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"
43 * @title: Perl-compatible regular expressions
44 * @short_description: matches strings against regular expressions
45 * @see_also: <xref linkend="glib-regex-syntax"/>
47 * The <function>g_regex_*()</function> functions implement regular
48 * expression pattern matching using syntax and semantics similar to
49 * Perl regular expression.
51 * Some functions accept a @start_position argument, setting it differs
52 * from just passing over a shortened string and setting #G_REGEX_MATCH_NOTBOL
53 * in the case of a pattern that begins with any kind of lookbehind assertion.
54 * For example, consider the pattern "\Biss\B" which finds occurrences of "iss"
55 * in the middle of words. ("\B" matches only if the current position in the
56 * subject is not a word boundary.) When applied to the string "Mississipi"
57 * from the fourth byte, namely "issipi", it does not match, because "\B" is
58 * always false at the start of the subject, which is deemed to be a word
59 * boundary. However, if the entire string is passed , but with
60 * @start_position set to 4, it finds the second occurrence of "iss" because
61 * it is able to look behind the starting point to discover that it is
62 * preceded by a letter.
64 * Note that, unless you set the #G_REGEX_RAW flag, all the strings passed
65 * to these functions must be encoded in UTF-8. The lengths and the positions
66 * inside the strings are in bytes and not in characters, so, for instance,
67 * "\xc3\xa0" (i.e. "à") is two bytes long but it is treated as a
68 * single character. If you set #G_REGEX_RAW the strings can be non-valid
69 * UTF-8 strings and a byte is treated as a character, so "\xc3\xa0" is two
70 * bytes and two characters long.
72 * When matching a pattern, "\n" matches only against a "\n" character in
73 * the string, and "\r" matches only a "\r" character. To match any newline
74 * sequence use "\R". This particular group matches either the two-character
75 * sequence CR + LF ("\r\n"), or one of the single characters LF (linefeed,
76 * U+000A, "\n"), VT vertical tab, U+000B, "\v"), FF (formfeed, U+000C, "\f"),
77 * CR (carriage return, U+000D, "\r"), NEL (next line, U+0085), LS (line
78 * separator, U+2028), or PS (paragraph separator, U+2029).
80 * The behaviour of the dot, circumflex, and dollar metacharacters are
81 * affected by newline characters, the default is to recognize any newline
82 * character (the same characters recognized by "\R"). This can be changed
83 * with #G_REGEX_NEWLINE_CR, #G_REGEX_NEWLINE_LF and #G_REGEX_NEWLINE_CRLF
84 * compile options, and with #G_REGEX_MATCH_NEWLINE_ANY,
85 * #G_REGEX_MATCH_NEWLINE_CR, #G_REGEX_MATCH_NEWLINE_LF and
86 * #G_REGEX_MATCH_NEWLINE_CRLF match options. These settings are also
87 * relevant when compiling a pattern if #G_REGEX_EXTENDED is set, and an
88 * unescaped "#" outside a character class is encountered. This indicates
89 * a comment that lasts until after the next newline.
91 * Creating and manipulating the same #GRegex structure from different
92 * threads is not a problem as #GRegex does not modify its internal
93 * state between creation and destruction, on the other hand #GMatchInfo
96 * The regular expressions low-level functionalities are obtained through
97 * the excellent <ulink url="http://www.pcre.org/">PCRE</ulink> library
98 * written by Philip Hazel.
101 /* Mask of all the possible values for GRegexCompileFlags. */
102 #define G_REGEX_COMPILE_MASK (G_REGEX_CASELESS | \
103 G_REGEX_MULTILINE | \
107 G_REGEX_DOLLAR_ENDONLY | \
110 G_REGEX_NO_AUTO_CAPTURE | \
113 G_REGEX_NEWLINE_CR | \
114 G_REGEX_NEWLINE_LF | \
115 G_REGEX_NEWLINE_CRLF)
117 /* Mask of all the possible values for GRegexMatchFlags. */
118 #define G_REGEX_MATCH_MASK (G_REGEX_MATCH_ANCHORED | \
119 G_REGEX_MATCH_NOTBOL | \
120 G_REGEX_MATCH_NOTEOL | \
121 G_REGEX_MATCH_NOTEMPTY | \
122 G_REGEX_MATCH_PARTIAL | \
123 G_REGEX_MATCH_NEWLINE_CR | \
124 G_REGEX_MATCH_NEWLINE_LF | \
125 G_REGEX_MATCH_NEWLINE_CRLF | \
126 G_REGEX_MATCH_NEWLINE_ANY)
128 /* if the string is in UTF-8 use g_utf8_ functions, else use
130 #define NEXT_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
131 g_utf8_next_char (s) : \
133 #define PREV_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
134 g_utf8_prev_char (s) : \
139 volatile gint ref_count; /* the ref count */
140 GRegex *regex; /* the regex */
141 GRegexMatchFlags match_opts; /* options used at match time on the regex */
142 gint matches; /* number of matching sub patterns */
143 gint pos; /* position in the string where last match left off */
144 gint n_offsets; /* number of offsets */
145 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
146 gint *workspace; /* workspace for pcre_dfa_exec() */
147 gint n_workspace; /* number of workspace elements */
148 const gchar *string; /* string passed to the match function */
149 gssize string_len; /* length of string */
154 volatile gint ref_count; /* the ref count for the immutable part */
155 gchar *pattern; /* the pattern */
156 pcre *pcre_re; /* compiled form of the pattern */
157 GRegexCompileFlags compile_opts; /* options used at compile time on the pattern */
158 GRegexMatchFlags match_opts; /* options used at match time on the regex */
159 pcre_extra *extra; /* data stored when G_REGEX_OPTIMIZE is used */
162 /* TRUE if ret is an error code, FALSE otherwise. */
163 #define IS_PCRE_ERROR(ret) ((ret) < PCRE_ERROR_NOMATCH && (ret) != PCRE_ERROR_PARTIAL)
165 typedef struct _InterpolationData InterpolationData;
166 static gboolean interpolation_list_needs_match (GList *list);
167 static gboolean interpolate_replacement (const GMatchInfo *match_info,
170 static GList *split_replacement (const gchar *replacement,
172 static void free_interpolation_data (InterpolationData *data);
176 match_error (gint errcode)
180 case PCRE_ERROR_NOMATCH:
183 case PCRE_ERROR_NULL:
184 /* NULL argument, this should not happen in GRegex */
185 g_warning ("A NULL argument was passed to PCRE");
187 case PCRE_ERROR_BADOPTION:
188 return "bad options";
189 case PCRE_ERROR_BADMAGIC:
190 return _("corrupted object");
191 case PCRE_ERROR_UNKNOWN_OPCODE:
192 return N_("internal error or corrupted object");
193 case PCRE_ERROR_NOMEMORY:
194 return _("out of memory");
195 case PCRE_ERROR_NOSUBSTRING:
196 /* not used by pcre_exec() */
198 case PCRE_ERROR_MATCHLIMIT:
199 return _("backtracking limit reached");
200 case PCRE_ERROR_CALLOUT:
201 /* callouts are not implemented */
203 case PCRE_ERROR_BADUTF8:
204 case PCRE_ERROR_BADUTF8_OFFSET:
205 /* we do not check if strings are valid */
207 case PCRE_ERROR_PARTIAL:
210 case PCRE_ERROR_BADPARTIAL:
211 return _("the pattern contains items not supported for partial matching");
212 case PCRE_ERROR_INTERNAL:
213 return _("internal error");
214 case PCRE_ERROR_BADCOUNT:
215 /* negative ovecsize, this should not happen in GRegex */
216 g_warning ("A negative ovecsize was passed to PCRE");
218 case PCRE_ERROR_DFA_UITEM:
219 return _("the pattern contains items not supported for partial matching");
220 case PCRE_ERROR_DFA_UCOND:
221 return _("back references as conditions are not supported for partial matching");
222 case PCRE_ERROR_DFA_UMLIMIT:
223 /* the match_field field is not used in GRegex */
225 case PCRE_ERROR_DFA_WSSIZE:
226 /* handled expanding the workspace */
228 case PCRE_ERROR_DFA_RECURSE:
229 case PCRE_ERROR_RECURSIONLIMIT:
230 return _("recursion limit reached");
231 case PCRE_ERROR_NULLWSLIMIT:
232 return _("workspace limit for empty substrings reached");
233 case PCRE_ERROR_BADNEWLINE:
234 return _("invalid combination of newline flags");
235 case PCRE_ERROR_BADOFFSET:
236 return _("bad offset");
237 case PCRE_ERROR_SHORTUTF8:
238 return _("short utf8");
242 return _("unknown error");
246 translate_compile_error (gint *errcode, const gchar **errmsg)
248 /* Compile errors are created adding 100 to the error code returned
250 * If errcode is known we put the translatable error message in
251 * erromsg. If errcode is unknown we put the generic
252 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
253 * untranslated error message returned by PCRE.
