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
26 #ifdef USE_SYSTEM_PCRE
29 #include "pcre/pcre.h"
36 #include "gmessages.h"
37 #include "gstrfuncs.h"
42 * @title: Perl-compatible regular expressions
43 * @short_description: matches strings against regular expressions
44 * @see_also: <xref linkend="glib-regex-syntax"/>
46 * The <function>g_regex_*()</function> functions implement regular
47 * expression pattern matching using syntax and semantics similar to
48 * Perl regular expression.
50 * Some functions accept a @start_position argument, setting it differs
51 * from just passing over a shortened string and setting #G_REGEX_MATCH_NOTBOL
52 * in the case of a pattern that begins with any kind of lookbehind assertion.
53 * For example, consider the pattern "\Biss\B" which finds occurrences of "iss"
54 * in the middle of words. ("\B" matches only if the current position in the
55 * subject is not a word boundary.) When applied to the string "Mississipi"
56 * from the fourth byte, namely "issipi", it does not match, because "\B" is
57 * always false at the start of the subject, which is deemed to be a word
58 * boundary. However, if the entire string is passed , but with
59 * @start_position set to 4, it finds the second occurrence of "iss" because
60 * it is able to look behind the starting point to discover that it is
61 * preceded by a letter.
63 * Note that, unless you set the #G_REGEX_RAW flag, all the strings passed
64 * to these functions must be encoded in UTF-8. The lengths and the positions
65 * inside the strings are in bytes and not in characters, so, for instance,
66 * "\xc3\xa0" (i.e. "à") is two bytes long but it is treated as a
67 * single character. If you set #G_REGEX_RAW the strings can be non-valid
68 * UTF-8 strings and a byte is treated as a character, so "\xc3\xa0" is two
69 * bytes and two characters long.
71 * When matching a pattern, "\n" matches only against a "\n" character in
72 * the string, and "\r" matches only a "\r" character. To match any newline
73 * sequence use "\R". This particular group matches either the two-character
74 * sequence CR + LF ("\r\n"), or one of the single characters LF (linefeed,
75 * U+000A, "\n"), VT vertical tab, U+000B, "\v"), FF (formfeed, U+000C, "\f"),
76 * CR (carriage return, U+000D, "\r"), NEL (next line, U+0085), LS (line
77 * separator, U+2028), or PS (paragraph separator, U+2029).
79 * The behaviour of the dot, circumflex, and dollar metacharacters are
80 * affected by newline characters, the default is to recognize any newline
81 * character (the same characters recognized by "\R"). This can be changed
82 * with #G_REGEX_NEWLINE_CR, #G_REGEX_NEWLINE_LF and #G_REGEX_NEWLINE_CRLF
83 * compile options, and with #G_REGEX_MATCH_NEWLINE_ANY,
84 * #G_REGEX_MATCH_NEWLINE_CR, #G_REGEX_MATCH_NEWLINE_LF and
85 * #G_REGEX_MATCH_NEWLINE_CRLF match options. These settings are also
86 * relevant when compiling a pattern if #G_REGEX_EXTENDED is set, and an
87 * unescaped "#" outside a character class is encountered. This indicates
88 * a comment that lasts until after the next newline.
90 * Creating and manipulating the same #GRegex structure from different
91 * threads is not a problem as #GRegex does not modify its internal
92 * state between creation and destruction, on the other hand #GMatchInfo
95 * The regular expressions low-level functionalities are obtained through
96 * the excellent <ulink url="http://www.pcre.org/">PCRE</ulink> library
97 * written by Philip Hazel.
100 /* Mask of all the possible values for GRegexCompileFlags. */
101 #define G_REGEX_COMPILE_MASK (G_REGEX_CASELESS | \
102 G_REGEX_MULTILINE | \
106 G_REGEX_DOLLAR_ENDONLY | \
109 G_REGEX_NO_AUTO_CAPTURE | \
112 G_REGEX_NEWLINE_CR | \
113 G_REGEX_NEWLINE_LF | \
114 G_REGEX_NEWLINE_CRLF)
116 /* Mask of all the possible values for GRegexMatchFlags. */
117 #define G_REGEX_MATCH_MASK (G_REGEX_MATCH_ANCHORED | \
118 G_REGEX_MATCH_NOTBOL | \
119 G_REGEX_MATCH_NOTEOL | \
120 G_REGEX_MATCH_NOTEMPTY | \
121 G_REGEX_MATCH_PARTIAL | \
122 G_REGEX_MATCH_NEWLINE_CR | \
123 G_REGEX_MATCH_NEWLINE_LF | \
124 G_REGEX_MATCH_NEWLINE_CRLF | \
125 G_REGEX_MATCH_NEWLINE_ANY)
127 /* if the string is in UTF-8 use g_utf8_ functions, else use
129 #define NEXT_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
130 g_utf8_next_char (s) : \
132 #define PREV_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
133 g_utf8_prev_char (s) : \
138 GRegex *regex; /* the regex */
139 GRegexMatchFlags match_opts; /* options used at match time on the regex */
140 gint matches; /* number of matching sub patterns */
141 gint pos; /* position in the string where last match left off */
142 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
143 gint n_offsets; /* number of offsets */
144 gint *workspace; /* workspace for pcre_dfa_exec() */
145 gint n_workspace; /* number of workspace elements */
146 const gchar *string; /* string passed to the match function */
147 gssize string_len; /* length of string */
152 volatile gint ref_count; /* the ref count for the immutable part */
153 gchar *pattern; /* the pattern */
154 pcre *pcre_re; /* compiled form of the pattern */
155 GRegexCompileFlags compile_opts; /* options used at compile time on the pattern */
156 GRegexMatchFlags match_opts; /* options used at match time on the regex */
157 pcre_extra *extra; /* data stored when G_REGEX_OPTIMIZE is used */
160 /* TRUE if ret is an error code, FALSE otherwise. */
161 #define IS_PCRE_ERROR(ret) ((ret) < PCRE_ERROR_NOMATCH && (ret) != PCRE_ERROR_PARTIAL)
163 typedef struct _InterpolationData InterpolationData;
164 static gboolean interpolation_list_needs_match (GList *list);
165 static gboolean interpolate_replacement (const GMatchInfo *match_info,
168 static GList *split_replacement (const gchar *replacement,
170 static void free_interpolation_data (InterpolationData *data);
174 match_error (gint errcode)
178 case PCRE_ERROR_NOMATCH:
181 case PCRE_ERROR_NULL:
182 /* NULL argument, this should not happen in GRegex */
183 g_warning ("A NULL argument was passed to PCRE");
185 case PCRE_ERROR_BADOPTION:
186 return "bad options";
187 case PCRE_ERROR_BADMAGIC:
188 return _("corrupted object");
189 case PCRE_ERROR_UNKNOWN_OPCODE:
190 return N_("internal error or corrupted object");
191 case PCRE_ERROR_NOMEMORY:
192 return _("out of memory");
193 case PCRE_ERROR_NOSUBSTRING:
194 /* not used by pcre_exec() */
196 case PCRE_ERROR_MATCHLIMIT:
197 return _("backtracking limit reached");
198 case PCRE_ERROR_CALLOUT:
199 /* callouts are not implemented */
201 case PCRE_ERROR_BADUTF8:
202 case PCRE_ERROR_BADUTF8_OFFSET:
203 /* we do not check if strings are valid */
205 case PCRE_ERROR_PARTIAL:
208 case PCRE_ERROR_BADPARTIAL:
209 return _("the pattern contains items not supported for partial matching");
210 case PCRE_ERROR_INTERNAL:
211 return _("internal error");
212 case PCRE_ERROR_BADCOUNT:
213 /* negative ovecsize, this should not happen in GRegex */
214 g_warning ("A negative ovecsize was passed to PCRE");
216 case PCRE_ERROR_DFA_UITEM:
217 return _("the pattern contains items not supported for partial matching");
218 case PCRE_ERROR_DFA_UCOND:
219 return _("back references as conditions are not supported for partial matching");
220 case PCRE_ERROR_DFA_UMLIMIT:
221 /* the match_field field is not used in GRegex */
223 case PCRE_ERROR_DFA_WSSIZE:
224 /* handled expanding the workspace */
226 case PCRE_ERROR_DFA_RECURSE:
227 case PCRE_ERROR_RECURSIONLIMIT:
228 return _("recursion limit reached");
229 case PCRE_ERROR_NULLWSLIMIT:
230 return _("workspace limit for empty substrings reached");
231 case PCRE_ERROR_BADNEWLINE:
232 return _("invalid combination of newline flags");
236 return _("unknown error");
240 translate_compile_error (gint *errcode, const gchar **errmsg)
242 /* Compile errors are created adding 100 to the error code returned
244 * If errcode is known we put the translatable error message in
245 * erromsg. If errcode is unknown we put the generic
246 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
247 * untranslated error message returned by PCRE.
