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"
38 * @title: Perl-compatible regular expressions
39 * @short_description: matches strings against regular expressions
40 * @see_also: <xref linkend="glib-regex-syntax"/>
42 * The <function>g_regex_*()</function> functions implement regular
43 * expression pattern matching using syntax and semantics similar to
44 * Perl regular expression.
46 * Some functions accept a @start_position argument, setting it differs
47 * from just passing over a shortened string and setting #G_REGEX_MATCH_NOTBOL
48 * in the case of a pattern that begins with any kind of lookbehind assertion.
49 * For example, consider the pattern "\Biss\B" which finds occurrences of "iss"
50 * in the middle of words. ("\B" matches only if the current position in the
51 * subject is not a word boundary.) When applied to the string "Mississipi"
52 * from the fourth byte, namely "issipi", it does not match, because "\B" is
53 * always false at the start of the subject, which is deemed to be a word
54 * boundary. However, if the entire string is passed , but with
55 * @start_position set to 4, it finds the second occurrence of "iss" because
56 * it is able to look behind the starting point to discover that it is
57 * preceded by a letter.
59 * Note that, unless you set the #G_REGEX_RAW flag, all the strings passed
60 * to these functions must be encoded in UTF-8. The lengths and the positions
61 * inside the strings are in bytes and not in characters, so, for instance,
62 * "\xc3\xa0" (i.e. "à") is two bytes long but it is treated as a
63 * single character. If you set #G_REGEX_RAW the strings can be non-valid
64 * UTF-8 strings and a byte is treated as a character, so "\xc3\xa0" is two
65 * bytes and two characters long.
67 * When matching a pattern, "\n" matches only against a "\n" character in
68 * the string, and "\r" matches only a "\r" character. To match any newline
69 * sequence use "\R". This particular group matches either the two-character
70 * sequence CR + LF ("\r\n"), or one of the single characters LF (linefeed,
71 * U+000A, "\n"), VT vertical tab, U+000B, "\v"), FF (formfeed, U+000C, "\f"),
72 * CR (carriage return, U+000D, "\r"), NEL (next line, U+0085), LS (line
73 * separator, U+2028), or PS (paragraph separator, U+2029).
75 * The behaviour of the dot, circumflex, and dollar metacharacters are
76 * affected by newline characters, the default is to recognize any newline
77 * character (the same characters recognized by "\R"). This can be changed
78 * with #G_REGEX_NEWLINE_CR, #G_REGEX_NEWLINE_LF and #G_REGEX_NEWLINE_CRLF
79 * compile options, and with #G_REGEX_MATCH_NEWLINE_ANY,
80 * #G_REGEX_MATCH_NEWLINE_CR, #G_REGEX_MATCH_NEWLINE_LF and
81 * #G_REGEX_MATCH_NEWLINE_CRLF match options. These settings are also
82 * relevant when compiling a pattern if #G_REGEX_EXTENDED is set, and an
83 * unescaped "#" outside a character class is encountered. This indicates
84 * a comment that lasts until after the next newline.
86 * Creating and manipulating the same #GRegex structure from different
87 * threads is not a problem as #GRegex does not modify its internal
88 * state between creation and destruction, on the other hand #GMatchInfo
91 * The regular expressions low-level functionalities are obtained through
92 * the excellent <ulink url="http://www.pcre.org/">PCRE</ulink> library
93 * written by Philip Hazel.
96 /* Mask of all the possible values for GRegexCompileFlags. */
97 #define G_REGEX_COMPILE_MASK (G_REGEX_CASELESS | \
102 G_REGEX_DOLLAR_ENDONLY | \
105 G_REGEX_NO_AUTO_CAPTURE | \
108 G_REGEX_NEWLINE_CR | \
109 G_REGEX_NEWLINE_LF | \
110 G_REGEX_NEWLINE_CRLF)
112 /* Mask of all the possible values for GRegexMatchFlags. */
113 #define G_REGEX_MATCH_MASK (G_REGEX_MATCH_ANCHORED | \
114 G_REGEX_MATCH_NOTBOL | \
115 G_REGEX_MATCH_NOTEOL | \
116 G_REGEX_MATCH_NOTEMPTY | \
117 G_REGEX_MATCH_PARTIAL | \
118 G_REGEX_MATCH_NEWLINE_CR | \
119 G_REGEX_MATCH_NEWLINE_LF | \
120 G_REGEX_MATCH_NEWLINE_CRLF | \
121 G_REGEX_MATCH_NEWLINE_ANY)
123 /* if the string is in UTF-8 use g_utf8_ functions, else use
125 #define NEXT_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
126 g_utf8_next_char (s) : \
128 #define PREV_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
129 g_utf8_prev_char (s) : \
134 GRegex *regex; /* the regex */
135 GRegexMatchFlags match_opts; /* options used at match time on the regex */
136 gint matches; /* number of matching sub patterns */
137 gint pos; /* position in the string where last match left off */
138 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
139 gint n_offsets; /* number of offsets */
140 gint *workspace; /* workspace for pcre_dfa_exec() */
141 gint n_workspace; /* number of workspace elements */
142 const gchar *string; /* string passed to the match function */
143 gssize string_len; /* length of string */
148 volatile gint ref_count; /* the ref count for the immutable part */
149 gchar *pattern; /* the pattern */
150 pcre *pcre_re; /* compiled form of the pattern */
151 GRegexCompileFlags compile_opts; /* options used at compile time on the pattern */
152 GRegexMatchFlags match_opts; /* options used at match time on the regex */
153 pcre_extra *extra; /* data stored when G_REGEX_OPTIMIZE is used */
156 /* TRUE if ret is an error code, FALSE otherwise. */
157 #define IS_PCRE_ERROR(ret) ((ret) < PCRE_ERROR_NOMATCH && (ret) != PCRE_ERROR_PARTIAL)
159 typedef struct _InterpolationData InterpolationData;
160 static gboolean interpolation_list_needs_match (GList *list);
161 static gboolean interpolate_replacement (const GMatchInfo *match_info,
164 static GList *split_replacement (const gchar *replacement,
166 static void free_interpolation_data (InterpolationData *data);
170 match_error (gint errcode)
174 case PCRE_ERROR_NOMATCH:
177 case PCRE_ERROR_NULL:
178 /* NULL argument, this should not happen in GRegex */
179 g_warning ("A NULL argument was passed to PCRE");
181 case PCRE_ERROR_BADOPTION:
182 return "bad options";
183 case PCRE_ERROR_BADMAGIC:
184 return _("corrupted object");
185 case PCRE_ERROR_UNKNOWN_OPCODE:
186 return N_("internal error or corrupted object");
187 case PCRE_ERROR_NOMEMORY:
188 return _("out of memory");
189 case PCRE_ERROR_NOSUBSTRING:
190 /* not used by pcre_exec() */
192 case PCRE_ERROR_MATCHLIMIT:
193 return _("backtracking limit reached");
194 case PCRE_ERROR_CALLOUT:
195 /* callouts are not implemented */
197 case PCRE_ERROR_BADUTF8:
198 case PCRE_ERROR_BADUTF8_OFFSET:
199 /* we do not check if strings are valid */
201 case PCRE_ERROR_PARTIAL:
204 case PCRE_ERROR_BADPARTIAL:
205 return _("the pattern contains items not supported for partial matching");
206 case PCRE_ERROR_INTERNAL:
207 return _("internal error");
208 case PCRE_ERROR_BADCOUNT:
209 /* negative ovecsize, this should not happen in GRegex */
210 g_warning ("A negative ovecsize was passed to PCRE");
212 case PCRE_ERROR_DFA_UITEM:
213 return _("the pattern contains items not supported for partial matching");
214 case PCRE_ERROR_DFA_UCOND:
215 return _("back references as conditions are not supported for partial matching");
216 case PCRE_ERROR_DFA_UMLIMIT:
217 /* the match_field field is not used in GRegex */
219 case PCRE_ERROR_DFA_WSSIZE:
220 /* handled expanding the workspace */
222 case PCRE_ERROR_DFA_RECURSE:
223 case PCRE_ERROR_RECURSIONLIMIT:
224 return _("recursion limit reached");
225 case PCRE_ERROR_NULLWSLIMIT:
226 return _("workspace limit for empty substrings reached");
227 case PCRE_ERROR_BADNEWLINE:
228 return _("invalid combination of newline flags");
232 return _("unknown error");
236 translate_compile_error (gint *errcode, const gchar **errmsg)
238 /* Compile errors are created adding 100 to the error code returned
240 * If errcode is known we put the translatable error message in
241 * erromsg. If errcode is unknown we put the generic
242 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
243 * untranslated error message returned by PCRE.