254 * Note that there can be more PCRE errors with the same GRegexError
255 * and that some PCRE errors are useless for us.
261 case G_REGEX_ERROR_STRAY_BACKSLASH:
262 *errmsg = _("\\ at end of pattern");
264 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
265 *errmsg = _("\\c at end of pattern");
267 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
268 *errmsg = _("unrecognized character follows \\");
271 /* A number of Perl escapes are not handled by PCRE.
272 * Therefore it explicitly raises ERR37.
274 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
275 *errmsg = _("case-changing escapes (\\l, \\L, \\u, \\U) are not allowed here");
277 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
278 *errmsg = _("numbers out of order in {} quantifier");
280 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
281 *errmsg = _("number too big in {} quantifier");
283 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
284 *errmsg = _("missing terminating ] for character class");
286 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
287 *errmsg = _("invalid escape sequence in character class");
289 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
290 *errmsg = _("range out of order in character class");
292 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
293 *errmsg = _("nothing to repeat");
295 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
296 *errmsg = _("unrecognized character after (?");
299 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
300 *errmsg = _("unrecognized character after (?<");
303 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
304 *errmsg = _("unrecognized character after (?P");
306 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
307 *errmsg = _("POSIX named classes are supported only within a class");
309 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
310 *errmsg = _("missing terminating )");
313 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
314 *errmsg = _(") without opening (");
317 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
318 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
319 * sequences here, '(?-54' would be an example for the second group.
321 *errmsg = _("(?R or (?[+-]digits must be followed by )");
323 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
324 *errmsg = _("reference to non-existent subpattern");
326 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
327 *errmsg = _("missing ) after comment");
329 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
330 *errmsg = _("regular expression too large");
332 case G_REGEX_ERROR_MEMORY_ERROR:
333 *errmsg = _("failed to get memory");
335 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
336 *errmsg = _("lookbehind assertion is not fixed length");
338 case G_REGEX_ERROR_MALFORMED_CONDITION:
339 *errmsg = _("malformed number or name after (?(");
341 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
342 *errmsg = _("conditional group contains more than two branches");
344 case G_REGEX_ERROR_ASSERTION_EXPECTED:
345 *errmsg = _("assertion expected after (?(");
347 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
348 *errmsg = _("unknown POSIX class name");
350 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
351 *errmsg = _("POSIX collating elements are not supported");
353 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
354 *errmsg = _("character value in \\x{...} sequence is too large");
356 case G_REGEX_ERROR_INVALID_CONDITION:
357 *errmsg = _("invalid condition (?(0)");
359 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
360 *errmsg = _("\\C not allowed in lookbehind assertion");
362 case G_REGEX_ERROR_INFINITE_LOOP:
363 *errmsg = _("recursive call could loop indefinitely");
365 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
366 *errmsg = _("missing terminator in subpattern name");
368 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
369 *errmsg = _("two named subpatterns have the same name");
371 case G_REGEX_ERROR_MALFORMED_PROPERTY:
372 *errmsg = _("malformed \\P or \\p sequence");
374 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
375 *errmsg = _("unknown property name after \\P or \\p");
377 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
378 *errmsg = _("subpattern name is too long (maximum 32 characters)");
380 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
381 *errmsg = _("too many named subpatterns (maximum 10,000)");
383 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
384 *errmsg = _("octal value is greater than \\377");
386 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
387 *errmsg = _("DEFINE group contains more than one branch");
389 case G_REGEX_ERROR_DEFINE_REPETION:
390 *errmsg = _("repeating a DEFINE group is not allowed");
392 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
393 *errmsg = _("inconsistent NEWLINE options");
395 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
396 *errmsg = _("\\g is not followed by a braced name or an optionally "
397 "braced non-zero number");
400 *errcode = G_REGEX_ERROR_INTERNAL;
401 *errmsg = _("unexpected repeat");
404 *errcode = G_REGEX_ERROR_INTERNAL;
405 *errmsg = _("code overflow");
408 *errcode = G_REGEX_ERROR_INTERNAL;
409 *errmsg = _("overran compiling workspace");
412 *errcode = G_REGEX_ERROR_INTERNAL;
413 *errmsg = _("previously-checked referenced subpattern not found");
416 /* This should not happen as we never pass a NULL erroffset */
417 g_warning ("erroffset passed as NULL");
418 *errcode = G_REGEX_ERROR_COMPILE;
421 /* This should not happen as we check options before passing them
422 * to pcre_compile2() */
423 g_warning ("unknown option bit(s) set");
424 *errcode = G_REGEX_ERROR_COMPILE;
429 /* These errors should not happen as we are using an UTF8-enabled PCRE
430 * and we do not check if strings are valid */
431 g_warning ("%s", *errmsg);
432 *errcode = G_REGEX_ERROR_COMPILE;
435 *errcode = G_REGEX_ERROR_COMPILE;
442 match_info_new (const GRegex *regex,
449 GMatchInfo *match_info;
452 string_len = strlen (string);
454 match_info = g_new0 (GMatchInfo, 1);
455 match_info->ref_count = 1;
456 match_info->regex = g_regex_ref ((GRegex *)regex);
457 match_info->string = string;
458 match_info->string_len = string_len;
459 match_info->matches = PCRE_ERROR_NOMATCH;
460 match_info->pos = start_position;
461 match_info->match_opts = match_options;
465 /* These values should be enough for most cases, if they are not
466 * enough g_regex_match_all_full() will expand them. */
467 match_info->n_offsets = 24;
468 match_info->n_workspace = 100;
469 match_info->workspace = g_new (gint, match_info->n_workspace);
474 pcre_fullinfo (regex->pcre_re, regex->extra,
475 PCRE_INFO_CAPTURECOUNT, &capture_count);
476 match_info->n_offsets = (capture_count + 1) * 3;
479 match_info->offsets = g_new0 (gint, match_info->n_offsets);
480 /* Set an invalid position for the previous match. */
481 match_info->offsets[0] = -1;
482 match_info->offsets[1] = -1;
488 * g_match_info_get_regex:
489 * @match_info: a #GMatchInfo
491 * Returns #GRegex object used in @match_info. It belongs to Glib
492 * and must not be freed. Use g_regex_ref() if you need to keep it
493 * after you free @match_info object.
495 * Returns: #GRegex object used in @match_info
500 g_match_info_get_regex (const GMatchInfo *match_info)
502 g_return_val_if_fail (match_info != NULL, NULL);
503 return match_info->regex;
507 * g_match_info_get_string:
508 * @match_info: a #GMatchInfo
510 * Returns the string searched with @match_info. This is the
511 * string passed to g_regex_match() or g_regex_replace() so
512 * you may not free it before calling this function.
514 * Returns: the string searched with @match_info
519 g_match_info_get_string (const GMatchInfo *match_info)
521 g_return_val_if_fail (match_info != NULL, NULL);
522 return match_info->string;
527 * @match_info: a #GMatchInfo
529 * Increases reference count of @match_info by 1.
531 * Returns: @match_info
536 g_match_info_ref (GMatchInfo *match_info)
538 g_return_val_if_fail (match_info != NULL, NULL);
539 g_atomic_int_inc (&match_info->ref_count);
544 * g_match_info_unref:
545 * @match_info: a #GMatchInfo
547 * Decreases reference count of @match_info by 1. When reference count drops
548 * to zero, it frees all the memory associated with the match_info structure.
553 g_match_info_unref (GMatchInfo *match_info)
555 if (g_atomic_int_dec_and_test (&match_info->ref_count))
557 g_regex_unref (match_info->regex);
558 g_free (match_info->offsets);
559 g_free (match_info->workspace);
566 * @match_info: (allow-none): a #GMatchInfo, or %NULL
568 * If @match_info is not %NULL, calls g_match_info_unref(); otherwise does
574 g_match_info_free (GMatchInfo *match_info)
576 if (match_info == NULL)
579 g_match_info_unref (match_info);
584 * @match_info: a #GMatchInfo structure
585 * @error: location to store the error occurring, or %NULL to ignore errors
587 * Scans for the next match using the same parameters of the previous
588 * call to g_regex_match_full() or g_regex_match() that returned
591 * The match is done on the string passed to the match function, so you
592 * cannot free it before calling this function.