248 * Note that there can be more PCRE errors with the same GRegexError
249 * and that some PCRE errors are useless for us.
255 case G_REGEX_ERROR_STRAY_BACKSLASH:
256 *errmsg = _("\\ at end of pattern");
258 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
259 *errmsg = _("\\c at end of pattern");
261 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
262 *errmsg = _("unrecognized character follows \\");
265 /* A number of Perl escapes are not handled by PCRE.
266 * Therefore it explicitly raises ERR37.
268 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
269 *errmsg = _("case-changing escapes (\\l, \\L, \\u, \\U) are not allowed here");
271 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
272 *errmsg = _("numbers out of order in {} quantifier");
274 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
275 *errmsg = _("number too big in {} quantifier");
277 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
278 *errmsg = _("missing terminating ] for character class");
280 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
281 *errmsg = _("invalid escape sequence in character class");
283 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
284 *errmsg = _("range out of order in character class");
286 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
287 *errmsg = _("nothing to repeat");
289 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
290 *errmsg = _("unrecognized character after (?");
293 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
294 *errmsg = _("unrecognized character after (?<");
297 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
298 *errmsg = _("unrecognized character after (?P");
300 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
301 *errmsg = _("POSIX named classes are supported only within a class");
303 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
304 *errmsg = _("missing terminating )");
307 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
308 *errmsg = _(") without opening (");
311 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
312 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
313 * sequences here, '(?-54' would be an example for the second group.
315 *errmsg = _("(?R or (?[+-]digits must be followed by )");
317 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
318 *errmsg = _("reference to non-existent subpattern");
320 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
321 *errmsg = _("missing ) after comment");
323 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
324 *errmsg = _("regular expression too large");
326 case G_REGEX_ERROR_MEMORY_ERROR:
327 *errmsg = _("failed to get memory");
329 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
330 *errmsg = _("lookbehind assertion is not fixed length");
332 case G_REGEX_ERROR_MALFORMED_CONDITION:
333 *errmsg = _("malformed number or name after (?(");
335 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
336 *errmsg = _("conditional group contains more than two branches");
338 case G_REGEX_ERROR_ASSERTION_EXPECTED:
339 *errmsg = _("assertion expected after (?(");
341 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
342 *errmsg = _("unknown POSIX class name");
344 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
345 *errmsg = _("POSIX collating elements are not supported");
347 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
348 *errmsg = _("character value in \\x{...} sequence is too large");
350 case G_REGEX_ERROR_INVALID_CONDITION:
351 *errmsg = _("invalid condition (?(0)");
353 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
354 *errmsg = _("\\C not allowed in lookbehind assertion");
356 case G_REGEX_ERROR_INFINITE_LOOP:
357 *errmsg = _("recursive call could loop indefinitely");
359 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
360 *errmsg = _("missing terminator in subpattern name");
362 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
363 *errmsg = _("two named subpatterns have the same name");
365 case G_REGEX_ERROR_MALFORMED_PROPERTY:
366 *errmsg = _("malformed \\P or \\p sequence");
368 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
369 *errmsg = _("unknown property name after \\P or \\p");
371 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
372 *errmsg = _("subpattern name is too long (maximum 32 characters)");
374 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
375 *errmsg = _("too many named subpatterns (maximum 10,000)");
377 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
378 *errmsg = _("octal value is greater than \\377");
380 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
381 *errmsg = _("DEFINE group contains more than one branch");
383 case G_REGEX_ERROR_DEFINE_REPETION:
384 *errmsg = _("repeating a DEFINE group is not allowed");
386 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
387 *errmsg = _("inconsistent NEWLINE options");
389 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
390 *errmsg = _("\\g is not followed by a braced name or an optionally "
391 "braced non-zero number");
394 *errcode = G_REGEX_ERROR_INTERNAL;
395 *errmsg = _("unexpected repeat");
398 *errcode = G_REGEX_ERROR_INTERNAL;
399 *errmsg = _("code overflow");
402 *errcode = G_REGEX_ERROR_INTERNAL;
403 *errmsg = _("overran compiling workspace");
406 *errcode = G_REGEX_ERROR_INTERNAL;
407 *errmsg = _("previously-checked referenced subpattern not found");
410 /* This should not happen as we never pass a NULL erroffset */
411 g_warning ("erroffset passed as NULL");
412 *errcode = G_REGEX_ERROR_COMPILE;
415 /* This should not happen as we check options before passing them
416 * to pcre_compile2() */
417 g_warning ("unknown option bit(s) set");
418 *errcode = G_REGEX_ERROR_COMPILE;
423 /* These errors should not happen as we are using an UTF8-enabled PCRE
424 * and we do not check if strings are valid */
425 g_warning ("%s", *errmsg);
426 *errcode = G_REGEX_ERROR_COMPILE;
429 *errcode = G_REGEX_ERROR_COMPILE;
436 match_info_new (const GRegex *regex,
443 GMatchInfo *match_info;
446 string_len = strlen (string);
448 match_info = g_new0 (GMatchInfo, 1);
449 match_info->regex = g_regex_ref ((GRegex *)regex);
450 match_info->string = string;
451 match_info->string_len = string_len;
452 match_info->matches = PCRE_ERROR_NOMATCH;
453 match_info->pos = start_position;
454 match_info->match_opts = match_options;
458 /* These values should be enough for most cases, if they are not
459 * enough g_regex_match_all_full() will expand them. */
460 match_info->n_offsets = 24;
461 match_info->n_workspace = 100;
462 match_info->workspace = g_new (gint, match_info->n_workspace);
467 pcre_fullinfo (regex->pcre_re, regex->extra,
468 PCRE_INFO_CAPTURECOUNT, &capture_count);
469 match_info->n_offsets = (capture_count + 1) * 3;
472 match_info->offsets = g_new0 (gint, match_info->n_offsets);
473 /* Set an invalid position for the previous match. */
474 match_info->offsets[0] = -1;
475 match_info->offsets[1] = -1;
481 * g_match_info_get_regex:
482 * @match_info: a #GMatchInfo
484 * Returns #GRegex object used in @match_info. It belongs to Glib
485 * and must not be freed. Use g_regex_ref() if you need to keep it
486 * after you free @match_info object.
488 * Returns: #GRegex object used in @match_info
493 g_match_info_get_regex (const GMatchInfo *match_info)
495 g_return_val_if_fail (match_info != NULL, NULL);
496 return match_info->regex;
500 * g_match_info_get_string:
501 * @match_info: a #GMatchInfo
503 * Returns the string searched with @match_info. This is the
504 * string passed to g_regex_match() or g_regex_replace() so
505 * you may not free it before calling this function.
507 * Returns: the string searched with @match_info
512 g_match_info_get_string (const GMatchInfo *match_info)
514 g_return_val_if_fail (match_info != NULL, NULL);
515 return match_info->string;
520 * @match_info: a #GMatchInfo
522 * Frees all the memory associated with the #GMatchInfo structure.
527 g_match_info_free (GMatchInfo *match_info)
531 g_regex_unref (match_info->regex);
532 g_free (match_info->offsets);
533 g_free (match_info->workspace);
540 * @match_info: a #GMatchInfo structure
541 * @error: location to store the error occuring, or %NULL to ignore errors
543 * Scans for the next match using the same parameters of the previous
544 * call to g_regex_match_full() or g_regex_match() that returned
547 * The match is done on the string passed to the match function, so you
548 * cannot free it before calling this function.
550 * Returns: %TRUE is the string matched, %FALSE otherwise
555 g_match_info_next (GMatchInfo *match_info,
558 gint prev_match_start;
561 g_return_val_if_fail (match_info != NULL, FALSE);
562 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
563 g_return_val_if_fail (match_info->pos >= 0, FALSE);
565 prev_match_start = match_info->offsets[0];
566 prev_match_end = match_info->offsets[1];
568 match_info->matches = pcre_exec (match_info->regex->pcre_re,
569 match_info->regex->extra,
571 match_info->string_len,
573 match_info->regex->match_opts | match_info->match_opts,
575 match_info->n_offsets);
576 if (IS_PCRE_ERROR (match_info->matches))
578 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
579 _("Error while matching regular expression %s: %s"),
580 match_info->regex->pattern, match_error (match_info->matches));
584 /* avoid infinite loops if the pattern is an empty string or something
586 if (match_info->pos == match_info->offsets[1])
588 if (match_info->pos > match_info->string_len)
590 /* we have reached the end of the string */
591 match_info->pos = -1;
592 match_info->matches = PCRE_ERROR_NOMATCH;
596 match_info->pos = NEXT_CHAR (match_info->regex,
597 &match_info->string[match_info->pos]) -
602 match_info->pos = match_info->offsets[1];
605 /* it's possible to get two identical matches when we are matching
606 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
607 * the string is "RegExTest" we have:
608 * - search at position 0: match from 0 to 0
609 * - search at position 1: match from 3 to 3
610 * - search at position 3: match from 3 to 3 (duplicate)
611 * - search at position 4: match from 5 to 5
612 * - search at position 5: match from 5 to 5 (duplicate)
613 * - search at position 6: no match -> stop
614 * so we have to ignore the duplicates.
