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 n_offsets; /* number of offsets */
143 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
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");
233 case PCRE_ERROR_BADOFFSET:
234 return _("bad offset");
235 case PCRE_ERROR_SHORTUTF8:
236 return _("short utf8");
240 return _("unknown error");
244 translate_compile_error (gint *errcode, const gchar **errmsg)
246 /* Compile errors are created adding 100 to the error code returned
248 * If errcode is known we put the translatable error message in
249 * erromsg. If errcode is unknown we put the generic
250 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
251 * untranslated error message returned by PCRE.
252 * Note that there can be more PCRE errors with the same GRegexError
253 * and that some PCRE errors are useless for us.
259 case G_REGEX_ERROR_STRAY_BACKSLASH:
260 *errmsg = _("\\ at end of pattern");
262 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
263 *errmsg = _("\\c at end of pattern");
265 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
266 *errmsg = _("unrecognized character follows \\");
269 /* A number of Perl escapes are not handled by PCRE.
270 * Therefore it explicitly raises ERR37.
272 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
273 *errmsg = _("case-changing escapes (\\l, \\L, \\u, \\U) are not allowed here");
275 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
276 *errmsg = _("numbers out of order in {} quantifier");
278 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
279 *errmsg = _("number too big in {} quantifier");
281 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
282 *errmsg = _("missing terminating ] for character class");
284 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
285 *errmsg = _("invalid escape sequence in character class");
287 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
288 *errmsg = _("range out of order in character class");
290 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
291 *errmsg = _("nothing to repeat");
293 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
294 *errmsg = _("unrecognized character after (?");
297 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
298 *errmsg = _("unrecognized character after (?<");
301 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
302 *errmsg = _("unrecognized character after (?P");
304 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
305 *errmsg = _("POSIX named classes are supported only within a class");
307 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
308 *errmsg = _("missing terminating )");
311 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
312 *errmsg = _(") without opening (");
315 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
316 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
317 * sequences here, '(?-54' would be an example for the second group.
319 *errmsg = _("(?R or (?[+-]digits must be followed by )");
321 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
322 *errmsg = _("reference to non-existent subpattern");
324 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
325 *errmsg = _("missing ) after comment");
327 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
328 *errmsg = _("regular expression too large");
330 case G_REGEX_ERROR_MEMORY_ERROR:
331 *errmsg = _("failed to get memory");
333 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
334 *errmsg = _("lookbehind assertion is not fixed length");
336 case G_REGEX_ERROR_MALFORMED_CONDITION:
337 *errmsg = _("malformed number or name after (?(");
339 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
340 *errmsg = _("conditional group contains more than two branches");
342 case G_REGEX_ERROR_ASSERTION_EXPECTED:
343 *errmsg = _("assertion expected after (?(");
345 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
346 *errmsg = _("unknown POSIX class name");
348 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
349 *errmsg = _("POSIX collating elements are not supported");
351 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
352 *errmsg = _("character value in \\x{...} sequence is too large");
354 case G_REGEX_ERROR_INVALID_CONDITION:
355 *errmsg = _("invalid condition (?(0)");
357 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
358 *errmsg = _("\\C not allowed in lookbehind assertion");
360 case G_REGEX_ERROR_INFINITE_LOOP:
361 *errmsg = _("recursive call could loop indefinitely");
363 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
364 *errmsg = _("missing terminator in subpattern name");
366 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
367 *errmsg = _("two named subpatterns have the same name");
369 case G_REGEX_ERROR_MALFORMED_PROPERTY:
370 *errmsg = _("malformed \\P or \\p sequence");
372 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
373 *errmsg = _("unknown property name after \\P or \\p");
375 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
376 *errmsg = _("subpattern name is too long (maximum 32 characters)");
378 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
379 *errmsg = _("too many named subpatterns (maximum 10,000)");
381 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
382 *errmsg = _("octal value is greater than \\377");
384 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
385 *errmsg = _("DEFINE group contains more than one branch");
387 case G_REGEX_ERROR_DEFINE_REPETION:
388 *errmsg = _("repeating a DEFINE group is not allowed");
390 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
391 *errmsg = _("inconsistent NEWLINE options");
393 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
394 *errmsg = _("\\g is not followed by a braced name or an optionally "
395 "braced non-zero number");
398 *errcode = G_REGEX_ERROR_INTERNAL;
399 *errmsg = _("unexpected repeat");
402 *errcode = G_REGEX_ERROR_INTERNAL;
403 *errmsg = _("code overflow");
406 *errcode = G_REGEX_ERROR_INTERNAL;
407 *errmsg = _("overran compiling workspace");
410 *errcode = G_REGEX_ERROR_INTERNAL;
411 *errmsg = _("previously-checked referenced subpattern not found");
414 /* This should not happen as we never pass a NULL erroffset */
415 g_warning ("erroffset passed as NULL");
416 *errcode = G_REGEX_ERROR_COMPILE;
419 /* This should not happen as we check options before passing them
420 * to pcre_compile2() */
421 g_warning ("unknown option bit(s) set");
422 *errcode = G_REGEX_ERROR_COMPILE;
427 /* These errors should not happen as we are using an UTF8-enabled PCRE
428 * and we do not check if strings are valid */
429 g_warning ("%s", *errmsg);
430 *errcode = G_REGEX_ERROR_COMPILE;
433 *errcode = G_REGEX_ERROR_COMPILE;
440 match_info_new (const GRegex *regex,
447 GMatchInfo *match_info;
450 string_len = strlen (string);
452 match_info = g_new0 (GMatchInfo, 1);
453 match_info->regex = g_regex_ref ((GRegex *)regex);
454 match_info->string = string;
455 match_info->string_len = string_len;
456 match_info->matches = PCRE_ERROR_NOMATCH;
457 match_info->pos = start_position;
458 match_info->match_opts = match_options;
462 /* These values should be enough for most cases, if they are not
463 * enough g_regex_match_all_full() will expand them. */
464 match_info->n_offsets = 24;
465 match_info->n_workspace = 100;
466 match_info->workspace = g_new (gint, match_info->n_workspace);
471 pcre_fullinfo (regex->pcre_re, regex->extra,
472 PCRE_INFO_CAPTURECOUNT, &capture_count);
473 match_info->n_offsets = (capture_count + 1) * 3;
476 match_info->offsets = g_new0 (gint, match_info->n_offsets);
477 /* Set an invalid position for the previous match. */
478 match_info->offsets[0] = -1;
479 match_info->offsets[1] = -1;
485 * g_match_info_get_regex:
486 * @match_info: a #GMatchInfo
488 * Returns #GRegex object used in @match_info. It belongs to Glib
489 * and must not be freed. Use g_regex_ref() if you need to keep it
490 * after you free @match_info object.
492 * Returns: #GRegex object used in @match_info
497 g_match_info_get_regex (const GMatchInfo *match_info)
499 g_return_val_if_fail (match_info != NULL, NULL);
500 return match_info->regex;
504 * g_match_info_get_string:
505 * @match_info: a #GMatchInfo
507 * Returns the string searched with @match_info. This is the
508 * string passed to g_regex_match() or g_regex_replace() so
509 * you may not free it before calling this function.
511 * Returns: the string searched with @match_info
516 g_match_info_get_string (const GMatchInfo *match_info)
518 g_return_val_if_fail (match_info != NULL, NULL);
519 return match_info->string;
524 * @match_info: a #GMatchInfo
526 * Frees all the memory associated with the #GMatchInfo structure.