244 * Note that there can be more PCRE errors with the same GRegexError
245 * and that some PCRE errors are useless for us.
251 case G_REGEX_ERROR_STRAY_BACKSLASH:
252 *errmsg = _("\\ at end of pattern");
254 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
255 *errmsg = _("\\c at end of pattern");
257 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
258 *errmsg = _("unrecognized character follows \\");
261 /* A number of Perl escapes are not handled by PCRE.
262 * Therefore it explicitly raises ERR37.
264 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
265 *errmsg = _("case-changing escapes (\\l, \\L, \\u, \\U) are not allowed here");
267 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
268 *errmsg = _("numbers out of order in {} quantifier");
270 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
271 *errmsg = _("number too big in {} quantifier");
273 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
274 *errmsg = _("missing terminating ] for character class");
276 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
277 *errmsg = _("invalid escape sequence in character class");
279 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
280 *errmsg = _("range out of order in character class");
282 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
283 *errmsg = _("nothing to repeat");
285 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
286 *errmsg = _("unrecognized character after (?");
289 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
290 *errmsg = _("unrecognized character after (?<");
293 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
294 *errmsg = _("unrecognized character after (?P");
296 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
297 *errmsg = _("POSIX named classes are supported only within a class");
299 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
300 *errmsg = _("missing terminating )");
303 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
304 *errmsg = _(") without opening (");
307 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
308 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
309 * sequences here, '(?-54' would be an example for the second group.
311 *errmsg = _("(?R or (?[+-]digits must be followed by )");
313 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
314 *errmsg = _("reference to non-existent subpattern");
316 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
317 *errmsg = _("missing ) after comment");
319 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
320 *errmsg = _("regular expression too large");
322 case G_REGEX_ERROR_MEMORY_ERROR:
323 *errmsg = _("failed to get memory");
325 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
326 *errmsg = _("lookbehind assertion is not fixed length");
328 case G_REGEX_ERROR_MALFORMED_CONDITION:
329 *errmsg = _("malformed number or name after (?(");
331 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
332 *errmsg = _("conditional group contains more than two branches");
334 case G_REGEX_ERROR_ASSERTION_EXPECTED:
335 *errmsg = _("assertion expected after (?(");
337 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
338 *errmsg = _("unknown POSIX class name");
340 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
341 *errmsg = _("POSIX collating elements are not supported");
343 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
344 *errmsg = _("character value in \\x{...} sequence is too large");
346 case G_REGEX_ERROR_INVALID_CONDITION:
347 *errmsg = _("invalid condition (?(0)");
349 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
350 *errmsg = _("\\C not allowed in lookbehind assertion");
352 case G_REGEX_ERROR_INFINITE_LOOP:
353 *errmsg = _("recursive call could loop indefinitely");
355 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
356 *errmsg = _("missing terminator in subpattern name");
358 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
359 *errmsg = _("two named subpatterns have the same name");
361 case G_REGEX_ERROR_MALFORMED_PROPERTY:
362 *errmsg = _("malformed \\P or \\p sequence");
364 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
365 *errmsg = _("unknown property name after \\P or \\p");
367 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
368 *errmsg = _("subpattern name is too long (maximum 32 characters)");
370 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
371 *errmsg = _("too many named subpatterns (maximum 10,000)");
373 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
374 *errmsg = _("octal value is greater than \\377");
376 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
377 *errmsg = _("DEFINE group contains more than one branch");
379 case G_REGEX_ERROR_DEFINE_REPETION:
380 *errmsg = _("repeating a DEFINE group is not allowed");
382 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
383 *errmsg = _("inconsistent NEWLINE options");
385 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
386 *errmsg = _("\\g is not followed by a braced name or an optionally "
387 "braced non-zero number");
390 *errcode = G_REGEX_ERROR_INTERNAL;
391 *errmsg = _("unexpected repeat");
394 *errcode = G_REGEX_ERROR_INTERNAL;
395 *errmsg = _("code overflow");
398 *errcode = G_REGEX_ERROR_INTERNAL;
399 *errmsg = _("overran compiling workspace");
402 *errcode = G_REGEX_ERROR_INTERNAL;
403 *errmsg = _("previously-checked referenced subpattern not found");
406 /* This should not happen as we never pass a NULL erroffset */
407 g_warning ("erroffset passed as NULL");
408 *errcode = G_REGEX_ERROR_COMPILE;
411 /* This should not happen as we check options before passing them
412 * to pcre_compile2() */
413 g_warning ("unknown option bit(s) set");
414 *errcode = G_REGEX_ERROR_COMPILE;
419 /* These errors should not happen as we are using an UTF8-enabled PCRE
420 * and we do not check if strings are valid */
421 g_warning ("%s", *errmsg);
422 *errcode = G_REGEX_ERROR_COMPILE;
425 *errcode = G_REGEX_ERROR_COMPILE;
432 match_info_new (const GRegex *regex,
439 GMatchInfo *match_info;
442 string_len = strlen (string);
444 match_info = g_new0 (GMatchInfo, 1);
445 match_info->regex = g_regex_ref ((GRegex *)regex);
446 match_info->string = string;
447 match_info->string_len = string_len;
448 match_info->matches = PCRE_ERROR_NOMATCH;
449 match_info->pos = start_position;
450 match_info->match_opts = match_options;
454 /* These values should be enough for most cases, if they are not
455 * enough g_regex_match_all_full() will expand them. */
456 match_info->n_offsets = 24;
457 match_info->n_workspace = 100;
458 match_info->workspace = g_new (gint, match_info->n_workspace);
463 pcre_fullinfo (regex->pcre_re, regex->extra,
464 PCRE_INFO_CAPTURECOUNT, &capture_count);
465 match_info->n_offsets = (capture_count + 1) * 3;
468 match_info->offsets = g_new0 (gint, match_info->n_offsets);
469 /* Set an invalid position for the previous match. */
470 match_info->offsets[0] = -1;
471 match_info->offsets[1] = -1;
477 * g_match_info_get_regex:
478 * @match_info: a #GMatchInfo
480 * Returns #GRegex object used in @match_info. It belongs to Glib
481 * and must not be freed. Use g_regex_ref() if you need to keep it
482 * after you free @match_info object.
484 * Returns: #GRegex object used in @match_info
489 g_match_info_get_regex (const GMatchInfo *match_info)
491 g_return_val_if_fail (match_info != NULL, NULL);
492 return match_info->regex;
496 * g_match_info_get_string:
497 * @match_info: a #GMatchInfo
499 * Returns the string searched with @match_info. This is the
500 * string passed to g_regex_match() or g_regex_replace() so
501 * you may not free it before calling this function.
503 * Returns: the string searched with @match_info
508 g_match_info_get_string (const GMatchInfo *match_info)
510 g_return_val_if_fail (match_info != NULL, NULL);
511 return match_info->string;
516 * @match_info: a #GMatchInfo
518 * Frees all the memory associated with the #GMatchInfo structure.
523 g_match_info_free (GMatchInfo *match_info)
527 g_regex_unref (match_info->regex);
528 g_free (match_info->offsets);
529 g_free (match_info->workspace);
536 * @match_info: a #GMatchInfo structure
537 * @error: location to store the error occuring, or %NULL to ignore errors
539 * Scans for the next match using the same parameters of the previous
540 * call to g_regex_match_full() or g_regex_match() that returned
543 * The match is done on the string passed to the match function, so you
544 * cannot free it before calling this function.
546 * Returns: %TRUE is the string matched, %FALSE otherwise
551 g_match_info_next (GMatchInfo *match_info,
554 gint prev_match_start;
557 g_return_val_if_fail (match_info != NULL, FALSE);
558 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
559 g_return_val_if_fail (match_info->pos >= 0, FALSE);
561 prev_match_start = match_info->offsets[0];
562 prev_match_end = match_info->offsets[1];
564 match_info->matches = pcre_exec (match_info->regex->pcre_re,
565 match_info->regex->extra,
567 match_info->string_len,
569 match_info->regex->match_opts | match_info->match_opts,
571 match_info->n_offsets);
572 if (IS_PCRE_ERROR (match_info->matches))
574 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
575 _("Error while matching regular expression %s: %s"),
576 match_info->regex->pattern, match_error (match_info->matches));
580 /* avoid infinite loops if the pattern is an empty string or something
582 if (match_info->pos == match_info->offsets[1])
584 if (match_info->pos > match_info->string_len)
586 /* we have reached the end of the string */
587 match_info->pos = -1;
588 match_info->matches = PCRE_ERROR_NOMATCH;
592 match_info->pos = NEXT_CHAR (match_info->regex,
593 &match_info->string[match_info->pos]) -
598 match_info->pos = match_info->offsets[1];
601 /* it's possible to get two identical matches when we are matching
602 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
603 * the string is "RegExTest" we have:
604 * - search at position 0: match from 0 to 0
605 * - search at position 1: match from 3 to 3
606 * - search at position 3: match from 3 to 3 (duplicate)
607 * - search at position 4: match from 5 to 5
608 * - search at position 5: match from 5 to 5 (duplicate)
609 * - search at position 6: no match -> stop
610 * so we have to ignore the duplicates.