615 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
616 if (match_info->matches >= 0 &&
617 prev_match_start == match_info->offsets[0] &&
618 prev_match_end == match_info->offsets[1])
620 /* ignore this match and search the next one */
621 return g_match_info_next (match_info, error);
624 return match_info->matches >= 0;
628 * g_match_info_matches:
629 * @match_info: a #GMatchInfo structure
631 * Returns whether the previous match operation succeeded.
633 * Returns: %TRUE if the previous match operation succeeded,
639 g_match_info_matches (const GMatchInfo *match_info)
641 g_return_val_if_fail (match_info != NULL, FALSE);
643 return match_info->matches >= 0;
647 * g_match_info_get_match_count:
648 * @match_info: a #GMatchInfo structure
650 * Retrieves the number of matched substrings (including substring 0,
651 * that is the whole matched text), so 1 is returned if the pattern
652 * has no substrings in it and 0 is returned if the match failed.
654 * If the last match was obtained using the DFA algorithm, that is
655 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
656 * count is not that of the number of capturing parentheses but that of
657 * the number of matched substrings.
659 * Returns: Number of matched substrings, or -1 if an error occurred
664 g_match_info_get_match_count (const GMatchInfo *match_info)
666 g_return_val_if_fail (match_info, -1);
668 if (match_info->matches == PCRE_ERROR_NOMATCH)
671 else if (match_info->matches < PCRE_ERROR_NOMATCH)
676 return match_info->matches;
680 * g_match_info_is_partial_match:
681 * @match_info: a #GMatchInfo structure
683 * Usually if the string passed to g_regex_match*() matches as far as
684 * it goes, but is too short to match the entire pattern, %FALSE is
685 * returned. There are circumstances where it might be helpful to
686 * distinguish this case from other cases in which there is no match.
688 * Consider, for example, an application where a human is required to
689 * type in data for a field with specific formatting requirements. An
690 * example might be a date in the form ddmmmyy, defined by the pattern
691 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
692 * If the application sees the user’s keystrokes one by one, and can
693 * check that what has been typed so far is potentially valid, it is
694 * able to raise an error as soon as a mistake is made.
696 * GRegex supports the concept of partial matching by means of the
697 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
698 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
699 * for a complete match, %FALSE otherwise. But, when these functions
700 * return %FALSE, you can check if the match was partial calling
701 * g_match_info_is_partial_match().
703 * When using partial matching you cannot use g_match_info_fetch*().
705 * Because of the way certain internal optimizations are implemented
706 * the partial matching algorithm cannot be used with all patterns.
707 * So repeated single characters such as "a{2,4}" and repeated single
708 * meta-sequences such as "\d+" are not permitted if the maximum number
709 * of occurrences is greater than one. Optional items such as "\d?"
710 * (where the maximum is one) are permitted. Quantifiers with any values
711 * are permitted after parentheses, so the invalid examples above can be
712 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
713 * for a pattern that does not conform to the restrictions, matching
714 * functions return an error.
716 * Returns: %TRUE if the match was partial, %FALSE otherwise
721 g_match_info_is_partial_match (const GMatchInfo *match_info)
723 g_return_val_if_fail (match_info != NULL, FALSE);
725 return match_info->matches == PCRE_ERROR_PARTIAL;
729 * g_match_info_expand_references:
730 * @match_info: a #GMatchInfo or %NULL
731 * @string_to_expand: the string to expand
732 * @error: location to store the error occuring, or %NULL to ignore errors
734 * Returns a new string containing the text in @string_to_expand with
735 * references and escape sequences expanded. References refer to the last
736 * match done with @string against @regex and have the same syntax used by
739 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
740 * passed to g_regex_new().
742 * The backreferences are extracted from the string passed to the match
743 * function, so you cannot call this function after freeing the string.
745 * @match_info may be %NULL in which case @string_to_expand must not
746 * contain references. For instance "foo\n" does not refer to an actual
747 * pattern and '\n' merely will be replaced with \n character,
748 * while to expand "\0" (whole match) one needs the result of a match.
749 * Use g_regex_check_replacement() to find out whether @string_to_expand
750 * contains references.
752 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
757 g_match_info_expand_references (const GMatchInfo *match_info,
758 const gchar *string_to_expand,
763 GError *tmp_error = NULL;
765 g_return_val_if_fail (string_to_expand != NULL, NULL);
766 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
768 list = split_replacement (string_to_expand, &tmp_error);
769 if (tmp_error != NULL)
771 g_propagate_error (error, tmp_error);
775 if (!match_info && interpolation_list_needs_match (list))
777 g_critical ("String '%s' contains references to the match, can't "
778 "expand references without GMatchInfo object",
783 result = g_string_sized_new (strlen (string_to_expand));
784 interpolate_replacement (match_info, result, list);
786 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
789 return g_string_free (result, FALSE);
793 * g_match_info_fetch:
794 * @match_info: #GMatchInfo structure
795 * @match_num: number of the sub expression
797 * Retrieves the text matching the @match_num<!-- -->'th capturing
798 * parentheses. 0 is the full text of the match, 1 is the first paren
799 * set, 2 the second, and so on.
801 * If @match_num is a valid sub pattern but it didn't match anything
802 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
803 * string is returned.
805 * If the match was obtained using the DFA algorithm, that is using
806 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
807 * string is not that of a set of parentheses but that of a matched
808 * substring. Substrings are matched in reverse order of length, so
809 * 0 is the longest match.
811 * The string is fetched from the string passed to the match function,
812 * so you cannot call this function after freeing the string.
814 * Returns: (allow-none): The matched substring, or %NULL if an error
815 * occurred. You have to free the string yourself
820 g_match_info_fetch (const GMatchInfo *match_info,
823 /* we cannot use pcre_get_substring() because it allocates the
824 * string using pcre_malloc(). */
828 g_return_val_if_fail (match_info != NULL, NULL);
829 g_return_val_if_fail (match_num >= 0, NULL);
831 /* match_num does not exist or it didn't matched, i.e. matching "b"
832 * against "(a)?b" then group 0 is empty. */
833 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
835 else if (start == -1)
836 match = g_strdup ("");
838 match = g_strndup (&match_info->string[start], end - start);
844 * g_match_info_fetch_pos:
845 * @match_info: #GMatchInfo structure
846 * @match_num: number of the sub expression
847 * @start_pos: (out) (allow-none): pointer to location where to store
848 * the start position, or %NULL
849 * @end_pos: (out) (allow-none): pointer to location where to store
850 * the end position, or %NULL
852 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
853 * parentheses. 0 is the full text of the match, 1 is the first
854 * paren set, 2 the second, and so on.
856 * If @match_num is a valid sub pattern but it didn't match anything
857 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
858 * and @end_pos are set to -1 and %TRUE is returned.
860 * If the match was obtained using the DFA algorithm, that is using
861 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
862 * position is not that of a set of parentheses but that of a matched
863 * substring. Substrings are matched in reverse order of length, so
864 * 0 is the longest match.
866 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
867 * the position cannot be fetched, @start_pos and @end_pos are left
873 g_match_info_fetch_pos (const GMatchInfo *match_info,
878 g_return_val_if_fail (match_info != NULL, FALSE);
879 g_return_val_if_fail (match_num >= 0, FALSE);
881 /* make sure the sub expression number they're requesting is less than
882 * the total number of sub expressions that were matched. */
883 if (match_num >= match_info->matches)
886 if (start_pos != NULL)
887 *start_pos = match_info->offsets[2 * match_num];
890 *end_pos = match_info->offsets[2 * match_num + 1];
896 * Returns number of first matched subpattern with name @name.
897 * There may be more than one in case when DUPNAMES is used,
898 * and not all subpatterns with that name match;
899 * pcre_get_stringnumber() does not work in that case.
