convert to lower case till \E</para>
2438 * <varlistentry><term>\U</term>
2440 * <para>Convert to upper case till \E</para>
2443 * <varlistentry><term>\E</term>
2445 * <para>End case modification</para>
2450 * If you do not need to use backreferences use g_regex_replace_literal().
2452 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2453 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2454 * you can use g_regex_replace_literal().
2456 * Setting @start_position differs from just passing over a shortened
2457 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2458 * begins with any kind of lookbehind assertion, such as "\b".
2460 * Returns: a newly allocated string containing the replacements
2465 g_regex_replace (const GRegex *regex,
2466 const gchar *string,
2468 gint start_position,
2469 const gchar *replacement,
2470 GRegexMatchFlags match_options,
2475 GError *tmp_error = NULL;
2477 g_return_val_if_fail (regex != NULL, NULL);
2478 g_return_val_if_fail (string != NULL, NULL);
2479 g_return_val_if_fail (start_position >= 0, NULL);
2480 g_return_val_if_fail (replacement != NULL, NULL);
2481 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2482 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2484 list = split_replacement (replacement, &tmp_error);
2485 if (tmp_error != NULL)
2487 g_propagate_error (error, tmp_error);
2491 result = g_regex_replace_eval (regex,
2492 string, string_len, start_position,
2494 interpolate_replacement,
2497 if (tmp_error != NULL)
2498 g_propagate_error (error, tmp_error);
2500 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2507 literal_replacement (const GMatchInfo *match_info,
2511 g_string_append (result, data);
2516 * g_regex_replace_literal:
2517 * @regex: a #GRegex structure
2518 * @string: (array length=string_len): the string to perform matches against
2519 * @string_len: the length of @string, or -1 if @string is nul-terminated
2520 * @start_position: starting index of the string to match
2521 * @replacement: text to replace each match with
2522 * @match_options: options for the match
2523 * @error: location to store the error occuring, or %NULL to ignore errors
2525 * Replaces all occurrences of the pattern in @regex with the
2526 * replacement text. @replacement is replaced literally, to
2527 * include backreferences use g_regex_replace().
2529 * Setting @start_position differs from just passing over a
2530 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2531 * case of a pattern that begins with any kind of lookbehind
2532 * assertion, such as "\b".
2534 * Returns: a newly allocated string containing the replacements
2539 g_regex_replace_literal (const GRegex *regex,
2540 const gchar *string,
2542 gint start_position,
2543 const gchar *replacement,
2544 GRegexMatchFlags match_options,
2547 g_return_val_if_fail (replacement != NULL, NULL);
2548 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2550 return g_regex_replace_eval (regex,
2551 string, string_len, start_position,
2553 literal_replacement,
2554 (gpointer)replacement,
2559 * g_regex_replace_eval:
2560 * @regex: a #GRegex structure from g_regex_new()
2561 * @string (array length=string_len): string to perform matches against
2562 * @string_len: the length of @string, or -1 if @string is nul-terminated
2563 * @start_position: starting index of the string to match
2564 * @match_options: options for the match
2565 * @eval: a function to call for each match
2566 * @user_data: user data to pass to the function
2567 * @error: location to store the error occuring, or %NULL to ignore errors
2569 * Replaces occurrences of the pattern in regex with the output of
2570 * @eval for that occurrence.
2572 * Setting @start_position differs from just passing over a shortened
2573 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2574 * that begins with any kind of lookbehind assertion, such as "\b".