531 g_match_info_free (GMatchInfo *match_info)
535 g_regex_unref (match_info->regex);
536 g_free (match_info->offsets);
537 g_free (match_info->workspace);
544 * @match_info: a #GMatchInfo structure
545 * @error: location to store the error occuring, or %NULL to ignore errors
547 * Scans for the next match using the same parameters of the previous
548 * call to g_regex_match_full() or g_regex_match() that returned
551 * The match is done on the string passed to the match function, so you
552 * cannot free it before calling this function.
554 * Returns: %TRUE is the string matched, %FALSE otherwise
559 g_match_info_next (GMatchInfo *match_info,
562 gint prev_match_start;
565 g_return_val_if_fail (match_info != NULL, FALSE);
566 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
567 g_return_val_if_fail (match_info->pos >= 0, FALSE);
569 prev_match_start = match_info->offsets[0];
570 prev_match_end = match_info->offsets[1];
572 if (match_info->pos > match_info->string_len)
574 /* we have reached the end of the string */
575 match_info->pos = -1;
576 match_info->matches = PCRE_ERROR_NOMATCH;
580 match_info->matches = pcre_exec (match_info->regex->pcre_re,
581 match_info->regex->extra,
583 match_info->string_len,
585 match_info->regex->match_opts | match_info->match_opts,
587 match_info->n_offsets);
588 if (IS_PCRE_ERROR (match_info->matches))
590 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
591 _("Error while matching regular expression %s: %s"),
592 match_info->regex->pattern, match_error (match_info->matches));
596 /* avoid infinite loops if the pattern is an empty string or something
598 if (match_info->pos == match_info->offsets[1])
600 if (match_info->pos > match_info->string_len)
602 /* we have reached the end of the string */
603 match_info->pos = -1;
604 match_info->matches = PCRE_ERROR_NOMATCH;
608 match_info->pos = NEXT_CHAR (match_info->regex,
609 &match_info->string[match_info->pos]) -
614 match_info->pos = match_info->offsets[1];
617 /* it's possible to get two identical matches when we are matching
618 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
619 * the string is "RegExTest" we have:
620 * - search at position 0: match from 0 to 0
621 * - search at position 1: match from 3 to 3
622 * - search at position 3: match from 3 to 3 (duplicate)
623 * - search at position 4: match from 5 to 5
624 * - search at position 5: match from 5 to 5 (duplicate)
625 * - search at position 6: no match -> stop
626 * so we have to ignore the duplicates.
627 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
628 if (match_info->matches >= 0 &&
629 prev_match_start == match_info->offsets[0] &&
630 prev_match_end == match_info->offsets[1])
632 /* ignore this match and search the next one */
633 return g_match_info_next (match_info, error);
636 return match_info->matches >= 0;
640 * g_match_info_matches:
641 * @match_info: a #GMatchInfo structure
643 * Returns whether the previous match operation succeeded.
645 * Returns: %TRUE if the previous match operation succeeded,
651 g_match_info_matches (const GMatchInfo *match_info)
653 g_return_val_if_fail (match_info != NULL, FALSE);
655 return match_info->matches >= 0;
659 * g_match_info_get_match_count:
660 * @match_info: a #GMatchInfo structure
662 * Retrieves the number of matched substrings (including substring 0,
663 * that is the whole matched text), so 1 is returned if the pattern
664 * has no substrings in it and 0 is returned if the match failed.
666 * If the last match was obtained using the DFA algorithm, that is
667 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
668 * count is not that of the number of capturing parentheses but that of
669 * the number of matched substrings.
671 * Returns: Number of matched substrings, or -1 if an error occurred
676 g_match_info_get_match_count (const GMatchInfo *match_info)
678 g_return_val_if_fail (match_info, -1);
680 if (match_info->matches == PCRE_ERROR_NOMATCH)
683 else if (match_info->matches < PCRE_ERROR_NOMATCH)
688 return match_info->matches;
692 * g_match_info_is_partial_match:
693 * @match_info: a #GMatchInfo structure
695 * Usually if the string passed to g_regex_match*() matches as far as
696 * it goes, but is too short to match the entire pattern, %FALSE is
697 * returned. There are circumstances where it might be helpful to
698 * distinguish this case from other cases in which there is no match.
700 * Consider, for example, an application where a human is required to
701 * type in data for a field with specific formatting requirements. An
702 * example might be a date in the form ddmmmyy, defined by the pattern
703 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
704 * If the application sees the user’s keystrokes one by one, and can
705 * check that what has been typed so far is potentially valid, it is
706 * able to raise an error as soon as a mistake is made.
708 * GRegex supports the concept of partial matching by means of the
709 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
710 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
711 * for a complete match, %FALSE otherwise. But, when these functions
712 * return %FALSE, you can check if the match was partial calling
713 * g_match_info_is_partial_match().
715 * When using partial matching you cannot use g_match_info_fetch*().
717 * Because of the way certain internal optimizations are implemented
718 * the partial matching algorithm cannot be used with all patterns.
719 * So repeated single characters such as "a{2,4}" and repeated single
720 * meta-sequences such as "\d+" are not permitted if the maximum number
721 * of occurrences is greater than one. Optional items such as "\d?"
722 * (where the maximum is one) are permitted. Quantifiers with any values
723 * are permitted after parentheses, so the invalid examples above can be
724 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
725 * for a pattern that does not conform to the restrictions, matching
726 * functions return an error.
728 * Returns: %TRUE if the match was partial, %FALSE otherwise
733 g_match_info_is_partial_match (const GMatchInfo *match_info)
735 g_return_val_if_fail (match_info != NULL, FALSE);
737 return match_info->matches == PCRE_ERROR_PARTIAL;
741 * g_match_info_expand_references:
742 * @match_info: a #GMatchInfo or %NULL
743 * @string_to_expand: the string to expand
744 * @error: location to store the error occuring, or %NULL to ignore errors
746 * Returns a new string containing the text in @string_to_expand with
747 * references and escape sequences expanded. References refer to the last
748 * match done with @string against @regex and have the same syntax used by
751 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
752 * passed to g_regex_new().
754 * The backreferences are extracted from the string passed to the match
755 * function, so you cannot call this function after freeing the string.
757 * @match_info may be %NULL in which case @string_to_expand must not
758 * contain references. For instance "foo\n" does not refer to an actual
759 * pattern and '\n' merely will be replaced with \n character,
760 * while to expand "\0" (whole match) one needs the result of a match.
761 * Use g_regex_check_replacement() to find out whether @string_to_expand
762 * contains references.
764 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
769 g_match_info_expand_references (const GMatchInfo *match_info,
770 const gchar *string_to_expand,
775 GError *tmp_error = NULL;
777 g_return_val_if_fail (string_to_expand != NULL, NULL);
778 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
780 list = split_replacement (string_to_expand, &tmp_error);
781 if (tmp_error != NULL)
783 g_propagate_error (error, tmp_error);
787 if (!match_info && interpolation_list_needs_match (list))
789 g_critical ("String '%s' contains references to the match, can't "
790 "expand references without GMatchInfo object",
795 result = g_string_sized_new (strlen (string_to_expand));
796 interpolate_replacement (match_info, result, list);
798 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
801 return g_string_free (result, FALSE);
805 * g_match_info_fetch:
806 * @match_info: #GMatchInfo structure
807 * @match_num: number of the sub expression
809 * Retrieves the text matching the @match_num<!-- -->'th capturing
810 * parentheses. 0 is the full text of the match, 1 is the first paren
811 * set, 2 the second, and so on.
813 * If @match_num is a valid sub pattern but it didn't match anything
814 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
815 * string is returned.
817 * If the match was obtained using the DFA algorithm, that is using
818 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
819 * string is not that of a set of parentheses but that of a matched
820 * substring. Substrings are matched in reverse order of length, so
821 * 0 is the longest match.