902 get_matched_substring_number (const GMatchInfo *match_info,
909 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
910 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
912 /* This code is copied from pcre_get.c: get_first_set() */
913 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
921 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
923 gint n = (entry[0] << 8) + entry[1];
924 if (match_info->offsets[n*2] >= 0)
928 return (first[0] << 8) + first[1];
932 * g_match_info_fetch_named:
933 * @match_info: #GMatchInfo structure
934 * @name: name of the subexpression
936 * Retrieves the text matching the capturing parentheses named @name.
938 * If @name is a valid sub pattern name but it didn't match anything
939 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
940 * then an empty string is returned.
942 * The string is fetched from the string passed to the match function,
943 * so you cannot call this function after freeing the string.
945 * Returns: (allow-none): The matched substring, or %NULL if an error
946 * occurred. You have to free the string yourself
951 g_match_info_fetch_named (const GMatchInfo *match_info,
954 /* we cannot use pcre_get_named_substring() because it allocates the
955 * string using pcre_malloc(). */
958 g_return_val_if_fail (match_info != NULL, NULL);
959 g_return_val_if_fail (name != NULL, NULL);
961 num = get_matched_substring_number (match_info, name);
965 return g_match_info_fetch (match_info, num);
969 * g_match_info_fetch_named_pos:
970 * @match_info: #GMatchInfo structure
971 * @name: name of the subexpression
972 * @start_pos: (out) (allow-none): pointer to location where to store
973 * the start position, or %NULL
974 * @end_pos: (out) (allow-none): pointer to location where to store
975 * the end position, or %NULL
977 * Retrieves the position in bytes of the capturing parentheses named @name.
979 * If @name is a valid sub pattern name but it didn't match anything
980 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
981 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
983 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
984 * If the position cannot be fetched, @start_pos and @end_pos
985 * are left unchanged.
990 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
997 g_return_val_if_fail (match_info != NULL, FALSE);
998 g_return_val_if_fail (name != NULL, FALSE);
1000 num = get_matched_substring_number (match_info, name);
1004 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1008 * g_match_info_fetch_all:
1009 * @match_info: a #GMatchInfo structure
1011 * Bundles up pointers to each of the matching substrings from a match
1012 * and stores them in an array of gchar pointers. The first element in
1013 * the returned array is the match number 0, i.e. the entire matched
1016 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1017 * "b" against "(a)?b") then an empty string is inserted.
1019 * If the last match was obtained using the DFA algorithm, that is using
1020 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1021 * strings are not that matched by sets of parentheses but that of the
1022 * matched substring. Substrings are matched in reverse order of length,
1023 * so the first one is the longest match.
1025 * The strings are fetched from the string passed to the match function,
1026 * so you cannot call this function after freeing the string.
1028 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1029 * It must be freed using g_strfreev(). If the previous match failed
1035 g_match_info_fetch_all (const GMatchInfo *match_info)
1037 /* we cannot use pcre_get_substring_list() because the returned value
1038 * isn't suitable for g_strfreev(). */
1042 g_return_val_if_fail (match_info != NULL, NULL);
1044 if (match_info->matches < 0)
1047 result = g_new (gchar *, match_info->matches + 1);
1048 for (i = 0; i < match_info->matches; i++)
1049 result[i] = g_match_info_fetch (match_info, i);
1059 g_regex_error_quark (void)
1061 static GQuark error_quark = 0;
1063 if (error_quark == 0)
1064 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1073 * Increases reference count of @regex by 1.
1080 g_regex_ref (GRegex *regex)
1082 g_return_val_if_fail (regex != NULL, NULL);
1083 g_atomic_int_inc (®ex->ref_count);
1091 * Decreases reference count of @regex by 1. When reference count drops
1092 * to zero, it frees all the memory associated with the regex structure.
1097 g_regex_unref (GRegex *regex)
1099 g_return_if_fail (regex != NULL);
1101 if (g_atomic_int_exchange_and_add (®ex->ref_count, -1) - 1 == 0)
1103 g_free (regex->pattern);
1104 if (regex->pcre_re != NULL)
1105 pcre_free (regex->pcre_re);
1106 if (regex->extra != NULL)
1107 pcre_free (regex->extra);
1114 * @pattern: the regular expression
1115 * @compile_options: compile options for the regular expression, or 0
1116 * @match_options: match options for the regular expression, or 0
1117 * @error: return location for a #GError
1119 * Compiles the regular expression to an internal form, and does
1120 * the initial setup of the #GRegex structure.
1122 * Returns: a #GRegex structure. Call g_regex_unref() when you
1128 g_regex_new (const gchar *pattern,
1129 GRegexCompileFlags compile_options,
1130 GRegexMatchFlags match_options,
1135 const gchar *errmsg;
1138 gboolean optimize = FALSE;
1139 static gboolean initialized = FALSE;
1140 unsigned long int pcre_compile_options;
1142 g_return_val_if_fail (pattern != NULL, NULL);
1143 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1144 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1145 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1152 pcre_config (PCRE_CONFIG_UTF8, &support);
1155 msg = N_("PCRE library is compiled without UTF8 support");
1156 g_critical ("%s", msg);
1157 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1161 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1164 msg = N_("PCRE library is compiled without UTF8 properties support");
1165 g_critical ("%s", msg);
1166 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1173 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1174 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1175 if (compile_options & G_REGEX_OPTIMIZE)
1178 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1179 * instead uses UTF-8 only if required with PCRE_UTF8. */
1180 if (compile_options & G_REGEX_RAW)
1183 compile_options &= ~G_REGEX_RAW;
1188 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1189 match_options |= PCRE_NO_UTF8_CHECK;
1192 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1193 * not for the system one. */
1194 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1195 !(compile_options & G_REGEX_NEWLINE_LF))
1197 compile_options |= PCRE_NEWLINE_ANY;
1200 /* compile the pattern */
1201 re = pcre_compile2 (pattern, compile_options, &errcode,
1202 &errmsg, &erroffset, NULL);
1204 /* if the compilation failed, set the error member and return
1210 /* Translate the PCRE error code to GRegexError and use a translated
1211 * error message if possible */
1212 translate_compile_error (&errcode, &errmsg);
1214 /* PCRE uses byte offsets but we want to show character offsets */
1215 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1217 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1218 _("Error while compiling regular "
1219 "expression %s at char %d: %s"),
1220 pattern, erroffset, errmsg);
1221 g_propagate_error (error, tmp_error);
1226 /* For options set at the beginning of the pattern, pcre puts them into
1227 * compile options, e.g. "(?i)foo" will make the pcre structure store
1228 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1229 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1230 compile_options = pcre_compile_options;
1232 if (!(compile_options & G_REGEX_DUPNAMES))
1234 gboolean jchanged = FALSE;
1235 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1237 compile_options |= G_REGEX_DUPNAMES;
1240 regex = g_new0 (GRegex, 1);
1241 regex->ref_count = 1;
1242 regex->pattern = g_strdup (pattern);
1243 regex->pcre_re = re;
1244 regex->compile_opts = compile_options;
1245 regex->match_opts = match_options;
1249 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1252 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1253 G_REGEX_ERROR_OPTIMIZE,
1254 _("Error while optimizing "
1255 "regular expression %s: %s"),
1258 g_propagate_error (error, tmp_error);
1260 g_regex_unref (regex);
1269 * g_regex_get_pattern:
1270 * @regex: a #GRegex structure
1272 * Gets the pattern string associated with @regex, i.e. a copy of
1273 * the string passed to g_regex_new().
1275 * Returns: the pattern of @regex
1280 g_regex_get_pattern (const GRegex *regex)
1282 g_return_val_if_fail (regex != NULL, NULL);
1284 return regex->pattern;
1288 * g_regex_get_max_backref:
1291 * Returns the number of the highest back reference
1292 * in the pattern, or 0 if the pattern does not contain
1295 * Returns: the number of the highest back reference
1300 g_regex_get_max_backref (const GRegex *regex)
1304 pcre_fullinfo (regex->pcre_re, regex->extra,
1305 PCRE_INFO_BACKREFMAX, &value);
1311 * g_regex_get_capture_count:
1314 * Returns the number of capturing subpatterns in the pattern.
1316 * Returns: the number of capturing subpatterns
1321 g_regex_get_capture_count (const GRegex *regex)
1325 pcre_fullinfo (regex->pcre_re, regex->extra,
1326 PCRE_INFO_CAPTURECOUNT, &value);
1332 * g_regex_get_compile_flags:
1335 * Returns the compile options that @regex was created with.