594 * Returns: %TRUE is the string matched, %FALSE otherwise
599 g_match_info_next (GMatchInfo *match_info,
602 gint prev_match_start;
605 g_return_val_if_fail (match_info != NULL, FALSE);
606 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
607 g_return_val_if_fail (match_info->pos >= 0, FALSE);
609 prev_match_start = match_info->offsets[0];
610 prev_match_end = match_info->offsets[1];
612 if (match_info->pos > match_info->string_len)
614 /* we have reached the end of the string */
615 match_info->pos = -1;
616 match_info->matches = PCRE_ERROR_NOMATCH;
620 match_info->matches = pcre_exec (match_info->regex->pcre_re,
621 match_info->regex->extra,
623 match_info->string_len,
625 match_info->regex->match_opts | match_info->match_opts,
627 match_info->n_offsets);
628 if (IS_PCRE_ERROR (match_info->matches))
630 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
631 _("Error while matching regular expression %s: %s"),
632 match_info->regex->pattern, match_error (match_info->matches));
636 /* avoid infinite loops if the pattern is an empty string or something
638 if (match_info->pos == match_info->offsets[1])
640 if (match_info->pos > match_info->string_len)
642 /* we have reached the end of the string */
643 match_info->pos = -1;
644 match_info->matches = PCRE_ERROR_NOMATCH;
648 match_info->pos = NEXT_CHAR (match_info->regex,
649 &match_info->string[match_info->pos]) -
654 match_info->pos = match_info->offsets[1];
657 /* it's possible to get two identical matches when we are matching
658 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
659 * the string is "RegExTest" we have:
660 * - search at position 0: match from 0 to 0
661 * - search at position 1: match from 3 to 3
662 * - search at position 3: match from 3 to 3 (duplicate)
663 * - search at position 4: match from 5 to 5
664 * - search at position 5: match from 5 to 5 (duplicate)
665 * - search at position 6: no match -> stop
666 * so we have to ignore the duplicates.
667 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
668 if (match_info->matches >= 0 &&
669 prev_match_start == match_info->offsets[0] &&
670 prev_match_end == match_info->offsets[1])
672 /* ignore this match and search the next one */
673 return g_match_info_next (match_info, error);
676 return match_info->matches >= 0;
680 * g_match_info_matches:
681 * @match_info: a #GMatchInfo structure
683 * Returns whether the previous match operation succeeded.
685 * Returns: %TRUE if the previous match operation succeeded,
691 g_match_info_matches (const GMatchInfo *match_info)
693 g_return_val_if_fail (match_info != NULL, FALSE);
695 return match_info->matches >= 0;
699 * g_match_info_get_match_count:
700 * @match_info: a #GMatchInfo structure
702 * Retrieves the number of matched substrings (including substring 0,
703 * that is the whole matched text), so 1 is returned if the pattern
704 * has no substrings in it and 0 is returned if the match failed.
706 * If the last match was obtained using the DFA algorithm, that is
707 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
708 * count is not that of the number of capturing parentheses but that of
709 * the number of matched substrings.
711 * Returns: Number of matched substrings, or -1 if an error occurred
716 g_match_info_get_match_count (const GMatchInfo *match_info)
718 g_return_val_if_fail (match_info, -1);
720 if (match_info->matches == PCRE_ERROR_NOMATCH)
723 else if (match_info->matches < PCRE_ERROR_NOMATCH)
728 return match_info->matches;
732 * g_match_info_is_partial_match:
733 * @match_info: a #GMatchInfo structure
735 * Usually if the string passed to g_regex_match*() matches as far as
736 * it goes, but is too short to match the entire pattern, %FALSE is
737 * returned. There are circumstances where it might be helpful to
738 * distinguish this case from other cases in which there is no match.
740 * Consider, for example, an application where a human is required to
741 * type in data for a field with specific formatting requirements. An
742 * example might be a date in the form ddmmmyy, defined by the pattern
743 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
744 * If the application sees the user’s keystrokes one by one, and can
745 * check that what has been typed so far is potentially valid, it is
746 * able to raise an error as soon as a mistake is made.
748 * GRegex supports the concept of partial matching by means of the
749 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
750 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
751 * for a complete match, %FALSE otherwise. But, when these functions
752 * return %FALSE, you can check if the match was partial calling
753 * g_match_info_is_partial_match().
755 * When using partial matching you cannot use g_match_info_fetch*().
757 * Because of the way certain internal optimizations are implemented
758 * the partial matching algorithm cannot be used with all patterns.
759 * So repeated single characters such as "a{2,4}" and repeated single
760 * meta-sequences such as "\d+" are not permitted if the maximum number
761 * of occurrences is greater than one. Optional items such as "\d?"
762 * (where the maximum is one) are permitted. Quantifiers with any values
763 * are permitted after parentheses, so the invalid examples above can be
764 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
765 * for a pattern that does not conform to the restrictions, matching
766 * functions return an error.
768 * Returns: %TRUE if the match was partial, %FALSE otherwise
773 g_match_info_is_partial_match (const GMatchInfo *match_info)
775 g_return_val_if_fail (match_info != NULL, FALSE);
777 return match_info->matches == PCRE_ERROR_PARTIAL;
781 * g_match_info_expand_references:
782 * @match_info: (allow-none): a #GMatchInfo or %NULL
783 * @string_to_expand: the string to expand
784 * @error: location to store the error occurring, or %NULL to ignore errors
786 * Returns a new string containing the text in @string_to_expand with
787 * references and escape sequences expanded. References refer to the last
788 * match done with @string against @regex and have the same syntax used by
791 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
792 * passed to g_regex_new().
794 * The backreferences are extracted from the string passed to the match
795 * function, so you cannot call this function after freeing the string.
797 * @match_info may be %NULL in which case @string_to_expand must not
798 * contain references. For instance "foo\n" does not refer to an actual
799 * pattern and '\n' merely will be replaced with \n character,
800 * while to expand "\0" (whole match) one needs the result of a match.
801 * Use g_regex_check_replacement() to find out whether @string_to_expand
802 * contains references.
804 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
809 g_match_info_expand_references (const GMatchInfo *match_info,
810 const gchar *string_to_expand,
815 GError *tmp_error = NULL;
817 g_return_val_if_fail (string_to_expand != NULL, NULL);
818 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
820 list = split_replacement (string_to_expand, &tmp_error);
821 if (tmp_error != NULL)
823 g_propagate_error (error, tmp_error);
827 if (!match_info && interpolation_list_needs_match (list))
829 g_critical ("String '%s' contains references to the match, can't "
830 "expand references without GMatchInfo object",
835 result = g_string_sized_new (strlen (string_to_expand));
836 interpolate_replacement (match_info, result, list);
838 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
840 return g_string_free (result, FALSE);
844 * g_match_info_fetch:
845 * @match_info: #GMatchInfo structure
846 * @match_num: number of the sub expression
848 * Retrieves the text matching the @match_num<!-- -->'th capturing
849 * parentheses. 0 is the full text of the match, 1 is the first paren
850 * 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 an empty
854 * string 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 * string 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 * The string is fetched from the string passed to the match function,
863 * so you cannot call this function after freeing the string.
865 * Returns: (allow-none): The matched substring, or %NULL if an error
866 * occurred. You have to free the string yourself
871 g_match_info_fetch (const GMatchInfo *match_info,
874 /* we cannot use pcre_get_substring() because it allocates the
875 * string using pcre_malloc(). */
879 g_return_val_if_fail (match_info != NULL, NULL);
880 g_return_val_if_fail (match_num >= 0, NULL);
882 /* match_num does not exist or it didn't matched, i.e. matching "b"
883 * against "(a)?b" then group 0 is empty. */
884 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
886 else if (start == -1)
887 match = g_strdup ("");
889 match = g_strndup (&match_info->string[start], end - start);
895 * g_match_info_fetch_pos:
896 * @match_info: #GMatchInfo structure
897 * @match_num: number of the sub expression
898 * @start_pos: (out) (allow-none): pointer to location where to store
899 * the start position, or %NULL
900 * @end_pos: (out) (allow-none): pointer to location where to store
901 * the end position, or %NULL
903 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
904 * parentheses. 0 is the full text of the match, 1 is the first
905 * paren set, 2 the second, and so on.
907 * If @match_num is a valid sub pattern but it didn't match anything
908 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
909 * and @end_pos are set to -1 and %TRUE is returned.
911 * If the match was obtained using the DFA algorithm, that is using
912 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
913 * position is not that of a set of parentheses but that of a matched
914 * substring. Substrings are matched in reverse order of length, so
915 * 0 is the longest match.
917 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
918 * the position cannot be fetched, @start_pos and @end_pos are left
924 g_match_info_fetch_pos (const GMatchInfo *match_info,
929 g_return_val_if_fail (match_info != NULL, FALSE);
930 g_return_val_if_fail (match_num >= 0, FALSE);
932 /* make sure the sub expression number they're requesting is less than
933 * the total number of sub expressions that were matched. */
934 if (match_num >= match_info->matches)
937 if (start_pos != NULL)
938 *start_pos = match_info->offsets[2 * match_num];
941 *end_pos = match_info->offsets[2 * match_num + 1];
947 * Returns number of first matched subpattern with name @name.
948 * There may be more than one in case when DUPNAMES is used,
949 * and not all subpatterns with that name match;
950 * pcre_get_stringnumber() does not work in that case.