2576 * The following example uses g_regex_replace_eval() to replace multiple
2580 * eval_cb (const GMatchInfo *info,
2587 * match = g_match_info_fetch (info, 0);
2588 * r = g_hash_table_lookup ((GHashTable *)data, match);
2589 * g_string_append (res, r);
2601 * h = g_hash_table_new (g_str_hash, g_str_equal);
2603 * g_hash_table_insert (h, "1", "ONE");
2604 * g_hash_table_insert (h, "2", "TWO");
2605 * g_hash_table_insert (h, "3", "THREE");
2606 * g_hash_table_insert (h, "4", "FOUR");
2608 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2609 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2610 * g_hash_table_destroy (h);
2615 * Returns: a newly allocated string containing the replacements
2620 g_regex_replace_eval (const GRegex *regex,
2621 const gchar *string,
2623 gint start_position,
2624 GRegexMatchFlags match_options,
2625 GRegexEvalCallback eval,
2629 GMatchInfo *match_info;
2632 gboolean done = FALSE;
2633 GError *tmp_error = NULL;
2635 g_return_val_if_fail (regex != NULL, NULL);
2636 g_return_val_if_fail (string != NULL, NULL);
2637 g_return_val_if_fail (start_position >= 0, NULL);
2638 g_return_val_if_fail (eval != NULL, NULL);
2639 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2642 string_len = strlen (string);
2644 result = g_string_sized_new (string_len);
2646 /* run down the string making matches. */
2647 g_regex_match_full (regex, string, string_len, start_position,
2648 match_options, &match_info, &tmp_error);
2649 while (!done && g_match_info_matches (match_info))
2651 g_string_append_len (result,
2653 match_info->offsets[0] - str_pos);
2654 done = (*eval) (match_info, result, user_data);
2655 str_pos = match_info->offsets[1];
2656 g_match_info_next (match_info, &tmp_error);
2658 g_match_info_free (match_info);
2659 if (tmp_error != NULL)
2661 g_propagate_error (error, tmp_error);
2662 g_string_free (result, TRUE);
2666 g_string_append_len (result, string + str_pos, string_len - str_pos);
2667 return g_string_free (result, FALSE);
2671 * g_regex_check_replacement:
2672 * @replacement: the replacement string
2673 * @has_references: (out) (allow-none): location to store information about
2674 * references in @replacement or %NULL
2675 * @error: location to store error
2677 * Checks whether @replacement is a valid replacement string
2678 * (see g_regex_replace()), i.e. that all escape sequences in
2681 * If @has_references is not %NULL then @replacement is checked
2682 * for pattern references. For instance, replacement text 'foo\n'
2683 * does not contain references and may be evaluated without information
2684 * about actual match, but '\0\1' (whole match followed by first
2685 * subpattern) requires valid #GMatchInfo object.
2687 * Returns: whether @replacement is a valid replacement string
2692 g_regex_check_replacement (const gchar *replacement,
2693 gboolean *has_references,
2699 list = split_replacement (replacement, &tmp);
2703 g_propagate_error (error, tmp);
2708 *has_references = interpolation_list_needs_match (list);
2710 g_list_foreach (list, (GFunc) free_interpolation_data, NULL);
2717 * g_regex_escape_string:
2718 * @string: (array length=length): the string to escape
2719 * @length: the length of @string, or -1 if @string is nul-terminated
2721 * Escapes the special characters used for regular expressions
2722 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2723 * function is useful to dynamically generate regular expressions.
2725 * @string can contain nul characters that are replaced with "\0",
2726 * in this case remember to specify the correct length of @string
2729 * Returns: a newly-allocated escaped string
2734 g_regex_escape_string (const gchar *string,
2738 const char *p, *piece_start, *end;
2740 g_return_val_if_fail (string != NULL, NULL);
2743 length = strlen (string);
2745 end = string + length;
2746 p = piece_start = string;
2747 escaped = g_string_sized_new (length + 1);
2768 if (p != piece_start)
2769 /* copy the previous piece. */
2770 g_string_append_len (escaped, piece_start, p - piece_start);
2771 g_string_append_c (escaped, '\\');
2773 g_string_append_c (escaped, '0');
2775 g_string_append_c (escaped, *p);
2779 p = g_utf8_next_char (p);
2784 if (piece_start < end)
2785 g_string_append_len (escaped, piece_start, end - piece_start);
2787 return g_string_free (escaped, FALSE);