1337 * Returns: flags from #GRegexCompileFlags
1342 g_regex_get_compile_flags (const GRegex *regex)
1344 g_return_val_if_fail (regex != NULL, 0);
1346 return regex->compile_opts;
1350 * g_regex_get_match_flags:
1353 * Returns the match options that @regex was created with.
1355 * Returns: flags from #GRegexMatchFlags
1360 g_regex_get_match_flags (const GRegex *regex)
1362 g_return_val_if_fail (regex != NULL, 0);
1364 return regex->match_opts;
1368 * g_regex_match_simple:
1369 * @pattern: the regular expression
1370 * @string: the string to scan for matches
1371 * @compile_options: compile options for the regular expression, or 0
1372 * @match_options: match options, or 0
1374 * Scans for a match in @string for @pattern.
1376 * This function is equivalent to g_regex_match() but it does not
1377 * require to compile the pattern with g_regex_new(), avoiding some
1378 * lines of code when you need just to do a match without extracting
1379 * substrings, capture counts, and so on.
1381 * If this function is to be called on the same @pattern more than
1382 * once, it's more efficient to compile the pattern once with
1383 * g_regex_new() and then use g_regex_match().
1385 * Returns: %TRUE if the string matched, %FALSE otherwise
1390 g_regex_match_simple (const gchar *pattern,
1391 const gchar *string,
1392 GRegexCompileFlags compile_options,
1393 GRegexMatchFlags match_options)
1398 regex = g_regex_new (pattern, compile_options, 0, NULL);
1401 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1402 g_regex_unref (regex);
1408 * @regex: a #GRegex structure from g_regex_new()
1409 * @string: the string to scan for matches
1410 * @match_options: match options
1411 * @match_info: (out) (allow-none): pointer to location where to store
1412 * the #GMatchInfo, or %NULL if you do not need it
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 * A #GMatchInfo structure, used to get information on the match,
1420 * is stored in @match_info if not %NULL. Note that if @match_info
1421 * is not %NULL then it is created even if the function returns %FALSE,
1422 * i.e. you must free it regardless if regular expression actually matched.
1424 * To retrieve all the non-overlapping matches of the pattern in
1425 * string you can use g_match_info_next().
1429 * print_uppercase_words (const gchar *string)
1431 * /* Print all uppercase-only words. */
1433 * GMatchInfo *match_info;
1435 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1436 * g_regex_match (regex, string, 0, &match_info);
1437 * while (g_match_info_matches (match_info))
1439 * gchar *word = g_match_info_fetch (match_info, 0);
1440 * g_print ("Found: %s\n", word);
1442 * g_match_info_next (match_info, NULL);
1444 * g_match_info_free (match_info);
1445 * g_regex_unref (regex);
1449 * @string is not copied and is used in #GMatchInfo internally. If
1450 * you use any #GMatchInfo method (except g_match_info_free()) after
1451 * freeing or modifying @string then the behaviour is undefined.
1453 * Returns: %TRUE is the string matched, %FALSE otherwise
1458 g_regex_match (const GRegex *regex,
1459 const gchar *string,
1460 GRegexMatchFlags match_options,
1461 GMatchInfo **match_info)
1463 return g_regex_match_full (regex, string, -1, 0, match_options,
1468 * g_regex_match_full:
1469 * @regex: a #GRegex structure from g_regex_new()
1470 * @string: (array length=string_len): the string to scan for matches
1471 * @string_len: the length of @string, or -1 if @string is nul-terminated
1472 * @start_position: starting index of the string to match
1473 * @match_options: match options
1474 * @match_info: (out) (allow-none): pointer to location where to store
1475 * the #GMatchInfo, or %NULL if you do not need it
1476 * @error: location to store the error occuring, or %NULL to ignore errors
1478 * Scans for a match in string for the pattern in @regex.
1479 * The @match_options are combined with the match options specified
1480 * when the @regex structure was created, letting you have more
1481 * flexibility in reusing #GRegex structures.
1483 * Setting @start_position differs from just passing over a shortened
1484 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1485 * that begins with any kind of lookbehind assertion, such as "\b".
1487 * A #GMatchInfo structure, used to get information on the match, is
1488 * stored in @match_info if not %NULL. Note that if @match_info is
1489 * not %NULL then it is created even if the function returns %FALSE,
1490 * i.e. you must free it regardless if regular expression actually
1493 * @string is not copied and is used in #GMatchInfo internally. If
1494 * you use any #GMatchInfo method (except g_match_info_free()) after
1495 * freeing or modifying @string then the behaviour is undefined.
1497 * To retrieve all the non-overlapping matches of the pattern in
1498 * string you can use g_match_info_next().
1502 * print_uppercase_words (const gchar *string)
1504 * /* Print all uppercase-only words. */
1506 * GMatchInfo *match_info;
1507 * GError *error = NULL;
1509 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1510 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1511 * while (g_match_info_matches (match_info))
1513 * gchar *word = g_match_info_fetch (match_info, 0);
1514 * g_print ("Found: %s\n", word);
1516 * g_match_info_next (match_info, &error);
1518 * g_match_info_free (match_info);
1519 * g_regex_unref (regex);
1520 * if (error != NULL)
1522 * g_printerr ("Error while matching: %s\n", error->message);
1523 * g_error_free (error);
1528 * Returns: %TRUE is the string matched, %FALSE otherwise
1533 g_regex_match_full (const GRegex *regex,
1534 const gchar *string,
1536 gint start_position,
1537 GRegexMatchFlags match_options,
1538 GMatchInfo **match_info,
1544 g_return_val_if_fail (regex != NULL, FALSE);
1545 g_return_val_if_fail (string != NULL, FALSE);
1546 g_return_val_if_fail (start_position >= 0, FALSE);
1547 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1548 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1550 info = match_info_new (regex, string, string_len, start_position,
1551 match_options, FALSE);
1552 match_ok = g_match_info_next (info, error);
1553 if (match_info != NULL)
1556 g_match_info_free (info);
1562 * g_regex_match_all:
1563 * @regex: a #GRegex structure from g_regex_new()
1564 * @string: the string to scan for matches
1565 * @match_options: match options
1566 * @match_info: (out) (allow-none): pointer to location where to store
1567 * the #GMatchInfo, or %NULL if you do not need it
1569 * Using the standard algorithm for regular expression matching only
1570 * the longest match in the string is retrieved. This function uses
1571 * a different algorithm so it can retrieve all the possible matches.
1572 * For more documentation see g_regex_match_all_full().
1574 * A #GMatchInfo structure, used to get information on the match, is
1575 * stored in @match_info if not %NULL. Note that if @match_info is
1576 * not %NULL then it is created even if the function returns %FALSE,
1577 * i.e. you must free it regardless if regular expression actually
1580 * @string is not copied and is used in #GMatchInfo internally. If
1581 * you use any #GMatchInfo method (except g_match_info_free()) after
1582 * freeing or modifying @string then the behaviour is undefined.
1584 * Returns: %TRUE is the string matched, %FALSE otherwise
1589 g_regex_match_all (const GRegex *regex,
1590 const gchar *string,
1591 GRegexMatchFlags match_options,
1592 GMatchInfo **match_info)
1594 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1599 * g_regex_match_all_full:
1600 * @regex: a #GRegex structure from g_regex_new()
1601 * @string: (array length=string_len): the string to scan for matches
1602 * @string_len: the length of @string, or -1 if @string is nul-terminated
1603 * @start_position: starting index of the string to match
1604 * @match_options: match options
1605 * @match_info: (out) (allow-none): pointer to location where to store
1606 * the #GMatchInfo, or %NULL if you do not need it
1607 * @error: location to store the error occuring, or %NULL to ignore errors
1609 * Using the standard algorithm for regular expression matching only
1610 * the longest match in the string is retrieved, it is not possibile
1611 * to obtain all the available matches. For instance matching
1612 * "<a> <b> <c>" against the pattern "<.*>"
1613 * you get "<a> <b> <c>".
1615 * This function uses a different algorithm (called DFA, i.e. deterministic
1616 * finite automaton), so it can retrieve all the possible matches, all
1617 * starting at the same point in the string. For instance matching
1618 * "<a> <b> <c>" against the pattern "<.*>"
1619 * you would obtain three matches: "<a> <b> <c>",
1620 * "<a> <b>" and "<a>".
1622 * The number of matched strings is retrieved using
1623 * g_match_info_get_match_count(). To obtain the matched strings and
1624 * their position you can use, respectively, g_match_info_fetch() and
1625 * g_match_info_fetch_pos(). Note that the strings are returned in
1626 * reverse order of length; that is, the longest matching string is
1629 * Note that the DFA algorithm is slower than the standard one and it
1630 * is not able to capture substrings, so backreferences do not work.