611 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
612 if (match_info->matches >= 0 &&
613 prev_match_start == match_info->offsets[0] &&
614 prev_match_end == match_info->offsets[1])
616 /* ignore this match and search the next one */
617 return g_match_info_next (match_info, error);
620 return match_info->matches >= 0;
624 * g_match_info_matches:
625 * @match_info: a #GMatchInfo structure
627 * Returns whether the previous match operation succeeded.
629 * Returns: %TRUE if the previous match operation succeeded,
635 g_match_info_matches (const GMatchInfo *match_info)
637 g_return_val_if_fail (match_info != NULL, FALSE);
639 return match_info->matches >= 0;
643 * g_match_info_get_match_count:
644 * @match_info: a #GMatchInfo structure
646 * Retrieves the number of matched substrings (including substring 0,
647 * that is the whole matched text), so 1 is returned if the pattern
648 * has no substrings in it and 0 is returned if the match failed.
650 * If the last match was obtained using the DFA algorithm, that is
651 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
652 * count is not that of the number of capturing parentheses but that of
653 * the number of matched substrings.
655 * Returns: Number of matched substrings, or -1 if an error occurred
660 g_match_info_get_match_count (const GMatchInfo *match_info)
662 g_return_val_if_fail (match_info, -1);
664 if (match_info->matches == PCRE_ERROR_NOMATCH)
667 else if (match_info->matches < PCRE_ERROR_NOMATCH)
672 return match_info->matches;
676 * g_match_info_is_partial_match:
677 * @match_info: a #GMatchInfo structure
679 * Usually if the string passed to g_regex_match*() matches as far as
680 * it goes, but is too short to match the entire pattern, %FALSE is
681 * returned. There are circumstances where it might be helpful to
682 * distinguish this case from other cases in which there is no match.
684 * Consider, for example, an application where a human is required to
685 * type in data for a field with specific formatting requirements. An
686 * example might be a date in the form ddmmmyy, defined by the pattern
687 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
688 * If the application sees the user’s keystrokes one by one, and can
689 * check that what has been typed so far is potentially valid, it is
690 * able to raise an error as soon as a mistake is made.
692 * GRegex supports the concept of partial matching by means of the
693 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
694 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
695 * for a complete match, %FALSE otherwise. But, when these functions
696 * return %FALSE, you can check if the match was partial calling
697 * g_match_info_is_partial_match().
699 * When using partial matching you cannot use g_match_info_fetch*().
701 * Because of the way certain internal optimizations are implemented
702 * the partial matching algorithm cannot be used with all patterns.
703 * So repeated single characters such as "a{2,4}" and repeated single
704 * meta-sequences such as "\d+" are not permitted if the maximum number
705 * of occurrences is greater than one. Optional items such as "\d?"
706 * (where the maximum is one) are permitted. Quantifiers with any values
707 * are permitted after parentheses, so the invalid examples above can be
708 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
709 * for a pattern that does not conform to the restrictions, matching
710 * functions return an error.
712 * Returns: %TRUE if the match was partial, %FALSE otherwise
717 g_match_info_is_partial_match (const GMatchInfo *match_info)
719 g_return_val_if_fail (match_info != NULL, FALSE);
721 return match_info->matches == PCRE_ERROR_PARTIAL;
725 * g_match_info_expand_references:
726 * @match_info: a #GMatchInfo or %NULL
727 * @string_to_expand: the string to expand
728 * @error: location to store the error occuring, or %NULL to ignore errors
730 * Returns a new string containing the text in @string_to_expand with
731 * references and escape sequences expanded. References refer to the last
732 * match done with @string against @regex and have the same syntax used by
735 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
736 * passed to g_regex_new().
738 * The backreferences are extracted from the string passed to the match
739 * function, so you cannot call this function after freeing the string.
741 * @match_info may be %NULL in which case @string_to_expand must not
742 * contain references. For instance "foo\n" does not refer to an actual
743 * pattern and '\n' merely will be replaced with \n character,
744 * while to expand "\0" (whole match) one needs the result of a match.
745 * Use g_regex_check_replacement() to find out whether @string_to_expand
746 * contains references.
748 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
753 g_match_info_expand_references (const GMatchInfo *match_info,
754 const gchar *string_to_expand,
759 GError *tmp_error = NULL;
761 g_return_val_if_fail (string_to_expand != NULL, NULL);
762 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
764 list = split_replacement (string_to_expand, &tmp_error);
765 if (tmp_error != NULL)
767 g_propagate_error (error, tmp_error);
771 if (!match_info && interpolation_list_needs_match (list))
773 g_critical ("String '%s' contains references to the match, can't "
774 "expand references without GMatchInfo object",
779 result = g_string_sized_new (strlen (string_to_expand));
780 interpolate_replacement (match_info, result, list);
782 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
785 return g_string_free (result, FALSE);
789 * g_match_info_fetch:
790 * @match_info: #GMatchInfo structure
791 * @match_num: number of the sub expression
793 * Retrieves the text matching the @match_num<!-- -->'th capturing
794 * parentheses. 0 is the full text of the match, 1 is the first paren
795 * set, 2 the second, and so on.
797 * If @match_num is a valid sub pattern but it didn't match anything
798 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
799 * string is returned.
801 * If the match was obtained using the DFA algorithm, that is using
802 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
803 * string is not that of a set of parentheses but that of a matched
804 * substring. Substrings are matched in reverse order of length, so
805 * 0 is the longest match.
807 * The string is fetched from the string passed to the match function,
808 * so you cannot call this function after freeing the string.
810 * Returns: (allow-none): The matched substring, or %NULL if an error
811 * occurred. You have to free the string yourself
816 g_match_info_fetch (const GMatchInfo *match_info,
819 /* we cannot use pcre_get_substring() because it allocates the
820 * string using pcre_malloc(). */
824 g_return_val_if_fail (match_info != NULL, NULL);
825 g_return_val_if_fail (match_num >= 0, NULL);
827 /* match_num does not exist or it didn't matched, i.e. matching "b"
828 * against "(a)?b" then group 0 is empty. */
829 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
831 else if (start == -1)
832 match = g_strdup ("");
834 match = g_strndup (&match_info->string[start], end - start);
840 * g_match_info_fetch_pos:
841 * @match_info: #GMatchInfo structure
842 * @match_num: number of the sub expression
843 * @start_pos: (out) (allow-none): pointer to location where to store
844 * the start position, or %NULL
845 * @end_pos: (out) (allow-none): pointer to location where to store
846 * the end position, or %NULL
848 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
849 * parentheses. 0 is the full text of the match, 1 is the first
850 * paren set, 2 the second, and so on.
852 * If @match_num is a valid sub pattern but it didn't match anything
853 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
854 * and @end_pos are set to -1 and %TRUE is returned.
856 * If the match was obtained using the DFA algorithm, that is using
857 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
858 * position is not that of a set of parentheses but that of a matched
859 * substring. Substrings are matched in reverse order of length, so
860 * 0 is the longest match.
862 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
863 * the position cannot be fetched, @start_pos and @end_pos are left
869 g_match_info_fetch_pos (const GMatchInfo *match_info,
874 g_return_val_if_fail (match_info != NULL, FALSE);
875 g_return_val_if_fail (match_num >= 0, FALSE);
877 /* make sure the sub expression number they're requesting is less than
878 * the total number of sub expressions that were matched. */
879 if (match_num >= match_info->matches)
882 if (start_pos != NULL)
883 *start_pos = match_info->offsets[2 * match_num];
886 *end_pos = match_info->offsets[2 * match_num + 1];
892 * Returns number of first matched subpattern with name @name.
893 * There may be more than one in case when DUPNAMES is used,
894 * and not all subpatterns with that name match;
895 * pcre_get_stringnumber() does not work in that case.