953 get_matched_substring_number (const GMatchInfo *match_info,
960 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
961 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
963 /* This code is copied from pcre_get.c: get_first_set() */
964 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
972 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
974 gint n = (entry[0] << 8) + entry[1];
975 if (match_info->offsets[n*2] >= 0)
979 return (first[0] << 8) + first[1];
983 * g_match_info_fetch_named:
984 * @match_info: #GMatchInfo structure
985 * @name: name of the subexpression
987 * Retrieves the text matching the capturing parentheses named @name.
989 * If @name is a valid sub pattern name but it didn't match anything
990 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
991 * then an empty string is returned.
993 * The string is fetched from the string passed to the match function,
994 * so you cannot call this function after freeing the string.
996 * Returns: (allow-none): The matched substring, or %NULL if an error
997 * occurred. You have to free the string yourself
1002 g_match_info_fetch_named (const GMatchInfo *match_info,
1005 /* we cannot use pcre_get_named_substring() because it allocates the
1006 * string using pcre_malloc(). */
1009 g_return_val_if_fail (match_info != NULL, NULL);
1010 g_return_val_if_fail (name != NULL, NULL);
1012 num = get_matched_substring_number (match_info, name);
1016 return g_match_info_fetch (match_info, num);
1020 * g_match_info_fetch_named_pos:
1021 * @match_info: #GMatchInfo structure
1022 * @name: name of the subexpression
1023 * @start_pos: (out) (allow-none): pointer to location where to store
1024 * the start position, or %NULL
1025 * @end_pos: (out) (allow-none): pointer to location where to store
1026 * the end position, or %NULL
1028 * Retrieves the position in bytes of the capturing parentheses named @name.
1030 * If @name is a valid sub pattern name but it didn't match anything
1031 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1032 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
1034 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
1035 * If the position cannot be fetched, @start_pos and @end_pos
1036 * are left unchanged.
1041 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
1048 g_return_val_if_fail (match_info != NULL, FALSE);
1049 g_return_val_if_fail (name != NULL, FALSE);
1051 num = get_matched_substring_number (match_info, name);
1055 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1059 * g_match_info_fetch_all:
1060 * @match_info: a #GMatchInfo structure
1062 * Bundles up pointers to each of the matching substrings from a match
1063 * and stores them in an array of gchar pointers. The first element in
1064 * the returned array is the match number 0, i.e. the entire matched
1067 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1068 * "b" against "(a)?b") then an empty string is inserted.
1070 * If the last match was obtained using the DFA algorithm, that is using
1071 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1072 * strings are not that matched by sets of parentheses but that of the
1073 * matched substring. Substrings are matched in reverse order of length,
1074 * so the first one is the longest match.
1076 * The strings are fetched from the string passed to the match function,
1077 * so you cannot call this function after freeing the string.
1079 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1080 * It must be freed using g_strfreev(). If the previous match failed
1086 g_match_info_fetch_all (const GMatchInfo *match_info)
1088 /* we cannot use pcre_get_substring_list() because the returned value
1089 * isn't suitable for g_strfreev(). */
1093 g_return_val_if_fail (match_info != NULL, NULL);
1095 if (match_info->matches < 0)
1098 result = g_new (gchar *, match_info->matches + 1);
1099 for (i = 0; i < match_info->matches; i++)
1100 result[i] = g_match_info_fetch (match_info, i);
1110 g_regex_error_quark (void)
1112 static GQuark error_quark = 0;
1114 if (error_quark == 0)
1115 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1124 * Increases reference count of @regex by 1.
1131 g_regex_ref (GRegex *regex)
1133 g_return_val_if_fail (regex != NULL, NULL);
1134 g_atomic_int_inc (®ex->ref_count);
1142 * Decreases reference count of @regex by 1. When reference count drops
1143 * to zero, it frees all the memory associated with the regex structure.
1148 g_regex_unref (GRegex *regex)
1150 g_return_if_fail (regex != NULL);
1152 if (g_atomic_int_dec_and_test (®ex->ref_count))
1154 g_free (regex->pattern);
1155 if (regex->pcre_re != NULL)
1156 pcre_free (regex->pcre_re);
1157 if (regex->extra != NULL)
1158 pcre_free (regex->extra);
1165 * @pattern: the regular expression
1166 * @compile_options: compile options for the regular expression, or 0
1167 * @match_options: match options for the regular expression, or 0
1168 * @error: return location for a #GError
1170 * Compiles the regular expression to an internal form, and does
1171 * the initial setup of the #GRegex structure.
1173 * Returns: a #GRegex structure. Call g_regex_unref() when you
1179 g_regex_new (const gchar *pattern,
1180 GRegexCompileFlags compile_options,
1181 GRegexMatchFlags match_options,
1186 const gchar *errmsg;
1189 gboolean optimize = FALSE;
1190 static gsize initialised;
1191 unsigned long int pcre_compile_options;
1193 g_return_val_if_fail (pattern != NULL, NULL);
1194 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1195 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1196 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1198 if (g_once_init_enter (&initialised))
1203 pcre_config (PCRE_CONFIG_UTF8, &support);
1206 msg = N_("PCRE library is compiled without UTF8 support");
1207 g_critical ("%s", msg);
1208 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1212 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1215 msg = N_("PCRE library is compiled without UTF8 properties support");
1216 g_critical ("%s", msg);
1217 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1221 g_once_init_leave (&initialised, TRUE);
1224 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1225 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1226 if (compile_options & G_REGEX_OPTIMIZE)
1229 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1230 * instead uses UTF-8 only if required with PCRE_UTF8. */
1231 if (compile_options & G_REGEX_RAW)
1234 compile_options &= ~G_REGEX_RAW;
1239 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1240 match_options |= PCRE_NO_UTF8_CHECK;
1243 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1244 * not for the system one. */
1245 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1246 !(compile_options & G_REGEX_NEWLINE_LF))
1248 compile_options |= PCRE_NEWLINE_ANY;
1251 compile_options |= PCRE_UCP;
1253 /* compile the pattern */
1254 re = pcre_compile2 (pattern, compile_options, &errcode,
1255 &errmsg, &erroffset, NULL);
1257 /* if the compilation failed, set the error member and return
1263 /* Translate the PCRE error code to GRegexError and use a translated
1264 * error message if possible */
1265 translate_compile_error (&errcode, &errmsg);
1267 /* PCRE uses byte offsets but we want to show character offsets */
1268 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1270 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1271 _("Error while compiling regular "
1272 "expression %s at char %d: %s"),
1273 pattern, erroffset, errmsg);
1274 g_propagate_error (error, tmp_error);
1279 /* For options set at the beginning of the pattern, pcre puts them into
1280 * compile options, e.g. "(?i)foo" will make the pcre structure store
1281 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1282 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1283 compile_options = pcre_compile_options;
1285 if (!(compile_options & G_REGEX_DUPNAMES))
1287 gboolean jchanged = FALSE;
1288 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1290 compile_options |= G_REGEX_DUPNAMES;
1293 regex = g_new0 (GRegex, 1);
1294 regex->ref_count = 1;
1295 regex->pattern = g_strdup (pattern);
1296 regex->pcre_re = re;
1297 regex->compile_opts = compile_options;
1298 regex->match_opts = match_options;
1302 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1305 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1306 G_REGEX_ERROR_OPTIMIZE,
1307 _("Error while optimizing "
1308 "regular expression %s: %s"),
1311 g_propagate_error (error, tmp_error);
1313 g_regex_unref (regex);
1322 * g_regex_get_pattern:
1323 * @regex: a #GRegex structure
1325 * Gets the pattern string associated with @regex, i.e. a copy of
1326 * the string passed to g_regex_new().
1328 * Returns: the pattern of @regex
1333 g_regex_get_pattern (const GRegex *regex)
1335 g_return_val_if_fail (regex != NULL, NULL);
1337 return regex->pattern;
1341 * g_regex_get_max_backref:
1344 * Returns the number of the highest back reference
1345 * in the pattern, or 0 if the pattern does not contain
1348 * Returns: the number of the highest back reference
1353 g_regex_get_max_backref (const GRegex *regex)
1357 pcre_fullinfo (regex->pcre_re, regex->extra,
1358 PCRE_INFO_BACKREFMAX, &value);
1364 * g_regex_get_capture_count:
1367 * Returns the number of capturing subpatterns in the pattern.
1369 * Returns: the number of capturing subpatterns
1374 g_regex_get_capture_count (const GRegex *regex)
1378 pcre_fullinfo (regex->pcre_re, regex->extra,
1379 PCRE_INFO_CAPTURECOUNT, &value);
1385 * g_regex_get_compile_flags:
1388 * Returns the compile options that @regex was created with.
1390 * Returns: flags from #GRegexCompileFlags
1395 g_regex_get_compile_flags (const GRegex *regex)
1397 g_return_val_if_fail (regex != NULL, 0);
1399 return regex->compile_opts;
1403 * g_regex_get_match_flags:
1406 * Returns the match options that @regex was created with.