823 * The string is fetched from the string passed to the match function,
824 * so you cannot call this function after freeing the string.
826 * Returns: (allow-none): The matched substring, or %NULL if an error
827 * occurred. You have to free the string yourself
832 g_match_info_fetch (const GMatchInfo *match_info,
835 /* we cannot use pcre_get_substring() because it allocates the
836 * string using pcre_malloc(). */
840 g_return_val_if_fail (match_info != NULL, NULL);
841 g_return_val_if_fail (match_num >= 0, NULL);
843 /* match_num does not exist or it didn't matched, i.e. matching "b"
844 * against "(a)?b" then group 0 is empty. */
845 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
847 else if (start == -1)
848 match = g_strdup ("");
850 match = g_strndup (&match_info->string[start], end - start);
856 * g_match_info_fetch_pos:
857 * @match_info: #GMatchInfo structure
858 * @match_num: number of the sub expression
859 * @start_pos: (out) (allow-none): pointer to location where to store
860 * the start position, or %NULL
861 * @end_pos: (out) (allow-none): pointer to location where to store
862 * the end position, or %NULL
864 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
865 * parentheses. 0 is the full text of the match, 1 is the first
866 * paren set, 2 the second, and so on.
868 * If @match_num is a valid sub pattern but it didn't match anything
869 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
870 * and @end_pos are set to -1 and %TRUE is returned.
872 * If the match was obtained using the DFA algorithm, that is using
873 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
874 * position is not that of a set of parentheses but that of a matched
875 * substring. Substrings are matched in reverse order of length, so
876 * 0 is the longest match.
878 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
879 * the position cannot be fetched, @start_pos and @end_pos are left
885 g_match_info_fetch_pos (const GMatchInfo *match_info,
890 g_return_val_if_fail (match_info != NULL, FALSE);
891 g_return_val_if_fail (match_num >= 0, FALSE);
893 /* make sure the sub expression number they're requesting is less than
894 * the total number of sub expressions that were matched. */
895 if (match_num >= match_info->matches)
898 if (start_pos != NULL)
899 *start_pos = match_info->offsets[2 * match_num];
902 *end_pos = match_info->offsets[2 * match_num + 1];
908 * Returns number of first matched subpattern with name @name.
909 * There may be more than one in case when DUPNAMES is used,
910 * and not all subpatterns with that name match;
911 * pcre_get_stringnumber() does not work in that case.
914 get_matched_substring_number (const GMatchInfo *match_info,
921 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
922 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
924 /* This code is copied from pcre_get.c: get_first_set() */
925 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
933 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
935 gint n = (entry[0] << 8) + entry[1];
936 if (match_info->offsets[n*2] >= 0)
940 return (first[0] << 8) + first[1];
944 * g_match_info_fetch_named:
945 * @match_info: #GMatchInfo structure
946 * @name: name of the subexpression
948 * Retrieves the text matching the capturing parentheses named @name.
950 * If @name is a valid sub pattern name but it didn't match anything
951 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
952 * then an empty string is returned.
954 * The string is fetched from the string passed to the match function,
955 * so you cannot call this function after freeing the string.
957 * Returns: (allow-none): The matched substring, or %NULL if an error
958 * occurred. You have to free the string yourself
963 g_match_info_fetch_named (const GMatchInfo *match_info,
966 /* we cannot use pcre_get_named_substring() because it allocates the
967 * string using pcre_malloc(). */
970 g_return_val_if_fail (match_info != NULL, NULL);
971 g_return_val_if_fail (name != NULL, NULL);
973 num = get_matched_substring_number (match_info, name);
977 return g_match_info_fetch (match_info, num);
981 * g_match_info_fetch_named_pos:
982 * @match_info: #GMatchInfo structure
983 * @name: name of the subexpression
984 * @start_pos: (out) (allow-none): pointer to location where to store
985 * the start position, or %NULL
986 * @end_pos: (out) (allow-none): pointer to location where to store
987 * the end position, or %NULL
989 * Retrieves the position in bytes of the capturing parentheses named @name.
991 * If @name is a valid sub pattern name but it didn't match anything
992 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
993 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
995 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
996 * If the position cannot be fetched, @start_pos and @end_pos
997 * are left unchanged.
1002 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
1009 g_return_val_if_fail (match_info != NULL, FALSE);
1010 g_return_val_if_fail (name != NULL, FALSE);
1012 num = get_matched_substring_number (match_info, name);
1016 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1020 * g_match_info_fetch_all:
1021 * @match_info: a #GMatchInfo structure
1023 * Bundles up pointers to each of the matching substrings from a match
1024 * and stores them in an array of gchar pointers. The first element in
1025 * the returned array is the match number 0, i.e. the entire matched
1028 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1029 * "b" against "(a)?b") then an empty string is inserted.
1031 * If the last match was obtained using the DFA algorithm, that is using
1032 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1033 * strings are not that matched by sets of parentheses but that of the
1034 * matched substring. Substrings are matched in reverse order of length,
1035 * so the first one is the longest match.
1037 * The strings are fetched from the string passed to the match function,
1038 * so you cannot call this function after freeing the string.
1040 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1041 * It must be freed using g_strfreev(). If the previous match failed
1047 g_match_info_fetch_all (const GMatchInfo *match_info)
1049 /* we cannot use pcre_get_substring_list() because the returned value
1050 * isn't suitable for g_strfreev(). */
1054 g_return_val_if_fail (match_info != NULL, NULL);
1056 if (match_info->matches < 0)
1059 result = g_new (gchar *, match_info->matches + 1);
1060 for (i = 0; i < match_info->matches; i++)
1061 result[i] = g_match_info_fetch (match_info, i);
1071 g_regex_error_quark (void)
1073 static GQuark error_quark = 0;
1075 if (error_quark == 0)
1076 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1085 * Increases reference count of @regex by 1.
1092 g_regex_ref (GRegex *regex)
1094 g_return_val_if_fail (regex != NULL, NULL);
1095 g_atomic_int_inc (®ex->ref_count);
1103 * Decreases reference count of @regex by 1. When reference count drops
1104 * to zero, it frees all the memory associated with the regex structure.
1109 g_regex_unref (GRegex *regex)
1111 g_return_if_fail (regex != NULL);
1113 if (g_atomic_int_dec_and_test (®ex->ref_count))
1115 g_free (regex->pattern);
1116 if (regex->pcre_re != NULL)
1117 pcre_free (regex->pcre_re);
1118 if (regex->extra != NULL)
1119 pcre_free (regex->extra);
1126 * @pattern: the regular expression
1127 * @compile_options: compile options for the regular expression, or 0
1128 * @match_options: match options for the regular expression, or 0
1129 * @error: return location for a #GError
1131 * Compiles the regular expression to an internal form, and does
1132 * the initial setup of the #GRegex structure.