1632 * Setting @start_position differs from just passing over a shortened
1633 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1634 * that begins with any kind of lookbehind assertion, such as "\b".
1636 * A #GMatchInfo structure, used to get information on the match, is
1637 * stored in @match_info if not %NULL. Note that if @match_info is
1638 * not %NULL then it is created even if the function returns %FALSE,
1639 * i.e. you must free it regardless if regular expression actually
1642 * @string is not copied and is used in #GMatchInfo internally. If
1643 * you use any #GMatchInfo method (except g_match_info_free()) after
1644 * freeing or modifying @string then the behaviour is undefined.
1646 * Returns: %TRUE is the string matched, %FALSE otherwise
1651 g_regex_match_all_full (const GRegex *regex,
1652 const gchar *string,
1654 gint start_position,
1655 GRegexMatchFlags match_options,
1656 GMatchInfo **match_info,
1662 g_return_val_if_fail (regex != NULL, FALSE);
1663 g_return_val_if_fail (string != NULL, FALSE);
1664 g_return_val_if_fail (start_position >= 0, FALSE);
1665 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1666 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1668 info = match_info_new (regex, string, string_len, start_position,
1669 match_options, TRUE);
1675 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1676 info->string, info->string_len,
1678 regex->match_opts | match_options,
1679 info->offsets, info->n_offsets,
1680 info->workspace, info->n_workspace);
1681 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1683 /* info->workspace is too small. */
1684 info->n_workspace *= 2;
1685 info->workspace = g_realloc (info->workspace,
1686 info->n_workspace * sizeof (gint));
1689 else if (info->matches == 0)
1691 /* info->offsets is too small. */
1692 info->n_offsets *= 2;
1693 info->offsets = g_realloc (info->offsets,
1694 info->n_offsets * sizeof (gint));
1697 else if (IS_PCRE_ERROR (info->matches))
1699 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1700 _("Error while matching regular expression %s: %s"),
1701 regex->pattern, match_error (info->matches));
1705 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1708 if (match_info != NULL)
1711 g_match_info_free (info);
1713 return info->matches >= 0;
1717 * g_regex_get_string_number:
1718 * @regex: #GRegex structure
1719 * @name: name of the subexpression
1721 * Retrieves the number of the subexpression named @name.
1723 * Returns: The number of the subexpression or -1 if @name
1729 g_regex_get_string_number (const GRegex *regex,
1734 g_return_val_if_fail (regex != NULL, -1);
1735 g_return_val_if_fail (name != NULL, -1);
1737 num = pcre_get_stringnumber (regex->pcre_re, name);
1738 if (num == PCRE_ERROR_NOSUBSTRING)
1745 * g_regex_split_simple:
1746 * @pattern: the regular expression
1747 * @string: the string to scan for matches
1748 * @compile_options: compile options for the regular expression, or 0
1749 * @match_options: match options, or 0
1751 * Breaks the string on the pattern, and returns an array of
1752 * the tokens. If the pattern contains capturing parentheses,
1753 * then the text for each of the substrings will also be returned.
1754 * If the pattern does not match anywhere in the string, then the
1755 * whole string is returned as the first token.
1757 * This function is equivalent to g_regex_split() but it does
1758 * not require to compile the pattern with g_regex_new(), avoiding
1759 * some lines of code when you need just to do a split without
1760 * extracting substrings, capture counts, and so on.
1762 * If this function is to be called on the same @pattern more than
1763 * once, it's more efficient to compile the pattern once with
1764 * g_regex_new() and then use g_regex_split().
1766 * As a special case, the result of splitting the empty string ""
1767 * is an empty vector, not a vector containing a single string.
1768 * The reason for this special case is that being able to represent
1769 * a empty vector is typically more useful than consistent handling
1770 * of empty elements. If you do need to represent empty elements,
1771 * you'll need to check for the empty string before calling this
1774 * A pattern that can match empty strings splits @string into
1775 * separate characters wherever it matches the empty string between
1776 * characters. For example splitting "ab c" using as a separator
1777 * "\s*", you will get "a", "b" and "c".
1779 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1784 g_regex_split_simple (const gchar *pattern,
1785 const gchar *string,
1786 GRegexCompileFlags compile_options,
1787 GRegexMatchFlags match_options)
1792 regex = g_regex_new (pattern, compile_options, 0, NULL);
1795 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1796 g_regex_unref (regex);
1802 * @regex: a #GRegex structure
1803 * @string: the string to split with the pattern
1804 * @match_options: match time option flags
1806 * Breaks the string on the pattern, and returns an array of the tokens.
1807 * If the pattern contains capturing parentheses, then the text for each
1808 * of the substrings will also be returned. If the pattern does not match
1809 * anywhere in the string, then the whole string is returned as the first
1812 * As a special case, the result of splitting the empty string "" is an
1813 * empty vector, not a vector containing a single string. The reason for
1814 * this special case is that being able to represent a empty vector is
1815 * typically more useful than consistent handling of empty elements. If
1816 * you do need to represent empty elements, you'll need to check for the
1817 * empty string before calling this function.
1819 * A pattern that can match empty strings splits @string into separate
1820 * characters wherever it matches the empty string between characters.
1821 * For example splitting "ab c" using as a separator "\s*", you will get
1824 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1829 g_regex_split (const GRegex *regex,
1830 const gchar *string,
1831 GRegexMatchFlags match_options)
1833 return g_regex_split_full (regex, string, -1, 0,
1834 match_options, 0, NULL);
1838 * g_regex_split_full:
1839 * @regex: a #GRegex structure
1840 * @string: (array length=string_len): the string to split with the pattern
1841 * @string_len: the length of @string, or -1 if @string is nul-terminated
1842 * @start_position: starting index of the string to match
1843 * @match_options: match time option flags
1844 * @max_tokens: the maximum number of tokens to split @string into.
1845 * If this is less than 1, the string is split completely
1846 * @error: return location for a #GError
1848 * Breaks the string on the pattern, and returns an array of the tokens.
1849 * If the pattern contains capturing parentheses, then the text for each
1850 * of the substrings will also be returned. If the pattern does not match
1851 * anywhere in the string, then the whole string is returned as the first
1854 * As a special case, the result of splitting the empty string "" is an
1855 * empty vector, not a vector containing a single string. The reason for
1856 * this special case is that being able to represent a empty vector is
1857 * typically more useful than consistent handling of empty elements. If
1858 * you do need to represent empty elements, you'll need to check for the
1859 * empty string before calling this function.
1861 * A pattern that can match empty strings splits @string into separate
1862 * characters wherever it matches the empty string between characters.
1863 * For example splitting "ab c" using as a separator "\s*", you will get
1866 * Setting @start_position differs from just passing over a shortened
1867 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1868 * that begins with any kind of lookbehind assertion, such as "\b".