898 get_matched_substring_number (const GMatchInfo *match_info,
905 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
906 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
908 /* This code is copied from pcre_get.c: get_first_set() */
909 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
917 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
919 gint n = (entry[0] << 8) + entry[1];
920 if (match_info->offsets[n*2] >= 0)
924 return (first[0] << 8) + first[1];
928 * g_match_info_fetch_named:
929 * @match_info: #GMatchInfo structure
930 * @name: name of the subexpression
932 * Retrieves the text matching the capturing parentheses named @name.
934 * If @name is a valid sub pattern name but it didn't match anything
935 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
936 * then an empty string is returned.
938 * The string is fetched from the string passed to the match function,
939 * so you cannot call this function after freeing the string.
941 * Returns: (allow-none): The matched substring, or %NULL if an error
942 * occurred. You have to free the string yourself
947 g_match_info_fetch_named (const GMatchInfo *match_info,
950 /* we cannot use pcre_get_named_substring() because it allocates the
951 * string using pcre_malloc(). */
954 g_return_val_if_fail (match_info != NULL, NULL);
955 g_return_val_if_fail (name != NULL, NULL);
957 num = get_matched_substring_number (match_info, name);
961 return g_match_info_fetch (match_info, num);
965 * g_match_info_fetch_named_pos:
966 * @match_info: #GMatchInfo structure
967 * @name: name of the subexpression
968 * @start_pos: (out) (allow-none): pointer to location where to store
969 * the start position, or %NULL
970 * @end_pos: (out) (allow-none): pointer to location where to store
971 * the end position, or %NULL
973 * Retrieves the position in bytes of the capturing parentheses named @name.
975 * If @name is a valid sub pattern name but it didn't match anything
976 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
977 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
979 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
980 * If the position cannot be fetched, @start_pos and @end_pos
981 * are left unchanged.
986 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
993 g_return_val_if_fail (match_info != NULL, FALSE);
994 g_return_val_if_fail (name != NULL, FALSE);
996 num = get_matched_substring_number (match_info, name);
1000 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1004 * g_match_info_fetch_all:
1005 * @match_info: a #GMatchInfo structure
1007 * Bundles up pointers to each of the matching substrings from a match
1008 * and stores them in an array of gchar pointers. The first element in
1009 * the returned array is the match number 0, i.e. the entire matched
1012 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1013 * "b" against "(a)?b") then an empty string is inserted.
1015 * If the last match was obtained using the DFA algorithm, that is using
1016 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1017 * strings are not that matched by sets of parentheses but that of the
1018 * matched substring. Substrings are matched in reverse order of length,
1019 * so the first one is the longest match.
1021 * The strings are fetched from the string passed to the match function,
1022 * so you cannot call this function after freeing the string.
1024 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1025 * It must be freed using g_strfreev(). If the previous match failed
1031 g_match_info_fetch_all (const GMatchInfo *match_info)
1033 /* we cannot use pcre_get_substring_list() because the returned value
1034 * isn't suitable for g_strfreev(). */
1038 g_return_val_if_fail (match_info != NULL, NULL);
1040 if (match_info->matches < 0)
1043 result = g_new (gchar *, match_info->matches + 1);
1044 for (i = 0; i < match_info->matches; i++)
1045 result[i] = g_match_info_fetch (match_info, i);
1055 g_regex_error_quark (void)
1057 static GQuark error_quark = 0;
1059 if (error_quark == 0)
1060 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1069 * Increases reference count of @regex by 1.
1076 g_regex_ref (GRegex *regex)
1078 g_return_val_if_fail (regex != NULL, NULL);
1079 g_atomic_int_inc (®ex->ref_count);
1087 * Decreases reference count of @regex by 1. When reference count drops
1088 * to zero, it frees all the memory associated with the regex structure.
1093 g_regex_unref (GRegex *regex)
1095 g_return_if_fail (regex != NULL);
1097 if (g_atomic_int_exchange_and_add (®ex->ref_count, -1) - 1 == 0)
1099 g_free (regex->pattern);
1100 if (regex->pcre_re != NULL)
1101 pcre_free (regex->pcre_re);
1102 if (regex->extra != NULL)
1103 pcre_free (regex->extra);
1110 * @pattern: the regular expression
1111 * @compile_options: compile options for the regular expression, or 0
1112 * @match_options: match options for the regular expression, or 0
1113 * @error: return location for a #GError
1115 * Compiles the regular expression to an internal form, and does
1116 * the initial setup of the #GRegex structure.
1118 * Returns: a #GRegex structure. Call g_regex_unref() when you
1124 g_regex_new (const gchar *pattern,
1125 GRegexCompileFlags compile_options,
1126 GRegexMatchFlags match_options,
1131 const gchar *errmsg;
1134 gboolean optimize = FALSE;
1135 static gboolean initialized = FALSE;
1136 unsigned long int pcre_compile_options;
1138 g_return_val_if_fail (pattern != NULL, NULL);
1139 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1140 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1141 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1148 pcre_config (PCRE_CONFIG_UTF8, &support);
1151 msg = N_("PCRE library is compiled without UTF8 support");
1152 g_critical ("%s", msg);
1153 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1157 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1160 msg = N_("PCRE library is compiled without UTF8 properties support");
1161 g_critical ("%s", msg);
1162 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1169 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1170 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1171 if (compile_options & G_REGEX_OPTIMIZE)
1174 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1175 * instead uses UTF-8 only if required with PCRE_UTF8. */
1176 if (compile_options & G_REGEX_RAW)
1179 compile_options &= ~G_REGEX_RAW;
1184 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1185 match_options |= PCRE_NO_UTF8_CHECK;
1188 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1189 * not for the system one. */
1190 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1191 !(compile_options & G_REGEX_NEWLINE_LF))
1193 compile_options |= PCRE_NEWLINE_ANY;
1196 /* compile the pattern */
1197 re = pcre_compile2 (pattern, compile_options, &errcode,
1198 &errmsg, &erroffset, NULL);
1200 /* if the compilation failed, set the error member and return
1206 /* Translate the PCRE error code to GRegexError and use a translated
1207 * error message if possible */
1208 translate_compile_error (&errcode, &errmsg);
1210 /* PCRE uses byte offsets but we want to show character offsets */
1211 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1213 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1214 _("Error while compiling regular "
1215 "expression %s at char %d: %s"),
1216 pattern, erroffset, errmsg);
1217 g_propagate_error (error, tmp_error);
1222 /* For options set at the beginning of the pattern, pcre puts them into
1223 * compile options, e.g. "(?i)foo" will make the pcre structure store
1224 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1225 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1226 compile_options = pcre_compile_options;
1228 if (!(compile_options & G_REGEX_DUPNAMES))
1230 gboolean jchanged = FALSE;
1231 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1233 compile_options |= G_REGEX_DUPNAMES;
1236 regex = g_new0 (GRegex, 1);
1237 regex->ref_count = 1;
1238 regex->pattern = g_strdup (pattern);
1239 regex->pcre_re = re;
1240 regex->compile_opts = compile_options;
1241 regex->match_opts = match_options;
1245 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1248 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1249 G_REGEX_ERROR_OPTIMIZE,
1250 _("Error while optimizing "
1251 "regular expression %s: %s"),
1254 g_propagate_error (error, tmp_error);
1256 g_regex_unref (regex);
1265 * g_regex_get_pattern:
1266 * @regex: a #GRegex structure
1268 * Gets the pattern string associated with @regex, i.e. a copy of
1269 * the string passed to g_regex_new().
1271 * Returns: the pattern of @regex
1276 g_regex_get_pattern (const GRegex *regex)
1278 g_return_val_if_fail (regex != NULL, NULL);
1280 return regex->pattern;
1284 * g_regex_get_max_backref:
1287 * Returns the number of the highest back reference
1288 * in the pattern, or 0 if the pattern does not contain
1291 * Returns: the number of the highest back reference
1296 g_regex_get_max_backref (const GRegex *regex)
1300 pcre_fullinfo (regex->pcre_re, regex->extra,
1301 PCRE_INFO_BACKREFMAX, &value);
1307 * g_regex_get_capture_count:
1310 * Returns the number of capturing subpatterns in the pattern.
1312 * Returns: the number of capturing subpatterns
1317 g_regex_get_capture_count (const GRegex *regex)
1321 pcre_fullinfo (regex->pcre_re, regex->extra,
1322 PCRE_INFO_CAPTURECOUNT, &value);
1328 * g_regex_get_compile_flags:
1331 * Returns the compile options that @regex was created with.
1333 * Returns: flags from #GRegexCompileFlags
1338 g_regex_get_compile_flags (const GRegex *regex)
1340 g_return_val_if_fail (regex != NULL, 0);
1342 return regex->compile_opts;
1346 * g_regex_get_match_flags:
1349 * Returns the match options that @regex was created with.