1134 * Returns: a #GRegex structure. Call g_regex_unref() when you
1140 g_regex_new (const gchar *pattern,
1141 GRegexCompileFlags compile_options,
1142 GRegexMatchFlags match_options,
1147 const gchar *errmsg;
1150 gboolean optimize = FALSE;
1151 static gboolean initialized = FALSE;
1152 unsigned long int pcre_compile_options;
1154 g_return_val_if_fail (pattern != NULL, NULL);
1155 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1156 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1157 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1164 pcre_config (PCRE_CONFIG_UTF8, &support);
1167 msg = N_("PCRE library is compiled without UTF8 support");
1168 g_critical ("%s", msg);
1169 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1173 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1176 msg = N_("PCRE library is compiled without UTF8 properties support");
1177 g_critical ("%s", msg);
1178 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1185 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1186 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1187 if (compile_options & G_REGEX_OPTIMIZE)
1190 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1191 * instead uses UTF-8 only if required with PCRE_UTF8. */
1192 if (compile_options & G_REGEX_RAW)
1195 compile_options &= ~G_REGEX_RAW;
1200 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1201 match_options |= PCRE_NO_UTF8_CHECK;
1204 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1205 * not for the system one. */
1206 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1207 !(compile_options & G_REGEX_NEWLINE_LF))
1209 compile_options |= PCRE_NEWLINE_ANY;
1212 compile_options |= PCRE_UCP;
1214 /* compile the pattern */
1215 re = pcre_compile2 (pattern, compile_options, &errcode,
1216 &errmsg, &erroffset, NULL);
1218 /* if the compilation failed, set the error member and return
1224 /* Translate the PCRE error code to GRegexError and use a translated
1225 * error message if possible */
1226 translate_compile_error (&errcode, &errmsg);
1228 /* PCRE uses byte offsets but we want to show character offsets */
1229 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1231 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1232 _("Error while compiling regular "
1233 "expression %s at char %d: %s"),
1234 pattern, erroffset, errmsg);
1235 g_propagate_error (error, tmp_error);
1240 /* For options set at the beginning of the pattern, pcre puts them into
1241 * compile options, e.g. "(?i)foo" will make the pcre structure store
1242 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1243 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1244 compile_options = pcre_compile_options;
1246 if (!(compile_options & G_REGEX_DUPNAMES))
1248 gboolean jchanged = FALSE;
1249 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1251 compile_options |= G_REGEX_DUPNAMES;
1254 regex = g_new0 (GRegex, 1);
1255 regex->ref_count = 1;
1256 regex->pattern = g_strdup (pattern);
1257 regex->pcre_re = re;
1258 regex->compile_opts = compile_options;
1259 regex->match_opts = match_options;
1263 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1266 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1267 G_REGEX_ERROR_OPTIMIZE,
1268 _("Error while optimizing "
1269 "regular expression %s: %s"),
1272 g_propagate_error (error, tmp_error);
1274 g_regex_unref (regex);
1283 * g_regex_get_pattern:
1284 * @regex: a #GRegex structure
1286 * Gets the pattern string associated with @regex, i.e. a copy of
1287 * the string passed to g_regex_new().
1289 * Returns: the pattern of @regex
1294 g_regex_get_pattern (const GRegex *regex)
1296 g_return_val_if_fail (regex != NULL, NULL);
1298 return regex->pattern;
1302 * g_regex_get_max_backref:
1305 * Returns the number of the highest back reference
1306 * in the pattern, or 0 if the pattern does not contain
1309 * Returns: the number of the highest back reference
1314 g_regex_get_max_backref (const GRegex *regex)
1318 pcre_fullinfo (regex->pcre_re, regex->extra,
1319 PCRE_INFO_BACKREFMAX, &value);
1325 * g_regex_get_capture_count:
1328 * Returns the number of capturing subpatterns in the pattern.
1330 * Returns: the number of capturing subpatterns
1335 g_regex_get_capture_count (const GRegex *regex)
1339 pcre_fullinfo (regex->pcre_re, regex->extra,
1340 PCRE_INFO_CAPTURECOUNT, &value);
1346 * g_regex_get_compile_flags:
1349 * Returns the compile options that @regex was created with.
1351 * Returns: flags from #GRegexCompileFlags
1356 g_regex_get_compile_flags (const GRegex *regex)
1358 g_return_val_if_fail (regex != NULL, 0);
1360 return regex->compile_opts;
1364 * g_regex_get_match_flags:
1367 * Returns the match options that @regex was created with.
1369 * Returns: flags from #GRegexMatchFlags
1374 g_regex_get_match_flags (const GRegex *regex)
1376 g_return_val_if_fail (regex != NULL, 0);
1378 return regex->match_opts;
1382 * g_regex_match_simple:
1383 * @pattern: the regular expression
1384 * @string: the string to scan for matches
1385 * @compile_options: compile options for the regular expression, or 0
1386 * @match_options: match options, or 0
1388 * Scans for a match in @string for @pattern.
1390 * This function is equivalent to g_regex_match() but it does not
1391 * require to compile the pattern with g_regex_new(), avoiding some
1392 * lines of code when you need just to do a match without extracting
1393 * substrings, capture counts, and so on.
1395 * If this function is to be called on the same @pattern more than
1396 * once, it's more efficient to compile the pattern once with
1397 * g_regex_new() and then use g_regex_match().
1399 * Returns: %TRUE if the string matched, %FALSE otherwise
1404 g_regex_match_simple (const gchar *pattern,
1405 const gchar *string,
1406 GRegexCompileFlags compile_options,
1407 GRegexMatchFlags match_options)
1412 regex = g_regex_new (pattern, compile_options, 0, NULL);
1415 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1416 g_regex_unref (regex);
1422 * @regex: a #GRegex structure from g_regex_new()
1423 * @string: the string to scan for matches
1424 * @match_options: match options
1425 * @match_info: (out) (allow-none): pointer to location where to store
1426 * the #GMatchInfo, or %NULL if you do not need it
1428 * Scans for a match in string for the pattern in @regex.
1429 * The @match_options are combined with the match options specified
1430 * when the @regex structure was created, letting you have more
1431 * flexibility in reusing #GRegex structures.
1433 * A #GMatchInfo structure, used to get information on the match,
1434 * is stored in @match_info if not %NULL. Note that if @match_info
1435 * is not %NULL then it is created even if the function returns %FALSE,
1436 * i.e. you must free it regardless if regular expression actually matched.
1438 * To retrieve all the non-overlapping matches of the pattern in
1439 * string you can use g_match_info_next().
1443 * print_uppercase_words (const gchar *string)
1445 * /* Print all uppercase-only words. */
1447 * GMatchInfo *match_info;
1449 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1450 * g_regex_match (regex, string, 0, &match_info);
1451 * while (g_match_info_matches (match_info))
1453 * gchar *word = g_match_info_fetch (match_info, 0);
1454 * g_print ("Found: %s\n", word);
1456 * g_match_info_next (match_info, NULL);
1458 * g_match_info_free (match_info);
1459 * g_regex_unref (regex);
1463 * @string is not copied and is used in #GMatchInfo internally. If
1464 * you use any #GMatchInfo method (except g_match_info_free()) after
1465 * freeing or modifying @string then the behaviour is undefined.
1467 * Returns: %TRUE is the string matched, %FALSE otherwise
1472 g_regex_match (const GRegex *regex,
1473 const gchar *string,
1474 GRegexMatchFlags match_options,
1475 GMatchInfo **match_info)
1477 return g_regex_match_full (regex, string, -1, 0, match_options,
1482 * g_regex_match_full:
1483 * @regex: a #GRegex structure from g_regex_new()
1484 * @string: (array length=string_len): the string to scan for matches
1485 * @string_len: the length of @string, or -1 if @string is nul-terminated
1486 * @start_position: starting index of the string to match
1487 * @match_options: match options
1488 * @match_info: (out) (allow-none): pointer to location where to store
1489 * the #GMatchInfo, or %NULL if you do not need it
1490 * @error: location to store the error occuring, or %NULL to ignore errors
1492 * Scans for a match in string for the pattern in @regex.
1493 * The @match_options are combined with the match options specified
1494 * when the @regex structure was created, letting you have more
1495 * flexibility in reusing #GRegex structures.
1497 * Setting @start_position differs from just passing over a shortened
1498 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1499 * that begins with any kind of lookbehind assertion, such as "\b".