1408 * Returns: flags from #GRegexMatchFlags
1413 g_regex_get_match_flags (const GRegex *regex)
1415 g_return_val_if_fail (regex != NULL, 0);
1417 return regex->match_opts;
1421 * g_regex_match_simple:
1422 * @pattern: the regular expression
1423 * @string: the string to scan for matches
1424 * @compile_options: compile options for the regular expression, or 0
1425 * @match_options: match options, or 0
1427 * Scans for a match in @string for @pattern.
1429 * This function is equivalent to g_regex_match() but it does not
1430 * require to compile the pattern with g_regex_new(), avoiding some
1431 * lines of code when you need just to do a match without extracting
1432 * substrings, capture counts, and so on.
1434 * If this function is to be called on the same @pattern more than
1435 * once, it's more efficient to compile the pattern once with
1436 * g_regex_new() and then use g_regex_match().
1438 * Returns: %TRUE if the string matched, %FALSE otherwise
1443 g_regex_match_simple (const gchar *pattern,
1444 const gchar *string,
1445 GRegexCompileFlags compile_options,
1446 GRegexMatchFlags match_options)
1451 regex = g_regex_new (pattern, compile_options, 0, NULL);
1454 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1455 g_regex_unref (regex);
1461 * @regex: a #GRegex structure from g_regex_new()
1462 * @string: the string to scan for matches
1463 * @match_options: match options
1464 * @match_info: (out) (allow-none): pointer to location where to store
1465 * the #GMatchInfo, or %NULL if you do not need it
1467 * Scans for a match in string for the pattern in @regex.
1468 * The @match_options are combined with the match options specified
1469 * when the @regex structure was created, letting you have more
1470 * flexibility in reusing #GRegex structures.
1472 * A #GMatchInfo structure, used to get information on the match,
1473 * is stored in @match_info if not %NULL. Note that if @match_info
1474 * is not %NULL then it is created even if the function returns %FALSE,
1475 * i.e. you must free it regardless if regular expression actually matched.
1477 * To retrieve all the non-overlapping matches of the pattern in
1478 * string you can use g_match_info_next().
1482 * print_uppercase_words (const gchar *string)
1484 * /* Print all uppercase-only words. */
1486 * GMatchInfo *match_info;
1488 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1489 * g_regex_match (regex, string, 0, &match_info);
1490 * while (g_match_info_matches (match_info))
1492 * gchar *word = g_match_info_fetch (match_info, 0);
1493 * g_print ("Found: %s\n", word);
1495 * g_match_info_next (match_info, NULL);
1497 * g_match_info_free (match_info);
1498 * g_regex_unref (regex);
1502 * @string is not copied and is used in #GMatchInfo internally. If
1503 * you use any #GMatchInfo method (except g_match_info_free()) after
1504 * freeing or modifying @string then the behaviour is undefined.
1506 * Returns: %TRUE is the string matched, %FALSE otherwise
1511 g_regex_match (const GRegex *regex,
1512 const gchar *string,
1513 GRegexMatchFlags match_options,
1514 GMatchInfo **match_info)
1516 return g_regex_match_full (regex, string, -1, 0, match_options,
1521 * g_regex_match_full:
1522 * @regex: a #GRegex structure from g_regex_new()
1523 * @string: (array length=string_len): the string to scan for matches
1524 * @string_len: the length of @string, or -1 if @string is nul-terminated
1525 * @start_position: starting index of the string to match
1526 * @match_options: match options
1527 * @match_info: (out) (allow-none): pointer to location where to store
1528 * the #GMatchInfo, or %NULL if you do not need it
1529 * @error: location to store the error occurring, or %NULL to ignore errors
1531 * Scans for a match in string for the pattern in @regex.
1532 * The @match_options are combined with the match options specified
1533 * when the @regex structure was created, letting you have more
1534 * flexibility in reusing #GRegex structures.
1536 * Setting @start_position differs from just passing over a shortened
1537 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1538 * that begins with any kind of lookbehind assertion, such as "\b".
1540 * A #GMatchInfo structure, used to get information on the match, is
1541 * stored in @match_info if not %NULL. Note that if @match_info is
1542 * not %NULL then it is created even if the function returns %FALSE,
1543 * i.e. you must free it regardless if regular expression actually
1546 * @string is not copied and is used in #GMatchInfo internally. If
1547 * you use any #GMatchInfo method (except g_match_info_free()) after
1548 * freeing or modifying @string then the behaviour is undefined.
1550 * To retrieve all the non-overlapping matches of the pattern in
1551 * string you can use g_match_info_next().
1555 * print_uppercase_words (const gchar *string)
1557 * /* Print all uppercase-only words. */
1559 * GMatchInfo *match_info;
1560 * GError *error = NULL;
1562 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1563 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1564 * while (g_match_info_matches (match_info))
1566 * gchar *word = g_match_info_fetch (match_info, 0);
1567 * g_print ("Found: %s\n", word);
1569 * g_match_info_next (match_info, &error);
1571 * g_match_info_free (match_info);
1572 * g_regex_unref (regex);
1573 * if (error != NULL)
1575 * g_printerr ("Error while matching: %s\n", error->message);
1576 * g_error_free (error);
1581 * Returns: %TRUE is the string matched, %FALSE otherwise
1586 g_regex_match_full (const GRegex *regex,
1587 const gchar *string,
1589 gint start_position,
1590 GRegexMatchFlags match_options,
1591 GMatchInfo **match_info,
1597 g_return_val_if_fail (regex != NULL, FALSE);
1598 g_return_val_if_fail (string != NULL, FALSE);
1599 g_return_val_if_fail (start_position >= 0, FALSE);
1600 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1601 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1603 info = match_info_new (regex, string, string_len, start_position,
1604 match_options, FALSE);
1605 match_ok = g_match_info_next (info, error);
1606 if (match_info != NULL)
1609 g_match_info_free (info);
1615 * g_regex_match_all:
1616 * @regex: a #GRegex structure from g_regex_new()
1617 * @string: the string to scan for matches
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
1622 * Using the standard algorithm for regular expression matching only
1623 * the longest match in the string is retrieved. This function uses
1624 * a different algorithm so it can retrieve all the possible matches.
1625 * For more documentation see g_regex_match_all_full().
1627 * A #GMatchInfo structure, used to get information on the match, is
1628 * stored in @match_info if not %NULL. Note that if @match_info is
1629 * not %NULL then it is created even if the function returns %FALSE,
1630 * i.e. you must free it regardless if regular expression actually
1633 * @string is not copied and is used in #GMatchInfo internally. If
1634 * you use any #GMatchInfo method (except g_match_info_free()) after
1635 * freeing or modifying @string then the behaviour is undefined.
1637 * Returns: %TRUE is the string matched, %FALSE otherwise
1642 g_regex_match_all (const GRegex *regex,
1643 const gchar *string,
1644 GRegexMatchFlags match_options,
1645 GMatchInfo **match_info)
1647 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1652 * g_regex_match_all_full:
1653 * @regex: a #GRegex structure from g_regex_new()
1654 * @string: (array length=string_len): the string to scan for matches
1655 * @string_len: the length of @string, or -1 if @string is nul-terminated
1656 * @start_position: starting index of the string to match
1657 * @match_options: match options
1658 * @match_info: (out) (allow-none): pointer to location where to store
1659 * the #GMatchInfo, or %NULL if you do not need it
1660 * @error: location to store the error occurring, or %NULL to ignore errors
1662 * Using the standard algorithm for regular expression matching only
1663 * the longest match in the string is retrieved, it is not possible
1664 * to obtain all the available matches. For instance matching
1665 * "<a> <b> <c>" against the pattern "<.*>"
1666 * you get "<a> <b> <c>".
1668 * This function uses a different algorithm (called DFA, i.e. deterministic
1669 * finite automaton), so it can retrieve all the possible matches, all
1670 * starting at the same point in the string. For instance matching
1671 * "<a> <b> <c>" against the pattern "<.*>"
1672 * you would obtain three matches: "<a> <b> <c>",
1673 * "<a> <b>" and "<a>".
1675 * The number of matched strings is retrieved using
1676 * g_match_info_get_match_count(). To obtain the matched strings and
1677 * their position you can use, respectively, g_match_info_fetch() and
1678 * g_match_info_fetch_pos(). Note that the strings are returned in
1679 * reverse order of length; that is, the longest matching string is
1682 * Note that the DFA algorithm is slower than the standard one and it
1683 * is not able to capture substrings, so backreferences do not work.
1685 * Setting @start_position differs from just passing over a shortened
1686 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1687 * that begins with any kind of lookbehind assertion, such as "\b".