1351 * Returns: flags from #GRegexMatchFlags
1356 g_regex_get_match_flags (const GRegex *regex)
1358 g_return_val_if_fail (regex != NULL, 0);
1360 return regex->match_opts;
1364 * g_regex_match_simple:
1365 * @pattern: the regular expression
1366 * @string: the string to scan for matches
1367 * @compile_options: compile options for the regular expression, or 0
1368 * @match_options: match options, or 0
1370 * Scans for a match in @string for @pattern.
1372 * This function is equivalent to g_regex_match() but it does not
1373 * require to compile the pattern with g_regex_new(), avoiding some
1374 * lines of code when you need just to do a match without extracting
1375 * substrings, capture counts, and so on.
1377 * If this function is to be called on the same @pattern more than
1378 * once, it's more efficient to compile the pattern once with
1379 * g_regex_new() and then use g_regex_match().
1381 * Returns: %TRUE if the string matched, %FALSE otherwise
1386 g_regex_match_simple (const gchar *pattern,
1387 const gchar *string,
1388 GRegexCompileFlags compile_options,
1389 GRegexMatchFlags match_options)
1394 regex = g_regex_new (pattern, compile_options, 0, NULL);
1397 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1398 g_regex_unref (regex);
1404 * @regex: a #GRegex structure from g_regex_new()
1405 * @string: the string to scan for matches
1406 * @match_options: match options
1407 * @match_info: (out) (allow-none): pointer to location where to store
1408 * the #GMatchInfo, or %NULL if you do not need it
1410 * Scans for a match in string for the pattern in @regex.
1411 * The @match_options are combined with the match options specified
1412 * when the @regex structure was created, letting you have more
1413 * flexibility in reusing #GRegex structures.
1415 * A #GMatchInfo structure, used to get information on the match,
1416 * is stored in @match_info if not %NULL. Note that if @match_info
1417 * is not %NULL then it is created even if the function returns %FALSE,
1418 * i.e. you must free it regardless if regular expression actually matched.
1420 * To retrieve all the non-overlapping matches of the pattern in
1421 * string you can use g_match_info_next().
1425 * print_uppercase_words (const gchar *string)
1427 * /* Print all uppercase-only words. */
1429 * GMatchInfo *match_info;
1431 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1432 * g_regex_match (regex, string, 0, &match_info);
1433 * while (g_match_info_matches (match_info))
1435 * gchar *word = g_match_info_fetch (match_info, 0);
1436 * g_print ("Found: %s\n", word);
1438 * g_match_info_next (match_info, NULL);
1440 * g_match_info_free (match_info);
1441 * g_regex_unref (regex);
1445 * @string is not copied and is used in #GMatchInfo internally. If
1446 * you use any #GMatchInfo method (except g_match_info_free()) after
1447 * freeing or modifying @string then the behaviour is undefined.
1449 * Returns: %TRUE is the string matched, %FALSE otherwise
1454 g_regex_match (const GRegex *regex,
1455 const gchar *string,
1456 GRegexMatchFlags match_options,
1457 GMatchInfo **match_info)
1459 return g_regex_match_full (regex, string, -1, 0, match_options,
1464 * g_regex_match_full:
1465 * @regex: a #GRegex structure from g_regex_new()
1466 * @string: (array length=string_len): the string to scan for matches
1467 * @string_len: the length of @string, or -1 if @string is nul-terminated
1468 * @start_position: starting index of the string to match
1469 * @match_options: match options
1470 * @match_info: (out) (allow-none): pointer to location where to store
1471 * the #GMatchInfo, or %NULL if you do not need it
1472 * @error: location to store the error occuring, or %NULL to ignore errors
1474 * Scans for a match in string for the pattern in @regex.
1475 * The @match_options are combined with the match options specified
1476 * when the @regex structure was created, letting you have more
1477 * flexibility in reusing #GRegex structures.
1479 * Setting @start_position differs from just passing over a shortened
1480 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1481 * that begins with any kind of lookbehind assertion, such as "\b".
1483 * A #GMatchInfo structure, used to get information on the match, is
1484 * stored in @match_info if not %NULL. Note that if @match_info is
1485 * not %NULL then it is created even if the function returns %FALSE,
1486 * i.e. you must free it regardless if regular expression actually
1489 * @string is not copied and is used in #GMatchInfo internally. If
1490 * you use any #GMatchInfo method (except g_match_info_free()) after
1491 * freeing or modifying @string then the behaviour is undefined.
1493 * To retrieve all the non-overlapping matches of the pattern in
1494 * string you can use g_match_info_next().
1498 * print_uppercase_words (const gchar *string)
1500 * /* Print all uppercase-only words. */
1502 * GMatchInfo *match_info;
1503 * GError *error = NULL;
1505 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1506 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1507 * while (g_match_info_matches (match_info))
1509 * gchar *word = g_match_info_fetch (match_info, 0);
1510 * g_print ("Found: %s\n", word);
1512 * g_match_info_next (match_info, &error);
1514 * g_match_info_free (match_info);
1515 * g_regex_unref (regex);
1516 * if (error != NULL)
1518 * g_printerr ("Error while matching: %s\n", error->message);
1519 * g_error_free (error);
1524 * Returns: %TRUE is the string matched, %FALSE otherwise
1529 g_regex_match_full (const GRegex *regex,
1530 const gchar *string,
1532 gint start_position,
1533 GRegexMatchFlags match_options,
1534 GMatchInfo **match_info,
1540 g_return_val_if_fail (regex != NULL, FALSE);
1541 g_return_val_if_fail (string != NULL, FALSE);
1542 g_return_val_if_fail (start_position >= 0, FALSE);
1543 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1544 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1546 info = match_info_new (regex, string, string_len, start_position,
1547 match_options, FALSE);
1548 match_ok = g_match_info_next (info, error);
1549 if (match_info != NULL)
1552 g_match_info_free (info);
1558 * g_regex_match_all:
1559 * @regex: a #GRegex structure from g_regex_new()
1560 * @string: the string to scan for matches
1561 * @match_options: match options
1562 * @match_info: (out) (allow-none): pointer to location where to store
1563 * the #GMatchInfo, or %NULL if you do not need it
1565 * Using the standard algorithm for regular expression matching only
1566 * the longest match in the string is retrieved. This function uses
1567 * a different algorithm so it can retrieve all the possible matches.
1568 * For more documentation see g_regex_match_all_full().
1570 * A #GMatchInfo structure, used to get information on the match, is
1571 * stored in @match_info if not %NULL. Note that if @match_info is
1572 * not %NULL then it is created even if the function returns %FALSE,
1573 * i.e. you must free it regardless if regular expression actually
1576 * @string is not copied and is used in #GMatchInfo internally. If
1577 * you use any #GMatchInfo method (except g_match_info_free()) after
1578 * freeing or modifying @string then the behaviour is undefined.
1580 * Returns: %TRUE is the string matched, %FALSE otherwise
1585 g_regex_match_all (const GRegex *regex,
1586 const gchar *string,
1587 GRegexMatchFlags match_options,
1588 GMatchInfo **match_info)
1590 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1595 * g_regex_match_all_full:
1596 * @regex: a #GRegex structure from g_regex_new()
1597 * @string: (array length=string_len): the string to scan for matches
1598 * @string_len: the length of @string, or -1 if @string is nul-terminated
1599 * @start_position: starting index of the string to match
1600 * @match_options: match options
1601 * @match_info: (out) (allow-none): pointer to location where to store
1602 * the #GMatchInfo, or %NULL if you do not need it
1603 * @error: location to store the error occuring, or %NULL to ignore errors
1605 * Using the standard algorithm for regular expression matching only
1606 * the longest match in the string is retrieved, it is not possibile
1607 * to obtain all the available matches. For instance matching
1608 * "<a> <b> <c>" against the pattern "<.*>"
1609 * you get "<a> <b> <c>".
1611 * This function uses a different algorithm (called DFA, i.e. deterministic
1612 * finite automaton), so it can retrieve all the possible matches, all
1613 * starting at the same point in the string. For instance matching
1614 * "<a> <b> <c>" against the pattern "<.*>"
1615 * you would obtain three matches: "<a> <b> <c>",
1616 * "<a> <b>" and "<a>".
1618 * The number of matched strings is retrieved using
1619 * g_match_info_get_match_count(). To obtain the matched strings and
1620 * their position you can use, respectively, g_match_info_fetch() and
1621 * g_match_info_fetch_pos(). Note that the strings are returned in
1622 * reverse order of length; that is, the longest matching string is
1625 * Note that the DFA algorithm is slower than the standard one and it
1626 * is not able to capture substrings, so backreferences do not work.