1501 * A #GMatchInfo structure, used to get information on the match, is
1502 * stored in @match_info if not %NULL. Note that if @match_info is
1503 * not %NULL then it is created even if the function returns %FALSE,
1504 * i.e. you must free it regardless if regular expression actually
1507 * @string is not copied and is used in #GMatchInfo internally. If
1508 * you use any #GMatchInfo method (except g_match_info_free()) after
1509 * freeing or modifying @string then the behaviour is undefined.
1511 * To retrieve all the non-overlapping matches of the pattern in
1512 * string you can use g_match_info_next().
1516 * print_uppercase_words (const gchar *string)
1518 * /* Print all uppercase-only words. */
1520 * GMatchInfo *match_info;
1521 * GError *error = NULL;
1523 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1524 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1525 * while (g_match_info_matches (match_info))
1527 * gchar *word = g_match_info_fetch (match_info, 0);
1528 * g_print ("Found: %s\n", word);
1530 * g_match_info_next (match_info, &error);
1532 * g_match_info_free (match_info);
1533 * g_regex_unref (regex);
1534 * if (error != NULL)
1536 * g_printerr ("Error while matching: %s\n", error->message);
1537 * g_error_free (error);
1542 * Returns: %TRUE is the string matched, %FALSE otherwise
1547 g_regex_match_full (const GRegex *regex,
1548 const gchar *string,
1550 gint start_position,
1551 GRegexMatchFlags match_options,
1552 GMatchInfo **match_info,
1558 g_return_val_if_fail (regex != NULL, FALSE);
1559 g_return_val_if_fail (string != NULL, FALSE);
1560 g_return_val_if_fail (start_position >= 0, FALSE);
1561 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1562 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1564 info = match_info_new (regex, string, string_len, start_position,
1565 match_options, FALSE);
1566 match_ok = g_match_info_next (info, error);
1567 if (match_info != NULL)
1570 g_match_info_free (info);
1576 * g_regex_match_all:
1577 * @regex: a #GRegex structure from g_regex_new()
1578 * @string: the string to scan for matches
1579 * @match_options: match options
1580 * @match_info: (out) (allow-none): pointer to location where to store
1581 * the #GMatchInfo, or %NULL if you do not need it
1583 * Using the standard algorithm for regular expression matching only
1584 * the longest match in the string is retrieved. This function uses
1585 * a different algorithm so it can retrieve all the possible matches.
1586 * For more documentation see g_regex_match_all_full().
1588 * A #GMatchInfo structure, used to get information on the match, is
1589 * stored in @match_info if not %NULL. Note that if @match_info is
1590 * not %NULL then it is created even if the function returns %FALSE,
1591 * i.e. you must free it regardless if regular expression actually
1594 * @string is not copied and is used in #GMatchInfo internally. If
1595 * you use any #GMatchInfo method (except g_match_info_free()) after
1596 * freeing or modifying @string then the behaviour is undefined.
1598 * Returns: %TRUE is the string matched, %FALSE otherwise
1603 g_regex_match_all (const GRegex *regex,
1604 const gchar *string,
1605 GRegexMatchFlags match_options,
1606 GMatchInfo **match_info)
1608 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1613 * g_regex_match_all_full:
1614 * @regex: a #GRegex structure from g_regex_new()
1615 * @string: (array length=string_len): the string to scan for matches
1616 * @string_len: the length of @string, or -1 if @string is nul-terminated
1617 * @start_position: starting index of the string to match
1618 * @match_options: match options
1619 * @match_info: (out) (allow-none): pointer to location where to store
1620 * the #GMatchInfo, or %NULL if you do not need it
1621 * @error: location to store the error occuring, or %NULL to ignore errors
1623 * Using the standard algorithm for regular expression matching only
1624 * the longest match in the string is retrieved, it is not possibile
1625 * to obtain all the available matches. For instance matching
1626 * "<a> <b> <c>" against the pattern "<.*>"
1627 * you get "<a> <b> <c>".
1629 * This function uses a different algorithm (called DFA, i.e. deterministic
1630 * finite automaton), so it can retrieve all the possible matches, all
1631 * starting at the same point in the string. For instance matching
1632 * "<a> <b> <c>" against the pattern "<.*>"
1633 * you would obtain three matches: "<a> <b> <c>",
1634 * "<a> <b>" and "<a>".
1636 * The number of matched strings is retrieved using
1637 * g_match_info_get_match_count(). To obtain the matched strings and
1638 * their position you can use, respectively, g_match_info_fetch() and
1639 * g_match_info_fetch_pos(). Note that the strings are returned in
1640 * reverse order of length; that is, the longest matching string is
1643 * Note that the DFA algorithm is slower than the standard one and it
1644 * is not able to capture substrings, so backreferences do not work.
1646 * Setting @start_position differs from just passing over a shortened
1647 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1648 * that begins with any kind of lookbehind assertion, such as "\b".
1650 * A #GMatchInfo structure, used to get information on the match, is
1651 * stored in @match_info if not %NULL. Note that if @match_info is
1652 * not %NULL then it is created even if the function returns %FALSE,
1653 * i.e. you must free it regardless if regular expression actually
1656 * @string is not copied and is used in #GMatchInfo internally. If
1657 * you use any #GMatchInfo method (except g_match_info_free()) after
1658 * freeing or modifying @string then the behaviour is undefined.
1660 * Returns: %TRUE is the string matched, %FALSE otherwise
1665 g_regex_match_all_full (const GRegex *regex,
1666 const gchar *string,
1668 gint start_position,
1669 GRegexMatchFlags match_options,
1670 GMatchInfo **match_info,
1676 g_return_val_if_fail (regex != NULL, FALSE);
1677 g_return_val_if_fail (string != NULL, FALSE);
1678 g_return_val_if_fail (start_position >= 0, FALSE);
1679 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1680 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1682 info = match_info_new (regex, string, string_len, start_position,
1683 match_options, TRUE);
1689 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1690 info->string, info->string_len,
1692 regex->match_opts | match_options,
1693 info->offsets, info->n_offsets,
1694 info->workspace, info->n_workspace);
1695 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1697 /* info->workspace is too small. */
1698 info->n_workspace *= 2;
1699 info->workspace = g_realloc (info->workspace,
1700 info->n_workspace * sizeof (gint));
1703 else if (info->matches == 0)
1705 /* info->offsets is too small. */
1706 info->n_offsets *= 2;
1707 info->offsets = g_realloc (info->offsets,
1708 info->n_offsets * sizeof (gint));
1711 else if (IS_PCRE_ERROR (info->matches))
1713 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1714 _("Error while matching regular expression %s: %s"),
1715 regex->pattern, match_error (info->matches));
1719 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1722 if (match_info != NULL)
1725 g_match_info_free (info);
1727 return info->matches >= 0;
1731 * g_regex_get_string_number:
1732 * @regex: #GRegex structure
1733 * @name: name of the subexpression
1735 * Retrieves the number of the subexpression named @name.
1737 * Returns: The number of the subexpression or -1 if @name
1743 g_regex_get_string_number (const GRegex *regex,
1748 g_return_val_if_fail (regex != NULL, -1);
1749 g_return_val_if_fail (name != NULL, -1);
1751 num = pcre_get_stringnumber (regex->pcre_re, name);
1752 if (num == PCRE_ERROR_NOSUBSTRING)
1759 * g_regex_split_simple:
1760 * @pattern: the regular expression
1761 * @string: the string to scan for matches
1762 * @compile_options: compile options for the regular expression, or 0
1763 * @match_options: match options, or 0
1765 * Breaks the string on the pattern, and returns an array of
1766 * the tokens. If the pattern contains capturing parentheses,
1767 * then the text for each of the substrings will also be returned.