1689 * A #GMatchInfo structure, used to get information on the match, is
1690 * stored in @match_info if not %NULL. Note that if @match_info is
1691 * not %NULL then it is created even if the function returns %FALSE,
1692 * i.e. you must free it regardless if regular expression actually
1695 * @string is not copied and is used in #GMatchInfo internally. If
1696 * you use any #GMatchInfo method (except g_match_info_free()) after
1697 * freeing or modifying @string then the behaviour is undefined.
1699 * Returns: %TRUE is the string matched, %FALSE otherwise
1704 g_regex_match_all_full (const GRegex *regex,
1705 const gchar *string,
1707 gint start_position,
1708 GRegexMatchFlags match_options,
1709 GMatchInfo **match_info,
1715 g_return_val_if_fail (regex != NULL, FALSE);
1716 g_return_val_if_fail (string != NULL, FALSE);
1717 g_return_val_if_fail (start_position >= 0, FALSE);
1718 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1719 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1721 info = match_info_new (regex, string, string_len, start_position,
1722 match_options, TRUE);
1728 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1729 info->string, info->string_len,
1731 regex->match_opts | match_options,
1732 info->offsets, info->n_offsets,
1733 info->workspace, info->n_workspace);
1734 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1736 /* info->workspace is too small. */
1737 info->n_workspace *= 2;
1738 info->workspace = g_realloc (info->workspace,
1739 info->n_workspace * sizeof (gint));
1742 else if (info->matches == 0)
1744 /* info->offsets is too small. */
1745 info->n_offsets *= 2;
1746 info->offsets = g_realloc (info->offsets,
1747 info->n_offsets * sizeof (gint));
1750 else if (IS_PCRE_ERROR (info->matches))
1752 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1753 _("Error while matching regular expression %s: %s"),
1754 regex->pattern, match_error (info->matches));
1758 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1761 if (match_info != NULL)
1764 g_match_info_free (info);
1766 return info->matches >= 0;
1770 * g_regex_get_string_number:
1771 * @regex: #GRegex structure
1772 * @name: name of the subexpression
1774 * Retrieves the number of the subexpression named @name.
1776 * Returns: The number of the subexpression or -1 if @name
1782 g_regex_get_string_number (const GRegex *regex,
1787 g_return_val_if_fail (regex != NULL, -1);
1788 g_return_val_if_fail (name != NULL, -1);
1790 num = pcre_get_stringnumber (regex->pcre_re, name);
1791 if (num == PCRE_ERROR_NOSUBSTRING)
1798 * g_regex_split_simple:
1799 * @pattern: the regular expression
1800 * @string: the string to scan for matches
1801 * @compile_options: compile options for the regular expression, or 0
1802 * @match_options: match options, or 0
1804 * Breaks the string on the pattern, and returns an array of
1805 * the tokens. If the pattern contains capturing parentheses,
1806 * then the text for each of the substrings will also be returned.
1807 * If the pattern does not match anywhere in the string, then the
1808 * whole string is returned as the first token.
1810 * This function is equivalent to g_regex_split() but it does
1811 * not require to compile the pattern with g_regex_new(), avoiding
1812 * some lines of code when you need just to do a split without
1813 * extracting substrings, capture counts, and so on.
1815 * If this function is to be called on the same @pattern more than
1816 * once, it's more efficient to compile the pattern once with
1817 * g_regex_new() and then use g_regex_split().
1819 * As a special case, the result of splitting the empty string ""
1820 * is an empty vector, not a vector containing a single string.
1821 * The reason for this special case is that being able to represent
1822 * a empty vector is typically more useful than consistent handling
1823 * of empty elements. If you do need to represent empty elements,
1824 * you'll need to check for the empty string before calling this
1827 * A pattern that can match empty strings splits @string into
1828 * separate characters wherever it matches the empty string between
1829 * characters. For example splitting "ab c" using as a separator
1830 * "\s*", you will get "a", "b" and "c".
1832 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1837 g_regex_split_simple (const gchar *pattern,
1838 const gchar *string,
1839 GRegexCompileFlags compile_options,
1840 GRegexMatchFlags match_options)
1845 regex = g_regex_new (pattern, compile_options, 0, NULL);
1849 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1850 g_regex_unref (regex);
1856 * @regex: a #GRegex structure
1857 * @string: the string to split with the pattern
1858 * @match_options: match time option flags
1860 * Breaks the string on the pattern, and returns an array of the tokens.
1861 * If the pattern contains capturing parentheses, then the text for each
1862 * of the substrings will also be returned. If the pattern does not match
1863 * anywhere in the string, then the whole string is returned as the first
1866 * As a special case, the result of splitting the empty string "" is an
1867 * empty vector, not a vector containing a single string. The reason for
1868 * this special case is that being able to represent a empty vector is
1869 * typically more useful than consistent handling of empty elements. If
1870 * you do need to represent empty elements, you'll need to check for the
1871 * empty string before calling this function.
1873 * A pattern that can match empty strings splits @string into separate
1874 * characters wherever it matches the empty string between characters.
1875 * For example splitting "ab c" using as a separator "\s*", you will get
1878 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1883 g_regex_split (const GRegex *regex,
1884 const gchar *string,
1885 GRegexMatchFlags match_options)
1887 return g_regex_split_full (regex, string, -1, 0,
1888 match_options, 0, NULL);
1892 * g_regex_split_full:
1893 * @regex: a #GRegex structure
1894 * @string: (array length=string_len): the string to split with the pattern
1895 * @string_len: the length of @string, or -1 if @string is nul-terminated
1896 * @start_position: starting index of the string to match
1897 * @match_options: match time option flags
1898 * @max_tokens: the maximum number of tokens to split @string into.
1899 * If this is less than 1, the string is split completely
1900 * @error: return location for a #GError
1902 * Breaks the string on the pattern, and returns an array of the tokens.
1903 * If the pattern contains capturing parentheses, then the text for each
1904 * of the substrings will also be returned. If the pattern does not match
1905 * anywhere in the string, then the whole string is returned as the first
1908 * As a special case, the result of splitting the empty string "" is an
1909 * empty vector, not a vector containing a single string. The reason for
1910 * this special case is that being able to represent a empty vector is
1911 * typically more useful than consistent handling of empty elements. If
1912 * you do need to represent empty elements, you'll need to check for the
1913 * empty string before calling this function.
1915 * A pattern that can match empty strings splits @string into separate
1916 * characters wherever it matches the empty string between characters.
1917 * For example splitting "ab c" using as a separator "\s*", you will get
1920 * Setting @start_position differs from just passing over a shortened
1921 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1922 * that begins with any kind of lookbehind assertion, such as "\b".