1628 * Setting @start_position differs from just passing over a shortened
1629 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1630 * that begins with any kind of lookbehind assertion, such as "\b".
1632 * A #GMatchInfo structure, used to get information on the match, is
1633 * stored in @match_info if not %NULL. Note that if @match_info is
1634 * not %NULL then it is created even if the function returns %FALSE,
1635 * i.e. you must free it regardless if regular expression actually
1638 * @string is not copied and is used in #GMatchInfo internally. If
1639 * you use any #GMatchInfo method (except g_match_info_free()) after
1640 * freeing or modifying @string then the behaviour is undefined.
1642 * Returns: %TRUE is the string matched, %FALSE otherwise
1647 g_regex_match_all_full (const GRegex *regex,
1648 const gchar *string,
1650 gint start_position,
1651 GRegexMatchFlags match_options,
1652 GMatchInfo **match_info,
1658 g_return_val_if_fail (regex != NULL, FALSE);
1659 g_return_val_if_fail (string != NULL, FALSE);
1660 g_return_val_if_fail (start_position >= 0, FALSE);
1661 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1662 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1664 info = match_info_new (regex, string, string_len, start_position,
1665 match_options, TRUE);
1671 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1672 info->string, info->string_len,
1674 regex->match_opts | match_options,
1675 info->offsets, info->n_offsets,
1676 info->workspace, info->n_workspace);
1677 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1679 /* info->workspace is too small. */
1680 info->n_workspace *= 2;
1681 info->workspace = g_realloc (info->workspace,
1682 info->n_workspace * sizeof (gint));
1685 else if (info->matches == 0)
1687 /* info->offsets is too small. */
1688 info->n_offsets *= 2;
1689 info->offsets = g_realloc (info->offsets,
1690 info->n_offsets * sizeof (gint));
1693 else if (IS_PCRE_ERROR (info->matches))
1695 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1696 _("Error while matching regular expression %s: %s"),
1697 regex->pattern, match_error (info->matches));
1701 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1704 if (match_info != NULL)
1707 g_match_info_free (info);
1709 return info->matches >= 0;
1713 * g_regex_get_string_number:
1714 * @regex: #GRegex structure
1715 * @name: name of the subexpression
1717 * Retrieves the number of the subexpression named @name.
1719 * Returns: The number of the subexpression or -1 if @name
1725 g_regex_get_string_number (const GRegex *regex,
1730 g_return_val_if_fail (regex != NULL, -1);
1731 g_return_val_if_fail (name != NULL, -1);
1733 num = pcre_get_stringnumber (regex->pcre_re, name);
1734 if (num == PCRE_ERROR_NOSUBSTRING)
1741 * g_regex_split_simple:
1742 * @pattern: the regular expression
1743 * @string: the string to scan for matches
1744 * @compile_options: compile options for the regular expression, or 0
1745 * @match_options: match options, or 0
1747 * Breaks the string on the pattern, and returns an array of
1748 * the tokens. If the pattern contains capturing parentheses,
1749 * then the text for each of the substrings will also be returned.
1750 * If the pattern does not match anywhere in the string, then the
1751 * whole string is returned as the first token.
1753 * This function is equivalent to g_regex_split() but it does
1754 * not require to compile the pattern with g_regex_new(), avoiding
1755 * some lines of code when you need just to do a split without
1756 * extracting substrings, capture counts, and so on.
1758 * If this function is to be called on the same @pattern more than
1759 * once, it's more efficient to compile the pattern once with
1760 * g_regex_new() and then use g_regex_split().
1762 * As a special case, the result of splitting the empty string ""
1763 * is an empty vector, not a vector containing a single string.
1764 * The reason for this special case is that being able to represent
1765 * a empty vector is typically more useful than consistent handling
1766 * of empty elements. If you do need to represent empty elements,
1767 * you'll need to check for the empty string before calling this
1770 * A pattern that can match empty strings splits @string into
1771 * separate characters wherever it matches the empty string between
1772 * characters. For example splitting "ab c" using as a separator
1773 * "\s*", you will get "a", "b" and "c".
1775 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1780 g_regex_split_simple (const gchar *pattern,
1781 const gchar *string,
1782 GRegexCompileFlags compile_options,
1783 GRegexMatchFlags match_options)
1788 regex = g_regex_new (pattern, compile_options, 0, NULL);
1791 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1792 g_regex_unref (regex);
1798 * @regex: a #GRegex structure
1799 * @string: the string to split with the pattern
1800 * @match_options: match time option flags
1802 * Breaks the string on the pattern, and returns an array of the tokens.
1803 * If the pattern contains capturing parentheses, then the text for each
1804 * of the substrings will also be returned. If the pattern does not match
1805 * anywhere in the string, then the whole string is returned as the first
1808 * As a special case, the result of splitting the empty string "" is an
1809 * empty vector, not a vector containing a single string. The reason for
1810 * this special case is that being able to represent a empty vector is
1811 * typically more useful than consistent handling of empty elements. If
1812 * you do need to represent empty elements, you'll need to check for the
1813 * empty string before calling this function.
1815 * A pattern that can match empty strings splits @string into separate
1816 * characters wherever it matches the empty string between characters.
1817 * For example splitting "ab c" using as a separator "\s*", you will get
1820 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1825 g_regex_split (const GRegex *regex,
1826 const gchar *string,
1827 GRegexMatchFlags match_options)
1829 return g_regex_split_full (regex, string, -1, 0,
1830 match_options, 0, NULL);
1834 * g_regex_split_full:
1835 * @regex: a #GRegex structure
1836 * @string: (array length=string_len): the string to split with the pattern
1837 * @string_len: the length of @string, or -1 if @string is nul-terminated
1838 * @start_position: starting index of the string to match
1839 * @match_options: match time option flags
1840 * @max_tokens: the maximum number of tokens to split @string into.
1841 * If this is less than 1, the string is split completely
1842 * @error: return location for a #GError
1844 * Breaks the string on the pattern, and returns an array of the tokens.
1845 * If the pattern contains capturing parentheses, then the text for each
1846 * of the substrings will also be returned. If the pattern does not match
1847 * anywhere in the string, then the whole string is returned as the first
1850 * As a special case, the result of splitting the empty string "" is an
1851 * empty vector, not a vector containing a single string. The reason for
1852 * this special case is that being able to represent a empty vector is
1853 * typically more useful than consistent handling of empty elements. If
1854 * you do need to represent empty elements, you'll need to check for the
1855 * empty string before calling this function.
1857 * A pattern that can match empty strings splits @string into separate
1858 * characters wherever it matches the empty string between characters.
1859 * For example splitting "ab c" using as a separator "\s*", you will get
1862 * Setting @start_position differs from just passing over a shortened
1863 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1864 * that begins with any kind of lookbehind assertion, such as "\b".