1768 * If the pattern does not match anywhere in the string, then the
1769 * whole string is returned as the first token.
1771 * This function is equivalent to g_regex_split() but it does
1772 * not require to compile the pattern with g_regex_new(), avoiding
1773 * some lines of code when you need just to do a split without
1774 * extracting substrings, capture counts, and so on.
1776 * If this function is to be called on the same @pattern more than
1777 * once, it's more efficient to compile the pattern once with
1778 * g_regex_new() and then use g_regex_split().
1780 * As a special case, the result of splitting the empty string ""
1781 * is an empty vector, not a vector containing a single string.
1782 * The reason for this special case is that being able to represent
1783 * a empty vector is typically more useful than consistent handling
1784 * of empty elements. If you do need to represent empty elements,
1785 * you'll need to check for the empty string before calling this
1788 * A pattern that can match empty strings splits @string into
1789 * separate characters wherever it matches the empty string between
1790 * characters. For example splitting "ab c" using as a separator
1791 * "\s*", you will get "a", "b" and "c".
1793 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1798 g_regex_split_simple (const gchar *pattern,
1799 const gchar *string,
1800 GRegexCompileFlags compile_options,
1801 GRegexMatchFlags match_options)
1806 regex = g_regex_new (pattern, compile_options, 0, NULL);
1810 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1811 g_regex_unref (regex);
1817 * @regex: a #GRegex structure
1818 * @string: the string to split with the pattern
1819 * @match_options: match time option flags
1821 * Breaks the string on the pattern, and returns an array of the tokens.
1822 * If the pattern contains capturing parentheses, then the text for each
1823 * of the substrings will also be returned. If the pattern does not match
1824 * anywhere in the string, then the whole string is returned as the first
1827 * As a special case, the result of splitting the empty string "" is an
1828 * empty vector, not a vector containing a single string. The reason for
1829 * this special case is that being able to represent a empty vector is
1830 * typically more useful than consistent handling of empty elements. If
1831 * you do need to represent empty elements, you'll need to check for the
1832 * empty string before calling this function.
1834 * A pattern that can match empty strings splits @string into separate
1835 * characters wherever it matches the empty string between characters.
1836 * For example splitting "ab c" using as a separator "\s*", you will get
1839 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1844 g_regex_split (const GRegex *regex,
1845 const gchar *string,
1846 GRegexMatchFlags match_options)
1848 return g_regex_split_full (regex, string, -1, 0,
1849 match_options, 0, NULL);
1853 * g_regex_split_full:
1854 * @regex: a #GRegex structure
1855 * @string: (array length=string_len): the string to split with the pattern
1856 * @string_len: the length of @string, or -1 if @string is nul-terminated
1857 * @start_position: starting index of the string to match
1858 * @match_options: match time option flags
1859 * @max_tokens: the maximum number of tokens to split @string into.
1860 * If this is less than 1, the string is split completely
1861 * @error: return location for a #GError
1863 * Breaks the string on the pattern, and returns an array of the tokens.
1864 * If the pattern contains capturing parentheses, then the text for each
1865 * of the substrings will also be returned. If the pattern does not match
1866 * anywhere in the string, then the whole string is returned as the first
1869 * As a special case, the result of splitting the empty string "" is an
1870 * empty vector, not a vector containing a single string. The reason for
1871 * this special case is that being able to represent a empty vector is
1872 * typically more useful than consistent handling of empty elements. If
1873 * you do need to represent empty elements, you'll need to check for the
1874 * empty string before calling this function.
1876 * A pattern that can match empty strings splits @string into separate
1877 * characters wherever it matches the empty string between characters.
1878 * For example splitting "ab c" using as a separator "\s*", you will get
1881 * Setting @start_position differs from just passing over a shortened
1882 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1883 * that begins with any kind of lookbehind assertion, such as "\b".
1885 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1890 g_regex_split_full (const GRegex *regex,
1891 const gchar *string,
1893 gint start_position,
1894 GRegexMatchFlags match_options,
1898 GError *tmp_error = NULL;
1899 GMatchInfo *match_info;
1904 /* position of the last separator. */
1905 gint last_separator_end;
1906 /* was the last match 0 bytes long? */
1907 gboolean last_match_is_empty;
1908 /* the returned array of char **s */
1909 gchar **string_list;
1911 g_return_val_if_fail (regex != NULL, NULL);
1912 g_return_val_if_fail (string != NULL, NULL);
1913 g_return_val_if_fail (start_position >= 0, NULL);
1914 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1915 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1917 if (max_tokens <= 0)
1918 max_tokens = G_MAXINT;
1921 string_len = strlen (string);
1923 /* zero-length string */
1924 if (string_len - start_position == 0)
1925 return g_new0 (gchar *, 1);
1927 if (max_tokens == 1)
1929 string_list = g_new0 (gchar *, 2);
1930 string_list[0] = g_strndup (&string[start_position],
1931 string_len - start_position);
1937 last_separator_end = start_position;
1938 last_match_is_empty = FALSE;
1940 match_ok = g_regex_match_full (regex, string, string_len, start_position,
1941 match_options, &match_info, &tmp_error);
1943 while (tmp_error == NULL)
1947 last_match_is_empty =
1948 (match_info->offsets[0] == match_info->offsets[1]);
1950 /* we need to skip empty separators at the same position of the end
1951 * of another separator. e.g. the string is "a b" and the separator
1952 * is " *", so from 1 to 2 we have a match and at position 2 we have
1953 * an empty match. */
1954 if (last_separator_end != match_info->offsets[1])
1959 token = g_strndup (string + last_separator_end,
1960 match_info->offsets[0] - last_separator_end);
1961 list = g_list_prepend (list, token);
1964 /* if there were substrings, these need to be added to
1966 match_count = g_match_info_get_match_count (match_info);
1967 if (match_count > 1)
1969 for (i = 1; i < match_count; i++)
1970 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
1976 /* if there was no match, copy to end of string. */
1977 if (!last_match_is_empty)
1979 gchar *token = g_strndup (string + last_separator_end,
1980 match_info->string_len - last_separator_end);
1981 list = g_list_prepend (list, token);
1983 /* no more tokens, end the loop. */
1987 /* -1 to leave room for the last part. */
1988 if (token_count >= max_tokens - 1)
1990 /* we have reached the maximum number of tokens, so we copy
1991 * the remaining part of the string. */
1992 if (last_match_is_empty)
1994 /* the last match was empty, so we have moved one char
1995 * after the real position to avoid empty matches at the
1997 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
1999 /* the if is needed in the case we have terminated the available
2000 * tokens, but we are at the end of the string, so there are no
2001 * characters left to copy. */
2002 if (string_len > match_info->pos)
2004 gchar *token = g_strndup (string + match_info->pos,
2005 string_len - match_info->pos);
2006 list = g_list_prepend (list, token);
2012 last_separator_end = match_info->pos;
2013 if (last_match_is_empty)
2014 /* if the last match was empty, g_match_info_next() has moved
2015 * forward to avoid infinite loops, but we still need to copy that