1924 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1929 g_regex_split_full (const GRegex *regex,
1930 const gchar *string,
1932 gint start_position,
1933 GRegexMatchFlags match_options,
1937 GError *tmp_error = NULL;
1938 GMatchInfo *match_info;
1943 /* position of the last separator. */
1944 gint last_separator_end;
1945 /* was the last match 0 bytes long? */
1946 gboolean last_match_is_empty;
1947 /* the returned array of char **s */
1948 gchar **string_list;
1950 g_return_val_if_fail (regex != NULL, NULL);
1951 g_return_val_if_fail (string != NULL, NULL);
1952 g_return_val_if_fail (start_position >= 0, NULL);
1953 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1954 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1956 if (max_tokens <= 0)
1957 max_tokens = G_MAXINT;
1960 string_len = strlen (string);
1962 /* zero-length string */
1963 if (string_len - start_position == 0)
1964 return g_new0 (gchar *, 1);
1966 if (max_tokens == 1)
1968 string_list = g_new0 (gchar *, 2);
1969 string_list[0] = g_strndup (&string[start_position],
1970 string_len - start_position);
1976 last_separator_end = start_position;
1977 last_match_is_empty = FALSE;
1979 match_ok = g_regex_match_full (regex, string, string_len, start_position,
1980 match_options, &match_info, &tmp_error);
1982 while (tmp_error == NULL)
1986 last_match_is_empty =
1987 (match_info->offsets[0] == match_info->offsets[1]);
1989 /* we need to skip empty separators at the same position of the end
1990 * of another separator. e.g. the string is "a b" and the separator
1991 * is " *", so from 1 to 2 we have a match and at position 2 we have
1992 * an empty match. */
1993 if (last_separator_end != match_info->offsets[1])
1998 token = g_strndup (string + last_separator_end,
1999 match_info->offsets[0] - last_separator_end);
2000 list = g_list_prepend (list, token);
2003 /* if there were substrings, these need to be added to
2005 match_count = g_match_info_get_match_count (match_info);
2006 if (match_count > 1)
2008 for (i = 1; i < match_count; i++)
2009 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
2015 /* if there was no match, copy to end of string. */
2016 if (!last_match_is_empty)
2018 gchar *token = g_strndup (string + last_separator_end,
2019 match_info->string_len - last_separator_end);
2020 list = g_list_prepend (list, token);
2022 /* no more tokens, end the loop. */
2026 /* -1 to leave room for the last part. */
2027 if (token_count >= max_tokens - 1)
2029 /* we have reached the maximum number of tokens, so we copy
2030 * the remaining part of the string. */
2031 if (last_match_is_empty)
2033 /* the last match was empty, so we have moved one char
2034 * after the real position to avoid empty matches at the
2036 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
2038 /* the if is needed in the case we have terminated the available
2039 * tokens, but we are at the end of the string, so there are no
2040 * characters left to copy. */
2041 if (string_len > match_info->pos)
2043 gchar *token = g_strndup (string + match_info->pos,
2044 string_len - match_info->pos);
2045 list = g_list_prepend (list, token);
2051 last_separator_end = match_info->pos;
2052 if (last_match_is_empty)
2053 /* if the last match was empty, g_match_info_next() has moved
2054 * forward to avoid infinite loops, but we still need to copy that
2056 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
2058 match_ok = g_match_info_next (match_info, &tmp_error);
2060 g_match_info_free (match_info);
2061 if (tmp_error != NULL)
2063 g_propagate_error (error, tmp_error);
2064 g_list_free_full (list, g_free);
2065 match_info->pos = -1;
2069 string_list = g_new (gchar *, g_list_length (list) + 1);
2071 for (last = g_list_last (list); last; last = g_list_previous (last))
2072 string_list[i++] = last->data;
2073 string_list[i] = NULL;
2082 REPL_TYPE_CHARACTER,
2083 REPL_TYPE_SYMBOLIC_REFERENCE,
2084 REPL_TYPE_NUMERIC_REFERENCE,
2085 REPL_TYPE_CHANGE_CASE
2090 CHANGE_CASE_NONE = 1 << 0,
2091 CHANGE_CASE_UPPER = 1 << 1,
2092 CHANGE_CASE_LOWER = 1 << 2,
2093 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2094 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2095 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2096 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2097 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2100 struct _InterpolationData
2106 ChangeCase change_case;
2110 free_interpolation_data (InterpolationData *data)
2112 g_free (data->text);
2116 static const gchar *
2117 expand_escape (const gchar *replacement,
2119 InterpolationData *data,
2124 const gchar *error_detail;
2126 GError *tmp_error = NULL;
2134 data->type = REPL_TYPE_CHARACTER;
2139 data->type = REPL_TYPE_CHARACTER;
2144 data->type = REPL_TYPE_CHARACTER;
2149 data->type = REPL_TYPE_CHARACTER;
2154 data->type = REPL_TYPE_CHARACTER;
2159 data->type = REPL_TYPE_CHARACTER;
2164 data->type = REPL_TYPE_CHARACTER;
2169 data->type = REPL_TYPE_CHARACTER;
2179 h = g_ascii_xdigit_value (*p);
2182 error_detail = _("hexadecimal digit or '}' expected");
2193 for (i = 0; i < 2; i++)
2195 h = g_ascii_xdigit_value (*p);
2198 error_detail = _("hexadecimal digit expected");
2205 data->type = REPL_TYPE_STRING;
2206 data->text = g_new0 (gchar, 8);
2207 g_unichar_to_utf8 (x, data->text);
2211 data->type = REPL_TYPE_CHANGE_CASE;
2212 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2216 data->type = REPL_TYPE_CHANGE_CASE;
2217 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2221 data->type = REPL_TYPE_CHANGE_CASE;
2222 data->change_case = CHANGE_CASE_LOWER;
2226 data->type = REPL_TYPE_CHANGE_CASE;
2227 data->change_case = CHANGE_CASE_UPPER;
2231 data->type = REPL_TYPE_CHANGE_CASE;
2232 data->change_case = CHANGE_CASE_NONE;
2238 error_detail = _("missing '<' in symbolic reference");
2247 error_detail = _("unfinished symbolic reference");
2254 error_detail = _("zero-length symbolic reference");
2257 if (g_ascii_isdigit (*q))
2262 h = g_ascii_digit_value (*q);
2265 error_detail = _("digit expected");
2274 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2281 if (!g_ascii_isalnum (*r))
2283 error_detail = _("illegal symbolic reference");
2290 data->text = g_strndup (q, p - q);
2291 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2296 /* if \0 is followed by a number is an octal number representing a
2297 * character, else it is a numeric reference. */
2298 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2301 p = g_utf8_next_char (p);
2314 for (i = 0; i < 3; i++)
2316 h = g_ascii_digit_value (*p);
2326 if (i == 2 && base == 10)
2332 if (base == 8 || i == 3)
2334 data->type = REPL_TYPE_STRING;
2335 data->text = g_new0 (gchar, 8);
2336 g_unichar_to_utf8 (x, data->text);
2340 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2345 error_detail = _("stray final '\\'");
2349 error_detail = _("unknown escape sequence");
2356 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2357 tmp_error = g_error_new (G_REGEX_ERROR,
2358 G_REGEX_ERROR_REPLACE,
2359 _("Error while parsing replacement "
2360 "text \"%s\" at char %lu: %s"),
2362 (gulong)(p - replacement),
2364 g_propagate_error (error, tmp_error);
2370 split_replacement (const gchar *replacement,
2374 InterpolationData *data;
2375 const gchar *p, *start;
2377 start = p = replacement;
2382 data = g_new0 (InterpolationData, 1);
2383 start = p = expand_escape (replacement, p, data, error);
2386 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2387 free_interpolation_data (data);
2391 list = g_list_prepend (list, data);
2396 if (*p == '\\' || *p == '\0')
2400 data = g_new0 (InterpolationData, 1);
2401 data->text = g_strndup (start, p - start);
2402 data->type = REPL_TYPE_STRING;
2403 list = g_list_prepend (list, data);
2409 return g_list_reverse (list);
2412 /* Change the case of c based on change_case. */
2413 #define CHANGE_CASE(c, change_case) \
2414 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2415 g_unichar_tolower (c) : \
2416 g_unichar_toupper (c))
2419 string_append (GString *string,
2421 ChangeCase *change_case)
2425 if (text[0] == '\0')
2428 if (*change_case == CHANGE_CASE_NONE)
2430 g_string_append (string, text);
2432 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2434 c = g_utf8_get_char (text);
2435 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2436 g_string_append (string, g_utf8_next_char (text));
2437 *change_case = CHANGE_CASE_NONE;
2441 while (*text != '\0')
2443 c = g_utf8_get_char (text);
2444 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2445 text = g_utf8_next_char (text);
2451 interpolate_replacement (const GMatchInfo *match_info,
2456 InterpolationData *idata;
2458 ChangeCase change_case = CHANGE_CASE_NONE;
2460 for (list = data; list; list = list->next)
2463 switch (idata->type)
2465 case REPL_TYPE_STRING:
2466 string_append (result, idata->text, &change_case);
2468 case REPL_TYPE_CHARACTER:
2469 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2470 if (change_case & CHANGE_CASE_SINGLE_MASK)
2471 change_case = CHANGE_CASE_NONE;
2473 case REPL_TYPE_NUMERIC_REFERENCE:
2474 match = g_match_info_fetch (match_info, idata->num);
2477 string_append (result, match, &change_case);
2481 case REPL_TYPE_SYMBOLIC_REFERENCE:
2482 match = g_match_info_fetch_named (match_info, idata->text);
2485 string_append (result, match, &change_case);
2489 case REPL_TYPE_CHANGE_CASE:
2490 change_case = idata->change_case;
2498 /* whether actual match_info is needed for replacement, i.e.
2499 * whether there are references
2502 interpolation_list_needs_match (GList *list)
2504 while (list != NULL)
2506 InterpolationData *data = list->data;
2508 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2509 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2522 * @regex: a #GRegex structure
2523 * @string: (array length=string_len): the string to perform matches against
2524 * @string_len: the length of @string, or -1 if @string is nul-terminated
2525 * @start_position: starting index of the string to match
2526 * @replacement: text to replace each match with
2527 * @match_options: options for the match
2528 * @error: location to store the error occurring, or %NULL to ignore errors
2530 * Replaces all occurrences of the pattern in @regex with the
2531 * replacement text. Backreferences of the form '\number' or
2532 * '\g<number>' in the replacement text are interpolated by the
2533 * number-th captured subexpression of the match, '\g<name>' refers
2534 * to the captured subexpression with the given name. '\0' refers to the
2535 * complete match, but '\0' followed by a number is the octal representation
2536 * of a character. To include a literal '\' in the replacement, write '\\'.
2537 * There are also escapes that changes the case of the following text:
2540 * <varlistentry><term>\l</term>
2542 * <para>Convert to lower case the next character</para>
2545 * <varlistentry><term>\u</term>
2547 * <para>Convert to upper case the next character</para>
2550 * <varlistentry><term>\L</term>
2552 * <para>Convert to lower case till \E</para>
2555 * <varlistentry><term>\U</term>
2557 * <para>Convert to upper case till \E</para>
2560 * <varlistentry><term>\E</term>
2562 * <para>End case modification</para>
2567 * If you do not need to use backreferences use g_regex_replace_literal().