1866 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1871 g_regex_split_full (const GRegex *regex,
1872 const gchar *string,
1874 gint start_position,
1875 GRegexMatchFlags match_options,
1879 GError *tmp_error = NULL;
1880 GMatchInfo *match_info;
1885 /* position of the last separator. */
1886 gint last_separator_end;
1887 /* was the last match 0 bytes long? */
1888 gboolean last_match_is_empty;
1889 /* the returned array of char **s */
1890 gchar **string_list;
1892 g_return_val_if_fail (regex != NULL, NULL);
1893 g_return_val_if_fail (string != NULL, NULL);
1894 g_return_val_if_fail (start_position >= 0, NULL);
1895 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1896 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1898 if (max_tokens <= 0)
1899 max_tokens = G_MAXINT;
1902 string_len = strlen (string);
1904 /* zero-length string */
1905 if (string_len - start_position == 0)
1906 return g_new0 (gchar *, 1);
1908 if (max_tokens == 1)
1910 string_list = g_new0 (gchar *, 2);
1911 string_list[0] = g_strndup (&string[start_position],
1912 string_len - start_position);
1918 last_separator_end = start_position;
1919 last_match_is_empty = FALSE;
1921 match_ok = g_regex_match_full (regex, string, string_len, start_position,
1922 match_options, &match_info, &tmp_error);
1923 while (tmp_error == NULL)
1927 last_match_is_empty =
1928 (match_info->offsets[0] == match_info->offsets[1]);
1930 /* we need to skip empty separators at the same position of the end
1931 * of another separator. e.g. the string is "a b" and the separator
1932 * is " *", so from 1 to 2 we have a match and at position 2 we have
1933 * an empty match. */
1934 if (last_separator_end != match_info->offsets[1])
1939 token = g_strndup (string + last_separator_end,
1940 match_info->offsets[0] - last_separator_end);
1941 list = g_list_prepend (list, token);
1944 /* if there were substrings, these need to be added to
1946 match_count = g_match_info_get_match_count (match_info);
1947 if (match_count > 1)
1949 for (i = 1; i < match_count; i++)
1950 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
1956 /* if there was no match, copy to end of string. */
1957 if (!last_match_is_empty)
1959 gchar *token = g_strndup (string + last_separator_end,
1960 match_info->string_len - last_separator_end);
1961 list = g_list_prepend (list, token);
1963 /* no more tokens, end the loop. */
1967 /* -1 to leave room for the last part. */
1968 if (token_count >= max_tokens - 1)
1970 /* we have reached the maximum number of tokens, so we copy
1971 * the remaining part of the string. */
1972 if (last_match_is_empty)
1974 /* the last match was empty, so we have moved one char
1975 * after the real position to avoid empty matches at the
1977 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
1979 /* the if is needed in the case we have terminated the available
1980 * tokens, but we are at the end of the string, so there are no
1981 * characters left to copy. */
1982 if (string_len > match_info->pos)
1984 gchar *token = g_strndup (string + match_info->pos,
1985 string_len - match_info->pos);
1986 list = g_list_prepend (list, token);
1992 last_separator_end = match_info->pos;
1993 if (last_match_is_empty)
1994 /* if the last match was empty, g_match_info_next() has moved
1995 * forward to avoid infinite loops, but we still need to copy that
1997 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
1999 match_ok = g_match_info_next (match_info, &tmp_error);
2001 g_match_info_free (match_info);
2002 if (tmp_error != NULL)
2004 g_propagate_error (error, tmp_error);
2005 g_list_foreach (list, (GFunc)g_free, NULL);
2007 match_info->pos = -1;
2011 string_list = g_new (gchar *, g_list_length (list) + 1);
2013 for (last = g_list_last (list); last; last = g_list_previous (last))
2014 string_list[i++] = last->data;
2015 string_list[i] = NULL;
2024 REPL_TYPE_CHARACTER,
2025 REPL_TYPE_SYMBOLIC_REFERENCE,
2026 REPL_TYPE_NUMERIC_REFERENCE,
2027 REPL_TYPE_CHANGE_CASE
2032 CHANGE_CASE_NONE = 1 << 0,
2033 CHANGE_CASE_UPPER = 1 << 1,
2034 CHANGE_CASE_LOWER = 1 << 2,
2035 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2036 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2037 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2038 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2039 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2042 struct _InterpolationData
2048 ChangeCase change_case;
2052 free_interpolation_data (InterpolationData *data)
2054 g_free (data->text);
2058 static const gchar *
2059 expand_escape (const gchar *replacement,
2061 InterpolationData *data,
2066 const gchar *error_detail;
2068 GError *tmp_error = NULL;
2076 data->type = REPL_TYPE_CHARACTER;
2081 data->type = REPL_TYPE_CHARACTER;
2086 data->type = REPL_TYPE_CHARACTER;
2091 data->type = REPL_TYPE_CHARACTER;
2096 data->type = REPL_TYPE_CHARACTER;
2101 data->type = REPL_TYPE_CHARACTER;
2106 data->type = REPL_TYPE_CHARACTER;
2111 data->type = REPL_TYPE_CHARACTER;
2121 h = g_ascii_xdigit_value (*p);
2124 error_detail = _("hexadecimal digit or '}' expected");
2135 for (i = 0; i < 2; i++)
2137 h = g_ascii_xdigit_value (*p);
2140 error_detail = _("hexadecimal digit expected");
2147 data->type = REPL_TYPE_STRING;
2148 data->text = g_new0 (gchar, 8);
2149 g_unichar_to_utf8 (x, data->text);
2153 data->type = REPL_TYPE_CHANGE_CASE;
2154 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2158 data->type = REPL_TYPE_CHANGE_CASE;
2159 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2163 data->type = REPL_TYPE_CHANGE_CASE;
2164 data->change_case = CHANGE_CASE_LOWER;
2168 data->type = REPL_TYPE_CHANGE_CASE;
2169 data->change_case = CHANGE_CASE_UPPER;
2173 data->type = REPL_TYPE_CHANGE_CASE;
2174 data->change_case = CHANGE_CASE_NONE;
2180 error_detail = _("missing '<' in symbolic reference");
2189 error_detail = _("unfinished symbolic reference");
2196 error_detail = _("zero-length symbolic reference");
2199 if (g_ascii_isdigit (*q))
2204 h = g_ascii_digit_value (*q);
2207 error_detail = _("digit expected");
2216 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2223 if (!g_ascii_isalnum (*r))
2225 error_detail = _("illegal symbolic reference");
2232 data->text = g_strndup (q, p - q);
2233 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2238 /* if \0 is followed by a number is an octal number representing a
2239 * character, else it is a numeric reference. */
2240 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2243 p = g_utf8_next_char (p);
2256 for (i = 0; i < 3; i++)
2258 h = g_ascii_digit_value (*p);
2268 if (i == 2 && base == 10)
2274 if (base == 8 || i == 3)
2276 data->type = REPL_TYPE_STRING;
2277 data->text = g_new0 (gchar, 8);
2278 g_unichar_to_utf8 (x, data->text);
2282 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2287 error_detail = _("stray final '\\'");
2291 error_detail = _("unknown escape sequence");
2298 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2299 tmp_error = g_error_new (G_REGEX_ERROR,
2300 G_REGEX_ERROR_REPLACE,
2301 _("Error while parsing replacement "
2302 "text \"%s\" at char %lu: %s"),
2304 (gulong)(p - replacement),
2306 g_propagate_error (error, tmp_error);
2312 split_replacement (const gchar *replacement,
2316 InterpolationData *data;
2317 const gchar *p, *start;
2319 start = p = replacement;
2324 data = g_new0 (InterpolationData, 1);
2325 start = p = expand_escape (replacement, p, data, error);
2328 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2330 free_interpolation_data (data);
2334 list = g_list_prepend (list, data);
2339 if (*p == '\\' || *p == '\0')
2343 data = g_new0 (InterpolationData, 1);
2344 data->text = g_strndup (start, p - start);
2345 data->type = REPL_TYPE_STRING;
2346 list = g_list_prepend (list, data);
2352 return g_list_reverse (list);
2355 /* Change the case of c based on change_case. */
2356 #define CHANGE_CASE(c, change_case) \
2357 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2358 g_unichar_tolower (c) : \
2359 g_unichar_toupper (c))
2362 string_append (GString *string,
2364 ChangeCase *change_case)
2368 if (text[0] == '\0')
2371 if (*change_case == CHANGE_CASE_NONE)
2373 g_string_append (string, text);
2375 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2377 c = g_utf8_get_char (text);
2378 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2379 g_string_append (string, g_utf8_next_char (text));
2380 *change_case = CHANGE_CASE_NONE;
2384 while (*text != '\0')
2386 c = g_utf8_get_char (text);
2387 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2388 text = g_utf8_next_char (text);
2394 interpolate_replacement (const GMatchInfo *match_info,
2399 InterpolationData *idata;
2401 ChangeCase change_case = CHANGE_CASE_NONE;
2403 for (list = data; list; list = list->next)
2406 switch (idata->type)
2408 case REPL_TYPE_STRING:
2409 string_append (result, idata->text, &change_case);
2411 case REPL_TYPE_CHARACTER:
2412 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2413 if (change_case & CHANGE_CASE_SINGLE_MASK)
2414 change_case = CHANGE_CASE_NONE;
2416 case REPL_TYPE_NUMERIC_REFERENCE:
2417 match = g_match_info_fetch (match_info, idata->num);
2420 string_append (result, match, &change_case);
2424 case REPL_TYPE_SYMBOLIC_REFERENCE:
2425 match = g_match_info_fetch_named (match_info, idata->text);
2428 string_append (result, match, &change_case);
2432 case REPL_TYPE_CHANGE_CASE:
2433 change_case = idata->change_case;
2441 /* whether actual match_info is needed for replacement, i.e.