2017 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
2019 match_ok = g_match_info_next (match_info, &tmp_error);
2021 g_match_info_free (match_info);
2022 if (tmp_error != NULL)
2024 g_propagate_error (error, tmp_error);
2025 g_list_foreach (list, (GFunc)g_free, NULL);
2027 match_info->pos = -1;
2031 string_list = g_new (gchar *, g_list_length (list) + 1);
2033 for (last = g_list_last (list); last; last = g_list_previous (last))
2034 string_list[i++] = last->data;
2035 string_list[i] = NULL;
2044 REPL_TYPE_CHARACTER,
2045 REPL_TYPE_SYMBOLIC_REFERENCE,
2046 REPL_TYPE_NUMERIC_REFERENCE,
2047 REPL_TYPE_CHANGE_CASE
2052 CHANGE_CASE_NONE = 1 << 0,
2053 CHANGE_CASE_UPPER = 1 << 1,
2054 CHANGE_CASE_LOWER = 1 << 2,
2055 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2056 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2057 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2058 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2059 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2062 struct _InterpolationData
2068 ChangeCase change_case;
2072 free_interpolation_data (InterpolationData *data)
2074 g_free (data->text);
2078 static const gchar *
2079 expand_escape (const gchar *replacement,
2081 InterpolationData *data,
2086 const gchar *error_detail;
2088 GError *tmp_error = NULL;
2096 data->type = REPL_TYPE_CHARACTER;
2101 data->type = REPL_TYPE_CHARACTER;
2106 data->type = REPL_TYPE_CHARACTER;
2111 data->type = REPL_TYPE_CHARACTER;
2116 data->type = REPL_TYPE_CHARACTER;
2121 data->type = REPL_TYPE_CHARACTER;
2126 data->type = REPL_TYPE_CHARACTER;
2131 data->type = REPL_TYPE_CHARACTER;
2141 h = g_ascii_xdigit_value (*p);
2144 error_detail = _("hexadecimal digit or '}' expected");
2155 for (i = 0; i < 2; i++)
2157 h = g_ascii_xdigit_value (*p);
2160 error_detail = _("hexadecimal digit expected");
2167 data->type = REPL_TYPE_STRING;
2168 data->text = g_new0 (gchar, 8);
2169 g_unichar_to_utf8 (x, data->text);
2173 data->type = REPL_TYPE_CHANGE_CASE;
2174 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2178 data->type = REPL_TYPE_CHANGE_CASE;
2179 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2183 data->type = REPL_TYPE_CHANGE_CASE;
2184 data->change_case = CHANGE_CASE_LOWER;
2188 data->type = REPL_TYPE_CHANGE_CASE;
2189 data->change_case = CHANGE_CASE_UPPER;
2193 data->type = REPL_TYPE_CHANGE_CASE;
2194 data->change_case = CHANGE_CASE_NONE;
2200 error_detail = _("missing '<' in symbolic reference");
2209 error_detail = _("unfinished symbolic reference");
2216 error_detail = _("zero-length symbolic reference");
2219 if (g_ascii_isdigit (*q))
2224 h = g_ascii_digit_value (*q);
2227 error_detail = _("digit expected");
2236 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2243 if (!g_ascii_isalnum (*r))
2245 error_detail = _("illegal symbolic reference");
2252 data->text = g_strndup (q, p - q);
2253 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2258 /* if \0 is followed by a number is an octal number representing a
2259 * character, else it is a numeric reference. */
2260 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2263 p = g_utf8_next_char (p);
2276 for (i = 0; i < 3; i++)
2278 h = g_ascii_digit_value (*p);
2288 if (i == 2 && base == 10)
2294 if (base == 8 || i == 3)
2296 data->type = REPL_TYPE_STRING;
2297 data->text = g_new0 (gchar, 8);
2298 g_unichar_to_utf8 (x, data->text);
2302 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2307 error_detail = _("stray final '\\'");
2311 error_detail = _("unknown escape sequence");
2318 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2319 tmp_error = g_error_new (G_REGEX_ERROR,
2320 G_REGEX_ERROR_REPLACE,
2321 _("Error while parsing replacement "
2322 "text \"%s\" at char %lu: %s"),
2324 (gulong)(p - replacement),
2326 g_propagate_error (error, tmp_error);
2332 split_replacement (const gchar *replacement,
2336 InterpolationData *data;
2337 const gchar *p, *start;
2339 start = p = replacement;
2344 data = g_new0 (InterpolationData, 1);
2345 start = p = expand_escape (replacement, p, data, error);
2348 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2350 free_interpolation_data (data);
2354 list = g_list_prepend (list, data);
2359 if (*p == '\\' || *p == '\0')
2363 data = g_new0 (InterpolationData, 1);
2364 data->text = g_strndup (start, p - start);
2365 data->type = REPL_TYPE_STRING;
2366 list = g_list_prepend (list, data);
2372 return g_list_reverse (list);
2375 /* Change the case of c based on change_case. */
2376 #define CHANGE_CASE(c, change_case) \
2377 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2378 g_unichar_tolower (c) : \
2379 g_unichar_toupper (c))
2382 string_append (GString *string,
2384 ChangeCase *change_case)
2388 if (text[0] == '\0')
2391 if (*change_case == CHANGE_CASE_NONE)
2393 g_string_append (string, text);
2395 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2397 c = g_utf8_get_char (text);
2398 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2399 g_string_append (string, g_utf8_next_char (text));
2400 *change_case = CHANGE_CASE_NONE;
2404 while (*text != '\0')
2406 c = g_utf8_get_char (text);
2407 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2408 text = g_utf8_next_char (text);
2414 interpolate_replacement (const GMatchInfo *match_info,
2419 InterpolationData *idata;
2421 ChangeCase change_case = CHANGE_CASE_NONE;
2423 for (list = data; list; list = list->next)
2426 switch (idata->type)
2428 case REPL_TYPE_STRING:
2429 string_append (result, idata->text, &change_case);
2431 case REPL_TYPE_CHARACTER:
2432 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2433 if (change_case & CHANGE_CASE_SINGLE_MASK)
2434 change_case = CHANGE_CASE_NONE;
2436 case REPL_TYPE_NUMERIC_REFERENCE:
2437 match = g_match_info_fetch (match_info, idata->num);
2440 string_append (result, match, &change_case);
2444 case REPL_TYPE_SYMBOLIC_REFERENCE:
2445 match = g_match_info_fetch_named (match_info, idata->text);
2448 string_append (result, match, &change_case);
2452 case REPL_TYPE_CHANGE_CASE:
2453 change_case = idata->change_case;
2461 /* whether actual match_info is needed for replacement, i.e.
2462 * whether there are references
2465 interpolation_list_needs_match (GList *list)
2467 while (list != NULL)
2469 InterpolationData *data = list->data;
2471 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2472 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2485 * @regex: a #GRegex structure
2486 * @string: (array length=string_len): the string to perform matches against
2487 * @string_len: the length of @string, or -1 if @string is nul-terminated
2488 * @start_position: starting index of the string to match
2489 * @replacement: text to replace each match with
2490 * @match_options: options for the match
2491 * @error: location to store the error occuring, or %NULL to ignore errors
2493 * Replaces all occurrences of the pattern in @regex with the
2494 * replacement text. Backreferences of the form '\number' or
2495 * '\g<number>' in the replacement text are interpolated by the
2496 * number-th captured subexpression of the match, '\g<name>' refers
2497 * to the captured subexpression with the given name. '\0' refers to the
2498 * complete match, but '\0' followed by a number is the octal representation
2499 * of a character. To include a literal '\' in the replacement, write '\\'.
2500 * There are also escapes that changes the case of the following text:
2503 * <varlistentry><term>\l</term>
2505 * <para>Convert to lower case the next character</para>
2508 * <varlistentry><term>\u</term>
2510 * <para>Convert to upper case the next character</para>
2513 * <varlistentry><term>\L</term>
2515 * <para>Convert to lower case till \E</para>
2518 * <varlistentry><term>\U</term>
2520 * <para>Convert to upper case till \E</para>
2523 * <varlistentry><term>\E</term>
2525 * <para>End case modification</para>
2530 * If you do not need to use backreferences use g_regex_replace_literal().