1870 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1875 g_regex_split_full (const GRegex *regex,
1876 const gchar *string,
1878 gint start_position,
1879 GRegexMatchFlags match_options,
1883 GError *tmp_error = NULL;
1884 GMatchInfo *match_info;
1889 /* position of the last separator. */
1890 gint last_separator_end;
1891 /* was the last match 0 bytes long? */
1892 gboolean last_match_is_empty;
1893 /* the returned array of char **s */
1894 gchar **string_list;
1896 g_return_val_if_fail (regex != NULL, NULL);
1897 g_return_val_if_fail (string != NULL, NULL);
1898 g_return_val_if_fail (start_position >= 0, NULL);
1899 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1900 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1902 if (max_tokens <= 0)
1903 max_tokens = G_MAXINT;
1906 string_len = strlen (string);
1908 /* zero-length string */
1909 if (string_len - start_position == 0)
1910 return g_new0 (gchar *, 1);
1912 if (max_tokens == 1)
1914 string_list = g_new0 (gchar *, 2);
1915 string_list[0] = g_strndup (&string[start_position],
1916 string_len - start_position);
1922 last_separator_end = start_position;
1923 last_match_is_empty = FALSE;
1925 match_ok = g_regex_match_full (regex, string, string_len, start_position,
1926 match_options, &match_info, &tmp_error);
1927 while (tmp_error == NULL)
1931 last_match_is_empty =
1932 (match_info->offsets[0] == match_info->offsets[1]);
1934 /* we need to skip empty separators at the same position of the end
1935 * of another separator. e.g. the string is "a b" and the separator
1936 * is " *", so from 1 to 2 we have a match and at position 2 we have
1937 * an empty match. */
1938 if (last_separator_end != match_info->offsets[1])
1943 token = g_strndup (string + last_separator_end,
1944 match_info->offsets[0] - last_separator_end);
1945 list = g_list_prepend (list, token);
1948 /* if there were substrings, these need to be added to
1950 match_count = g_match_info_get_match_count (match_info);
1951 if (match_count > 1)
1953 for (i = 1; i < match_count; i++)
1954 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
1960 /* if there was no match, copy to end of string. */
1961 if (!last_match_is_empty)
1963 gchar *token = g_strndup (string + last_separator_end,
1964 match_info->string_len - last_separator_end);
1965 list = g_list_prepend (list, token);
1967 /* no more tokens, end the loop. */
1971 /* -1 to leave room for the last part. */
1972 if (token_count >= max_tokens - 1)
1974 /* we have reached the maximum number of tokens, so we copy
1975 * the remaining part of the string. */
1976 if (last_match_is_empty)
1978 /* the last match was empty, so we have moved one char
1979 * after the real position to avoid empty matches at the
1981 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
1983 /* the if is needed in the case we have terminated the available
1984 * tokens, but we are at the end of the string, so there are no
1985 * characters left to copy. */
1986 if (string_len > match_info->pos)
1988 gchar *token = g_strndup (string + match_info->pos,
1989 string_len - match_info->pos);
1990 list = g_list_prepend (list, token);
1996 last_separator_end = match_info->pos;
1997 if (last_match_is_empty)
1998 /* if the last match was empty, g_match_info_next() has moved
1999 * forward to avoid infinite loops, but we still need to copy that
2001 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
2003 match_ok = g_match_info_next (match_info, &tmp_error);
2005 g_match_info_free (match_info);
2006 if (tmp_error != NULL)
2008 g_propagate_error (error, tmp_error);
2009 g_list_foreach (list, (GFunc)g_free, NULL);
2011 match_info->pos = -1;
2015 string_list = g_new (gchar *, g_list_length (list) + 1);
2017 for (last = g_list_last (list); last; last = g_list_previous (last))
2018 string_list[i++] = last->data;
2019 string_list[i] = NULL;
2028 REPL_TYPE_CHARACTER,
2029 REPL_TYPE_SYMBOLIC_REFERENCE,
2030 REPL_TYPE_NUMERIC_REFERENCE,
2031 REPL_TYPE_CHANGE_CASE
2036 CHANGE_CASE_NONE = 1 << 0,
2037 CHANGE_CASE_UPPER = 1 << 1,
2038 CHANGE_CASE_LOWER = 1 << 2,
2039 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2040 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2041 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2042 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2043 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2046 struct _InterpolationData
2052 ChangeCase change_case;
2056 free_interpolation_data (InterpolationData *data)
2058 g_free (data->text);
2062 static const gchar *
2063 expand_escape (const gchar *replacement,
2065 InterpolationData *data,
2070 const gchar *error_detail;
2072 GError *tmp_error = NULL;
2080 data->type = REPL_TYPE_CHARACTER;
2085 data->type = REPL_TYPE_CHARACTER;
2090 data->type = REPL_TYPE_CHARACTER;
2095 data->type = REPL_TYPE_CHARACTER;
2100 data->type = REPL_TYPE_CHARACTER;
2105 data->type = REPL_TYPE_CHARACTER;
2110 data->type = REPL_TYPE_CHARACTER;
2115 data->type = REPL_TYPE_CHARACTER;
2125 h = g_ascii_xdigit_value (*p);
2128 error_detail = _("hexadecimal digit or '}' expected");
2139 for (i = 0; i < 2; i++)
2141 h = g_ascii_xdigit_value (*p);
2144 error_detail = _("hexadecimal digit expected");
2151 data->type = REPL_TYPE_STRING;
2152 data->text = g_new0 (gchar, 8);
2153 g_unichar_to_utf8 (x, data->text);
2157 data->type = REPL_TYPE_CHANGE_CASE;
2158 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2162 data->type = REPL_TYPE_CHANGE_CASE;
2163 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2167 data->type = REPL_TYPE_CHANGE_CASE;
2168 data->change_case = CHANGE_CASE_LOWER;
2172 data->type = REPL_TYPE_CHANGE_CASE;
2173 data->change_case = CHANGE_CASE_UPPER;
2177 data->type = REPL_TYPE_CHANGE_CASE;
2178 data->change_case = CHANGE_CASE_NONE;
2184 error_detail = _("missing '<' in symbolic reference");
2193 error_detail = _("unfinished symbolic reference");
2200 error_detail = _("zero-length symbolic reference");
2203 if (g_ascii_isdigit (*q))
2208 h = g_ascii_digit_value (*q);
2211 error_detail = _("digit expected");
2220 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2227 if (!g_ascii_isalnum (*r))
2229 error_detail = _("illegal symbolic reference");
2236 data->text = g_strndup (q, p - q);
2237 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2242 /* if \0 is followed by a number is an octal number representing a
2243 * character, else it is a numeric reference. */
2244 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2247 p = g_utf8_next_char (p);
2260 for (i = 0; i < 3; i++)
2262 h = g_ascii_digit_value (*p);
2272 if (i == 2 && base == 10)
2278 if (base == 8 || i == 3)
2280 data->type = REPL_TYPE_STRING;
2281 data->text = g_new0 (gchar, 8);
2282 g_unichar_to_utf8 (x, data->text);
2286 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2291 error_detail = _("stray final '\\'");
2295 error_detail = _("unknown escape sequence");
2302 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2303 tmp_error = g_error_new (G_REGEX_ERROR,
2304 G_REGEX_ERROR_REPLACE,
2305 _("Error while parsing replacement "
2306 "text \"%s\" at char %lu: %s"),
2308 (gulong)(p - replacement),
2310 g_propagate_error (error, tmp_error);
2316 split_replacement (const gchar *replacement,
2320 InterpolationData *data;
2321 const gchar *p, *start;
2323 start = p = replacement;
2328 data = g_new0 (InterpolationData, 1);
2329 start = p = expand_escape (replacement, p, data, error);
2332 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2334 free_interpolation_data (data);
2338 list = g_list_prepend (list, data);
2343 if (*p == '\\' || *p == '\0')
2347 data = g_new0 (InterpolationData, 1);
2348 data->text = g_strndup (start, p - start);
2349 data->type = REPL_TYPE_STRING;
2350 list = g_list_prepend (list, data);
2356 return g_list_reverse (list);
2359 /* Change the case of c based on change_case. */
2360 #define CHANGE_CASE(c, change_case) \
2361 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2362 g_unichar_tolower (c) : \
2363 g_unichar_toupper (c))
2366 string_append (GString *string,
2368 ChangeCase *change_case)
2372 if (text[0] == '\0')
2375 if (*change_case == CHANGE_CASE_NONE)
2377 g_string_append (string, text);
2379 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2381 c = g_utf8_get_char (text);
2382 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2383 g_string_append (string, g_utf8_next_char (text));
2384 *change_case = CHANGE_CASE_NONE;
2388 while (*text != '\0')
2390 c = g_utf8_get_char (text);
2391 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2392 text = g_utf8_next_char (text);
2398 interpolate_replacement (const GMatchInfo *match_info,
2403 InterpolationData *idata;
2405 ChangeCase change_case = CHANGE_CASE_NONE;
2407 for (list = data; list; list = list->next)
2410 switch (idata->type)
2412 case REPL_TYPE_STRING:
2413 string_append (result, idata->text, &change_case);
2415 case REPL_TYPE_CHARACTER:
2416 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2417 if (change_case & CHANGE_CASE_SINGLE_MASK)
2418 change_case = CHANGE_CASE_NONE;
2420 case REPL_TYPE_NUMERIC_REFERENCE:
2421 match = g_match_info_fetch (match_info, idata->num);
2424 string_append (result, match, &change_case);
2428 case REPL_TYPE_SYMBOLIC_REFERENCE:
2429 match = g_match_info_fetch_named (match_info, idata->text);
2432 string_append (result, match, &change_case);
2436 case REPL_TYPE_CHANGE_CASE:
2437 change_case = idata->change_case;
2445 /* whether actual match_info is needed for replacement, i.e.