2442 * whether there are references
2445 interpolation_list_needs_match (GList *list)
2447 while (list != NULL)
2449 InterpolationData *data = list->data;
2451 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2452 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2465 * @regex: a #GRegex structure
2466 * @string: (array length=string_len): the string to perform matches against
2467 * @string_len: the length of @string, or -1 if @string is nul-terminated
2468 * @start_position: starting index of the string to match
2469 * @replacement: text to replace each match with
2470 * @match_options: options for the match
2471 * @error: location to store the error occuring, or %NULL to ignore errors
2473 * Replaces all occurrences of the pattern in @regex with the
2474 * replacement text. Backreferences of the form '\number' or
2475 * '\g<number>' in the replacement text are interpolated by the
2476 * number-th captured subexpression of the match, '\g<name>' refers
2477 * to the captured subexpression with the given name. '\0' refers to the
2478 * complete match, but '\0' followed by a number is the octal representation
2479 * of a character. To include a literal '\' in the replacement, write '\\'.
2480 * There are also escapes that changes the case of the following text:
2483 * <varlistentry><term>\l</term>
2485 * <para>Convert to lower case the next character</para>
2488 * <varlistentry><term>\u</term>
2490 * <para>Convert to upper case the next character</para>
2493 * <varlistentry><term>\L</term>
2495 * <para>Convert to lower case till \E</para>
2498 * <varlistentry><term>\U</term>
2500 * <para>Convert to upper case till \E</para>
2503 * <varlistentry><term>\E</term>
2505 * <para>End case modification</para>
2510 * If you do not need to use backreferences use g_regex_replace_literal().
2512 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2513 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2514 * you can use g_regex_replace_literal().
2516 * Setting @start_position differs from just passing over a shortened
2517 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2518 * begins with any kind of lookbehind assertion, such as "\b".
2520 * Returns: a newly allocated string containing the replacements
2525 g_regex_replace (const GRegex *regex,
2526 const gchar *string,
2528 gint start_position,
2529 const gchar *replacement,
2530 GRegexMatchFlags match_options,
2535 GError *tmp_error = NULL;
2537 g_return_val_if_fail (regex != NULL, NULL);
2538 g_return_val_if_fail (string != NULL, NULL);
2539 g_return_val_if_fail (start_position >= 0, NULL);
2540 g_return_val_if_fail (replacement != NULL, NULL);
2541 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2542 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2544 list = split_replacement (replacement, &tmp_error);
2545 if (tmp_error != NULL)
2547 g_propagate_error (error, tmp_error);
2551 result = g_regex_replace_eval (regex,
2552 string, string_len, start_position,
2554 interpolate_replacement,
2557 if (tmp_error != NULL)
2558 g_propagate_error (error, tmp_error);
2560 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2567 literal_replacement (const GMatchInfo *match_info,
2571 g_string_append (result, data);
2576 * g_regex_replace_literal:
2577 * @regex: a #GRegex structure
2578 * @string: (array length=string_len): the string to perform matches against
2579 * @string_len: the length of @string, or -1 if @string is nul-terminated
2580 * @start_position: starting index of the string to match
2581 * @replacement: text to replace each match with
2582 * @match_options: options for the match
2583 * @error: location to store the error occuring, or %NULL to ignore errors
2585 * Replaces all occurrences of the pattern in @regex with the
2586 * replacement text. @replacement is replaced literally, to
2587 * include backreferences use g_regex_replace().
2589 * Setting @start_position differs from just passing over a
2590 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2591 * case of a pattern that begins with any kind of lookbehind
2592 * assertion, such as "\b".
2594 * Returns: a newly allocated string containing the replacements
2599 g_regex_replace_literal (const GRegex *regex,
2600 const gchar *string,
2602 gint start_position,
2603 const gchar *replacement,
2604 GRegexMatchFlags match_options,
2607 g_return_val_if_fail (replacement != NULL, NULL);
2608 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2610 return g_regex_replace_eval (regex,
2611 string, string_len, start_position,
2613 literal_replacement,
2614 (gpointer)replacement,
2619 * g_regex_replace_eval:
2620 * @regex: a #GRegex structure from g_regex_new()
2621 * @string: (array length=string_len): string to perform matches against
2622 * @string_len: the length of @string, or -1 if @string is nul-terminated
2623 * @start_position: starting index of the string to match
2624 * @match_options: options for the match
2625 * @eval: a function to call for each match
2626 * @user_data: user data to pass to the function
2627 * @error: location to store the error occuring, or %NULL to ignore errors
2629 * Replaces occurrences of the pattern in regex with the output of
2630 * @eval for that occurrence.
2632 * Setting @start_position differs from just passing over a shortened
2633 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2634 * that begins with any kind of lookbehind assertion, such as "\b".
2636 * The following example uses g_regex_replace_eval() to replace multiple
2640 * eval_cb (const GMatchInfo *info,
2647 * match = g_match_info_fetch (info, 0);
2648 * r = g_hash_table_lookup ((GHashTable *)data, match);
2649 * g_string_append (res, r);
2661 * h = g_hash_table_new (g_str_hash, g_str_equal);
2663 * g_hash_table_insert (h, "1", "ONE");
2664 * g_hash_table_insert (h, "2", "TWO");
2665 * g_hash_table_insert (h, "3", "THREE");
2666 * g_hash_table_insert (h, "4", "FOUR");
2668 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2669 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2670 * g_hash_table_destroy (h);
2675 * Returns: a newly allocated string containing the replacements
2680 g_regex_replace_eval (const GRegex *regex,
2681 const gchar *string,
2683 gint start_position,
2684 GRegexMatchFlags match_options,
2685 GRegexEvalCallback eval,
2689 GMatchInfo *match_info;
2692 gboolean done = FALSE;
2693 GError *tmp_error = NULL;
2695 g_return_val_if_fail (regex != NULL, NULL);
2696 g_return_val_if_fail (string != NULL, NULL);
2697 g_return_val_if_fail (start_position >= 0, NULL);
2698 g_return_val_if_fail (eval != NULL, NULL);
2699 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2702 string_len = strlen (string);
2704 result = g_string_sized_new (string_len);
2706 /* run down the string making matches. */
2707 g_regex_match_full (regex, string, string_len, start_position,
2708 match_options, &match_info, &tmp_error);
2709 while (!done && g_match_info_matches (match_info))
2711 g_string_append_len (result,
2713 match_info->offsets[0] - str_pos);
2714 done = (*eval) (match_info, result, user_data);
2715 str_pos = match_info->offsets[1];
2716 g_match_info_next (match_info, &tmp_error);
2718 g_match_info_free (match_info);
2719 if (tmp_error != NULL)
2721 g_propagate_error (error, tmp_error);
2722 g_string_free (result, TRUE);
2726 g_string_append_len (result, string + str_pos, string_len - str_pos);
2727 return g_string_free (result, FALSE);
2731 * g_regex_check_replacement:
2732 * @replacement: the replacement string
2733 * @has_references: (out) (allow-none): location to store information about
2734 * references in @replacement or %NULL
2735 * @error: location to store error
2737 * Checks whether @replacement is a valid replacement string
2738 * (see g_regex_replace()), i.e. that all escape sequences in
2741 * If @has_references is not %NULL then @replacement is checked
2742 * for pattern references. For instance, replacement text 'foo\n'
2743 * does not contain references and may be evaluated without information
2744 * about actual match, but '\0\1' (whole match followed by first
2745 * subpattern) requires valid #GMatchInfo object.
2747 * Returns: whether @replacement is a valid replacement string
2752 g_regex_check_replacement (const gchar *replacement,
2753 gboolean *has_references,
2759 list = split_replacement (replacement, &tmp);
2763 g_propagate_error (error, tmp);
2768 *has_references = interpolation_list_needs_match (list);
2770 g_list_foreach (list, (GFunc) free_interpolation_data, NULL);
2777 * g_regex_escape_string:
2778 * @string: (array length=length): the string to escape
2779 * @length: the length of @string, or -1 if @string is nul-terminated
2781 * Escapes the special characters used for regular expressions
2782 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2783 * function is useful to dynamically generate regular expressions.
2785 * @string can contain nul characters that are replaced with "\0",
2786 * in this case remember to specify the correct length of @string
2789 * Returns: a newly-allocated escaped string
2794 g_regex_escape_string (const gchar *string,
2798 const char *p, *piece_start, *end;
2800 g_return_val_if_fail (string != NULL, NULL);
2803 length = strlen (string);
2805 end = string + length;
2806 p = piece_start = string;
2807 escaped = g_string_sized_new (length + 1);
2828 if (p != piece_start)
2829 /* copy the previous piece. */
2830 g_string_append_len (escaped, piece_start, p - piece_start);
2831 g_string_append_c (escaped, '\\');
2833 g_string_append_c (escaped, '0');
2835 g_string_append_c (escaped, *p);
2839 p = g_utf8_next_char (p);
2844 if (piece_start < end)
2845 g_string_append_len (escaped, piece_start, end - piece_start);
2847 return g_string_free (escaped, FALSE);