2569 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2570 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2571 * you can use g_regex_replace_literal().
2573 * Setting @start_position differs from just passing over a shortened
2574 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2575 * begins with any kind of lookbehind assertion, such as "\b".
2577 * Returns: a newly allocated string containing the replacements
2582 g_regex_replace (const GRegex *regex,
2583 const gchar *string,
2585 gint start_position,
2586 const gchar *replacement,
2587 GRegexMatchFlags match_options,
2592 GError *tmp_error = NULL;
2594 g_return_val_if_fail (regex != NULL, NULL);
2595 g_return_val_if_fail (string != NULL, NULL);
2596 g_return_val_if_fail (start_position >= 0, NULL);
2597 g_return_val_if_fail (replacement != NULL, NULL);
2598 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2599 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2601 list = split_replacement (replacement, &tmp_error);
2602 if (tmp_error != NULL)
2604 g_propagate_error (error, tmp_error);
2608 result = g_regex_replace_eval (regex,
2609 string, string_len, start_position,
2611 interpolate_replacement,
2614 if (tmp_error != NULL)
2615 g_propagate_error (error, tmp_error);
2617 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2623 literal_replacement (const GMatchInfo *match_info,
2627 g_string_append (result, data);
2632 * g_regex_replace_literal:
2633 * @regex: a #GRegex structure
2634 * @string: (array length=string_len): the string to perform matches against
2635 * @string_len: the length of @string, or -1 if @string is nul-terminated
2636 * @start_position: starting index of the string to match
2637 * @replacement: text to replace each match with
2638 * @match_options: options for the match
2639 * @error: location to store the error occurring, or %NULL to ignore errors
2641 * Replaces all occurrences of the pattern in @regex with the
2642 * replacement text. @replacement is replaced literally, to
2643 * include backreferences use g_regex_replace().
2645 * Setting @start_position differs from just passing over a
2646 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2647 * case of a pattern that begins with any kind of lookbehind
2648 * assertion, such as "\b".
2650 * Returns: a newly allocated string containing the replacements
2655 g_regex_replace_literal (const GRegex *regex,
2656 const gchar *string,
2658 gint start_position,
2659 const gchar *replacement,
2660 GRegexMatchFlags match_options,
2663 g_return_val_if_fail (replacement != NULL, NULL);
2664 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2666 return g_regex_replace_eval (regex,
2667 string, string_len, start_position,
2669 literal_replacement,
2670 (gpointer)replacement,
2675 * g_regex_replace_eval:
2676 * @regex: a #GRegex structure from g_regex_new()
2677 * @string: (array length=string_len): string to perform matches against
2678 * @string_len: the length of @string, or -1 if @string is nul-terminated
2679 * @start_position: starting index of the string to match
2680 * @match_options: options for the match
2681 * @eval: a function to call for each match
2682 * @user_data: user data to pass to the function
2683 * @error: location to store the error occurring, or %NULL to ignore errors
2685 * Replaces occurrences of the pattern in regex with the output of
2686 * @eval for that occurrence.
2688 * Setting @start_position differs from just passing over a shortened
2689 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2690 * that begins with any kind of lookbehind assertion, such as "\b".
2692 * The following example uses g_regex_replace_eval() to replace multiple
2696 * eval_cb (const GMatchInfo *info,
2703 * match = g_match_info_fetch (info, 0);
2704 * r = g_hash_table_lookup ((GHashTable *)data, match);
2705 * g_string_append (res, r);
2717 * h = g_hash_table_new (g_str_hash, g_str_equal);
2719 * g_hash_table_insert (h, "1", "ONE");
2720 * g_hash_table_insert (h, "2", "TWO");
2721 * g_hash_table_insert (h, "3", "THREE");
2722 * g_hash_table_insert (h, "4", "FOUR");
2724 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2725 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2726 * g_hash_table_destroy (h);
2731 * Returns: a newly allocated string containing the replacements
2736 g_regex_replace_eval (const GRegex *regex,
2737 const gchar *string,
2739 gint start_position,
2740 GRegexMatchFlags match_options,
2741 GRegexEvalCallback eval,
2745 GMatchInfo *match_info;
2748 gboolean done = FALSE;
2749 GError *tmp_error = NULL;
2751 g_return_val_if_fail (regex != NULL, NULL);
2752 g_return_val_if_fail (string != NULL, NULL);
2753 g_return_val_if_fail (start_position >= 0, NULL);
2754 g_return_val_if_fail (eval != NULL, NULL);
2755 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2758 string_len = strlen (string);
2760 result = g_string_sized_new (string_len);
2762 /* run down the string making matches. */
2763 g_regex_match_full (regex, string, string_len, start_position,
2764 match_options, &match_info, &tmp_error);
2765 while (!done && g_match_info_matches (match_info))
2767 g_string_append_len (result,
2769 match_info->offsets[0] - str_pos);
2770 done = (*eval) (match_info, result, user_data);
2771 str_pos = match_info->offsets[1];
2772 g_match_info_next (match_info, &tmp_error);
2774 g_match_info_free (match_info);
2775 if (tmp_error != NULL)
2777 g_propagate_error (error, tmp_error);
2778 g_string_free (result, TRUE);
2782 g_string_append_len (result, string + str_pos, string_len - str_pos);
2783 return g_string_free (result, FALSE);
2787 * g_regex_check_replacement:
2788 * @replacement: the replacement string
2789 * @has_references: (out) (allow-none): location to store information about
2790 * references in @replacement or %NULL
2791 * @error: location to store error
2793 * Checks whether @replacement is a valid replacement string
2794 * (see g_regex_replace()), i.e. that all escape sequences in
2797 * If @has_references is not %NULL then @replacement is checked
2798 * for pattern references. For instance, replacement text 'foo\n'
2799 * does not contain references and may be evaluated without information
2800 * about actual match, but '\0\1' (whole match followed by first
2801 * subpattern) requires valid #GMatchInfo object.
2803 * Returns: whether @replacement is a valid replacement string
2808 g_regex_check_replacement (const gchar *replacement,
2809 gboolean *has_references,
2815 list = split_replacement (replacement, &tmp);
2819 g_propagate_error (error, tmp);
2824 *has_references = interpolation_list_needs_match (list);
2826 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2832 * g_regex_escape_nul:
2833 * @string: the string to escape
2834 * @length: the length of @string
2836 * Escapes the nul characters in @string to "\x00". It can be used
2837 * to compile a regex with embedded nul characters.
2839 * For completeness, @length can be -1 for a nul-terminated string.
2840 * In this case the output string will be of course equal to @string.
2842 * Returns: a newly-allocated escaped string
2847 g_regex_escape_nul (const gchar *string,
2851 const gchar *p, *piece_start, *end;
2854 g_return_val_if_fail (string != NULL, NULL);
2857 return g_strdup (string);
2859 end = string + length;
2860 p = piece_start = string;
2861 escaped = g_string_sized_new (length + 1);
2869 if (p != piece_start)
2871 /* copy the previous piece. */
2872 g_string_append_len (escaped, piece_start, p - piece_start);
2874 if ((backslashes & 1) == 0)
2875 g_string_append_c (escaped, '\\');
2876 g_string_append_c (escaped, 'x');
2877 g_string_append_c (escaped, '0');
2878 g_string_append_c (escaped, '0');
2888 p = g_utf8_next_char (p);
2893 if (piece_start < end)
2894 g_string_append_len (escaped, piece_start, end - piece_start);
2896 return g_string_free (escaped, FALSE);
2900 * g_regex_escape_string:
2901 * @string: (array length=length): the string to escape
2902 * @length: the length of @string, or -1 if @string is nul-terminated
2904 * Escapes the special characters used for regular expressions
2905 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2906 * function is useful to dynamically generate regular expressions.
2908 * @string can contain nul characters that are replaced with "\0",
2909 * in this case remember to specify the correct length of @string
2912 * Returns: a newly-allocated escaped string
2917 g_regex_escape_string (const gchar *string,
2921 const char *p, *piece_start, *end;
2923 g_return_val_if_fail (string != NULL, NULL);
2926 length = strlen (string);
2928 end = string + length;
2929 p = piece_start = string;
2930 escaped = g_string_sized_new (length + 1);
2951 if (p != piece_start)
2952 /* copy the previous piece. */
2953 g_string_append_len (escaped, piece_start, p - piece_start);
2954 g_string_append_c (escaped, '\\');
2956 g_string_append_c (escaped, '0');
2958 g_string_append_c (escaped, *p);
2962 p = g_utf8_next_char (p);
2967 if (piece_start < end)
2968 g_string_append_len (escaped, piece_start, end - piece_start);
2970 return g_string_free (escaped, FALSE);