2446 * whether there are references
2449 interpolation_list_needs_match (GList *list)
2451 while (list != NULL)
2453 InterpolationData *data = list->data;
2455 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2456 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2469 * @regex: a #GRegex structure
2470 * @string: (array length=string_len): the string to perform matches against
2471 * @string_len: the length of @string, or -1 if @string is nul-terminated
2472 * @start_position: starting index of the string to match
2473 * @replacement: text to replace each match with
2474 * @match_options: options for the match
2475 * @error: location to store the error occuring, or %NULL to ignore errors
2477 * Replaces all occurrences of the pattern in @regex with the
2478 * replacement text. Backreferences of the form '\number' or
2479 * '\g<number>' in the replacement text are interpolated by the
2480 * number-th captured subexpression of the match, '\g<name>' refers
2481 * to the captured subexpression with the given name. '\0' refers to the
2482 * complete match, but '\0' followed by a number is the octal representation
2483 * of a character. To include a literal '\' in the replacement, write '\\'.
2484 * There are also escapes that changes the case of the following text:
2487 * <varlistentry><term>\l</term>
2489 * <para>Convert to lower case the next character</para>
2492 * <varlistentry><term>\u</term>
2494 * <para>Convert to upper case the next character</para>
2497 * <varlistentry><term>\L</term>
2499 * <para>Convert to lower case till \E</para>
2502 * <varlistentry><term>\U</term>
2504 * <para>Convert to upper case till \E</para>
2507 * <varlistentry><term>\E</term>
2509 * <para>End case modification</para>
2514 * If you do not need to use backreferences use g_regex_replace_literal().
2516 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2517 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2518 * you can use g_regex_replace_literal().
2520 * Setting @start_position differs from just passing over a shortened
2521 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2522 * begins with any kind of lookbehind assertion, such as "\b".
2524 * Returns: a newly allocated string containing the replacements
2529 g_regex_replace (const GRegex *regex,
2530 const gchar *string,
2532 gint start_position,
2533 const gchar *replacement,
2534 GRegexMatchFlags match_options,
2539 GError *tmp_error = NULL;
2541 g_return_val_if_fail (regex != NULL, NULL);
2542 g_return_val_if_fail (string != NULL, NULL);
2543 g_return_val_if_fail (start_position >= 0, NULL);
2544 g_return_val_if_fail (replacement != NULL, NULL);
2545 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2546 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2548 list = split_replacement (replacement, &tmp_error);
2549 if (tmp_error != NULL)
2551 g_propagate_error (error, tmp_error);
2555 result = g_regex_replace_eval (regex,
2556 string, string_len, start_position,
2558 interpolate_replacement,
2561 if (tmp_error != NULL)
2562 g_propagate_error (error, tmp_error);
2564 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2571 literal_replacement (const GMatchInfo *match_info,
2575 g_string_append (result, data);
2580 * g_regex_replace_literal:
2581 * @regex: a #GRegex structure
2582 * @string: (array length=string_len): the string to perform matches against
2583 * @string_len: the length of @string, or -1 if @string is nul-terminated
2584 * @start_position: starting index of the string to match
2585 * @replacement: text to replace each match with
2586 * @match_options: options for the match
2587 * @error: location to store the error occuring, or %NULL to ignore errors
2589 * Replaces all occurrences of the pattern in @regex with the
2590 * replacement text. @replacement is replaced literally, to
2591 * include backreferences use g_regex_replace().
2593 * Setting @start_position differs from just passing over a
2594 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2595 * case of a pattern that begins with any kind of lookbehind
2596 * assertion, such as "\b".
2598 * Returns: a newly allocated string containing the replacements
2603 g_regex_replace_literal (const GRegex *regex,
2604 const gchar *string,
2606 gint start_position,
2607 const gchar *replacement,
2608 GRegexMatchFlags match_options,
2611 g_return_val_if_fail (replacement != NULL, NULL);
2612 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2614 return g_regex_replace_eval (regex,
2615 string, string_len, start_position,
2617 literal_replacement,
2618 (gpointer)replacement,
2623 * g_regex_replace_eval:
2624 * @regex: a #GRegex structure from g_regex_new()
2625 * @string: (array length=string_len): string to perform matches against
2626 * @string_len: the length of @string, or -1 if @string is nul-terminated
2627 * @start_position: starting index of the string to match
2628 * @match_options: options for the match
2629 * @eval: a function to call for each match
2630 * @user_data: user data to pass to the function
2631 * @error: location to store the error occuring, or %NULL to ignore errors
2633 * Replaces occurrences of the pattern in regex with the output of
2634 * @eval for that occurrence.
2636 * Setting @start_position differs from just passing over a shortened
2637 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2638 * that begins with any kind of lookbehind assertion, such as "\b".
2640 * The following example uses g_regex_replace_eval() to replace multiple
2644 * eval_cb (const GMatchInfo *info,
2651 * match = g_match_info_fetch (info, 0);
2652 * r = g_hash_table_lookup ((GHashTable *)data, match);
2653 * g_string_append (res, r);
2665 * h = g_hash_table_new (g_str_hash, g_str_equal);
2667 * g_hash_table_insert (h, "1", "ONE");
2668 * g_hash_table_insert (h, "2", "TWO");
2669 * g_hash_table_insert (h, "3", "THREE");
2670 * g_hash_table_insert (h, "4", "FOUR");
2672 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2673 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2674 * g_hash_table_destroy (h);
2679 * Returns: a newly allocated string containing the replacements
2684 g_regex_replace_eval (const GRegex *regex,
2685 const gchar *string,
2687 gint start_position,
2688 GRegexMatchFlags match_options,
2689 GRegexEvalCallback eval,
2693 GMatchInfo *match_info;
2696 gboolean done = FALSE;
2697 GError *tmp_error = NULL;
2699 g_return_val_if_fail (regex != NULL, NULL);
2700 g_return_val_if_fail (string != NULL, NULL);
2701 g_return_val_if_fail (start_position >= 0, NULL);
2702 g_return_val_if_fail (eval != NULL, NULL);
2703 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2706 string_len = strlen (string);
2708 result = g_string_sized_new (string_len);
2710 /* run down the string making matches. */
2711 g_regex_match_full (regex, string, string_len, start_position,
2712 match_options, &match_info, &tmp_error);
2713 while (!done && g_match_info_matches (match_info))
2715 g_string_append_len (result,
2717 match_info->offsets[0] - str_pos);
2718 done = (*eval) (match_info, result, user_data);
2719 str_pos = match_info->offsets[1];
2720 g_match_info_next (match_info, &tmp_error);
2722 g_match_info_free (match_info);
2723 if (tmp_error != NULL)
2725 g_propagate_error (error, tmp_error);
2726 g_string_free (result, TRUE);
2730 g_string_append_len (result, string + str_pos, string_len - str_pos);
2731 return g_string_free (result, FALSE);
2735 * g_regex_check_replacement:
2736 * @replacement: the replacement string
2737 * @has_references: (out) (allow-none): location to store information about
2738 * references in @replacement or %NULL
2739 * @error: location to store error
2741 * Checks whether @replacement is a valid replacement string
2742 * (see g_regex_replace()), i.e. that all escape sequences in
2745 * If @has_references is not %NULL then @replacement is checked
2746 * for pattern references. For instance, replacement text 'foo\n'
2747 * does not contain references and may be evaluated without information
2748 * about actual match, but '\0\1' (whole match followed by first
2749 * subpattern) requires valid #GMatchInfo object.
2751 * Returns: whether @replacement is a valid replacement string
2756 g_regex_check_replacement (const gchar *replacement,
2757 gboolean *has_references,
2763 list = split_replacement (replacement, &tmp);
2767 g_propagate_error (error, tmp);
2772 *has_references = interpolation_list_needs_match (list);
2774 g_list_foreach (list, (GFunc) free_interpolation_data, NULL);
2781 * g_regex_escape_string:
2782 * @string: (array length=length): the string to escape
2783 * @length: the length of @string, or -1 if @string is nul-terminated
2785 * Escapes the special characters used for regular expressions
2786 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2787 * function is useful to dynamically generate regular expressions.
2789 * @string can contain nul characters that are replaced with "\0",
2790 * in this case remember to specify the correct length of @string
2793 * Returns: a newly-allocated escaped string
2798 g_regex_escape_string (const gchar *string,
2802 const char *p, *piece_start, *end;
2804 g_return_val_if_fail (string != NULL, NULL);
2807 length = strlen (string);
2809 end = string + length;
2810 p = piece_start = string;
2811 escaped = g_string_sized_new (length + 1);
2832 if (p != piece_start)
2833 /* copy the previous piece. */
2834 g_string_append_len (escaped, piece_start, p - piece_start);
2835 g_string_append_c (escaped, '\\');
2837 g_string_append_c (escaped, '0');
2839 g_string_append_c (escaped, *p);
2843 p = g_utf8_next_char (p);
2848 if (piece_start < end)
2849 g_string_append_len (escaped, piece_start, end - piece_start);
2851 return g_string_free (escaped, FALSE);