2532 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2533 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2534 * you can use g_regex_replace_literal().
2536 * Setting @start_position differs from just passing over a shortened
2537 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2538 * begins with any kind of lookbehind assertion, such as "\b".
2540 * Returns: a newly allocated string containing the replacements
2545 g_regex_replace (const GRegex *regex,
2546 const gchar *string,
2548 gint start_position,
2549 const gchar *replacement,
2550 GRegexMatchFlags match_options,
2555 GError *tmp_error = NULL;
2557 g_return_val_if_fail (regex != NULL, NULL);
2558 g_return_val_if_fail (string != NULL, NULL);
2559 g_return_val_if_fail (start_position >= 0, NULL);
2560 g_return_val_if_fail (replacement != NULL, NULL);
2561 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2562 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2564 list = split_replacement (replacement, &tmp_error);
2565 if (tmp_error != NULL)
2567 g_propagate_error (error, tmp_error);
2571 result = g_regex_replace_eval (regex,
2572 string, string_len, start_position,
2574 interpolate_replacement,
2577 if (tmp_error != NULL)
2578 g_propagate_error (error, tmp_error);
2580 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2587 literal_replacement (const GMatchInfo *match_info,
2591 g_string_append (result, data);
2596 * g_regex_replace_literal:
2597 * @regex: a #GRegex structure
2598 * @string: (array length=string_len): the string to perform matches against
2599 * @string_len: the length of @string, or -1 if @string is nul-terminated
2600 * @start_position: starting index of the string to match
2601 * @replacement: text to replace each match with
2602 * @match_options: options for the match
2603 * @error: location to store the error occuring, or %NULL to ignore errors
2605 * Replaces all occurrences of the pattern in @regex with the
2606 * replacement text. @replacement is replaced literally, to
2607 * include backreferences use g_regex_replace().
2609 * Setting @start_position differs from just passing over a
2610 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2611 * case of a pattern that begins with any kind of lookbehind
2612 * assertion, such as "\b".
2614 * Returns: a newly allocated string containing the replacements
2619 g_regex_replace_literal (const GRegex *regex,
2620 const gchar *string,
2622 gint start_position,
2623 const gchar *replacement,
2624 GRegexMatchFlags match_options,
2627 g_return_val_if_fail (replacement != NULL, NULL);
2628 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2630 return g_regex_replace_eval (regex,
2631 string, string_len, start_position,
2633 literal_replacement,
2634 (gpointer)replacement,
2639 * g_regex_replace_eval:
2640 * @regex: a #GRegex structure from g_regex_new()
2641 * @string: (array length=string_len): string to perform matches against
2642 * @string_len: the length of @string, or -1 if @string is nul-terminated
2643 * @start_position: starting index of the string to match
2644 * @match_options: options for the match
2645 * @eval: a function to call for each match
2646 * @user_data: user data to pass to the function
2647 * @error: location to store the error occuring, or %NULL to ignore errors
2649 * Replaces occurrences of the pattern in regex with the output of
2650 * @eval for that occurrence.
2652 * Setting @start_position differs from just passing over a shortened
2653 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2654 * that begins with any kind of lookbehind assertion, such as "\b".
2656 * The following example uses g_regex_replace_eval() to replace multiple
2660 * eval_cb (const GMatchInfo *info,
2667 * match = g_match_info_fetch (info, 0);
2668 * r = g_hash_table_lookup ((GHashTable *)data, match);
2669 * g_string_append (res, r);
2681 * h = g_hash_table_new (g_str_hash, g_str_equal);
2683 * g_hash_table_insert (h, "1", "ONE");
2684 * g_hash_table_insert (h, "2", "TWO");
2685 * g_hash_table_insert (h, "3", "THREE");
2686 * g_hash_table_insert (h, "4", "FOUR");
2688 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2689 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2690 * g_hash_table_destroy (h);
2695 * Returns: a newly allocated string containing the replacements
2700 g_regex_replace_eval (const GRegex *regex,
2701 const gchar *string,
2703 gint start_position,
2704 GRegexMatchFlags match_options,
2705 GRegexEvalCallback eval,
2709 GMatchInfo *match_info;
2712 gboolean done = FALSE;
2713 GError *tmp_error = NULL;
2715 g_return_val_if_fail (regex != NULL, NULL);
2716 g_return_val_if_fail (string != NULL, NULL);
2717 g_return_val_if_fail (start_position >= 0, NULL);
2718 g_return_val_if_fail (eval != NULL, NULL);
2719 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2722 string_len = strlen (string);
2724 result = g_string_sized_new (string_len);
2726 /* run down the string making matches. */
2727 g_regex_match_full (regex, string, string_len, start_position,
2728 match_options, &match_info, &tmp_error);
2729 while (!done && g_match_info_matches (match_info))
2731 g_string_append_len (result,
2733 match_info->offsets[0] - str_pos);
2734 done = (*eval) (match_info, result, user_data);
2735 str_pos = match_info->offsets[1];
2736 g_match_info_next (match_info, &tmp_error);
2738 g_match_info_free (match_info);
2739 if (tmp_error != NULL)
2741 g_propagate_error (error, tmp_error);
2742 g_string_free (result, TRUE);
2746 g_string_append_len (result, string + str_pos, string_len - str_pos);
2747 return g_string_free (result, FALSE);
2751 * g_regex_check_replacement:
2752 * @replacement: the replacement string
2753 * @has_references: (out) (allow-none): location to store information about
2754 * references in @replacement or %NULL
2755 * @error: location to store error
2757 * Checks whether @replacement is a valid replacement string
2758 * (see g_regex_replace()), i.e. that all escape sequences in
2761 * If @has_references is not %NULL then @replacement is checked
2762 * for pattern references. For instance, replacement text 'foo\n'
2763 * does not contain references and may be evaluated without information
2764 * about actual match, but '\0\1' (whole match followed by first
2765 * subpattern) requires valid #GMatchInfo object.
2767 * Returns: whether @replacement is a valid replacement string
2772 g_regex_check_replacement (const gchar *replacement,
2773 gboolean *has_references,
2779 list = split_replacement (replacement, &tmp);
2783 g_propagate_error (error, tmp);
2788 *has_references = interpolation_list_needs_match (list);
2790 g_list_foreach (list, (GFunc) free_interpolation_data, NULL);
2797 * g_regex_escape_string:
2798 * @string: (array length=length): the string to escape
2799 * @length: the length of @string, or -1 if @string is nul-terminated
2801 * Escapes the special characters used for regular expressions
2802 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2803 * function is useful to dynamically generate regular expressions.
2805 * @string can contain nul characters that are replaced with "\0",
2806 * in this case remember to specify the correct length of @string
2809 * Returns: a newly-allocated escaped string
2814 g_regex_escape_string (const gchar *string,
2818 const char *p, *piece_start, *end;
2820 g_return_val_if_fail (string != NULL, NULL);
2823 length = strlen (string);
2825 end = string + length;
2826 p = piece_start = string;
2827 escaped = g_string_sized_new (length + 1);
2848 if (p != piece_start)
2849 /* copy the previous piece. */
2850 g_string_append_len (escaped, piece_start, p - piece_start);
2851 g_string_append_c (escaped, '\\');
2853 g_string_append_c (escaped, '0');
2855 g_string_append_c (escaped, *p);
2859 p = g_utf8_next_char (p);
2864 if (piece_start < end)
2865 g_string_append_len (escaped, piece_start, end - piece_start);
2867 return g_string_free (escaped, FALSE);