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_INCONSISTENT_NEWLINE_OPTIONS:
390 *errmsg = _("inconsistent NEWLINE options");
392 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
393 *errmsg = _("\\g is not followed by a braced name or an optionally "
394 "braced non-zero number");
397 *errcode = G_REGEX_ERROR_INTERNAL;
398 *errmsg = _("unexpected repeat");
401 *errcode = G_REGEX_ERROR_INTERNAL;
402 *errmsg = _("code overflow");
405 *errcode = G_REGEX_ERROR_INTERNAL;
406 *errmsg = _("overran compiling workspace");
409 *errcode = G_REGEX_ERROR_INTERNAL;
410 *errmsg = _("previously-checked referenced subpattern not found");
413 /* This should not happen as we never pass a NULL erroffset */
414 g_warning ("erroffset passed as NULL");
415 *errcode = G_REGEX_ERROR_COMPILE;
418 /* This should not happen as we check options before passing them
419 * to pcre_compile2() */
420 g_warning ("unknown option bit(s) set");
421 *errcode = G_REGEX_ERROR_COMPILE;
426 /* These errors should not happen as we are using an UTF8-enabled PCRE
427 * and we do not check if strings are valid */
428 g_warning ("%s", *errmsg);
429 *errcode = G_REGEX_ERROR_COMPILE;
432 *errcode = G_REGEX_ERROR_COMPILE;
439 match_info_new (const GRegex *regex,
446 GMatchInfo *match_info;
449 string_len = strlen (string);
451 match_info = g_new0 (GMatchInfo, 1);
452 match_info->ref_count = 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 * Increases reference count of @match_info by 1.
528 * Returns: @match_info
533 g_match_info_ref (GMatchInfo *match_info)
535 g_return_val_if_fail (match_info != NULL, NULL);
536 g_atomic_int_inc (&match_info->ref_count);
541 * g_match_info_unref:
542 * @match_info: a #GMatchInfo
544 * Decreases reference count of @match_info by 1. When reference count drops
545 * to zero, it frees all the memory associated with the match_info structure.
550 g_match_info_unref (GMatchInfo *match_info)
552 if (g_atomic_int_dec_and_test (&match_info->ref_count))
554 g_regex_unref (match_info->regex);
555 g_free (match_info->offsets);
556 g_free (match_info->workspace);
563 * @match_info: (allow-none): a #GMatchInfo, or %NULL
565 * If @match_info is not %NULL, calls g_match_info_unref(); otherwise does
571 g_match_info_free (GMatchInfo *match_info)
573 if (match_info == NULL)
576 g_match_info_unref (match_info);
581 * @match_info: a #GMatchInfo structure
582 * @error: location to store the error occurring, or %NULL to ignore errors
584 * Scans for the next match using the same parameters of the previous
585 * call to g_regex_match_full() or g_regex_match() that returned
588 * The match is done on the string passed to the match function, so you
589 * cannot free it before calling this function.
591 * Returns: %TRUE is the string matched, %FALSE otherwise
596 g_match_info_next (GMatchInfo *match_info,
599 gint prev_match_start;
602 g_return_val_if_fail (match_info != NULL, FALSE);
603 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
604 g_return_val_if_fail (match_info->pos >= 0, FALSE);
606 prev_match_start = match_info->offsets[0];
607 prev_match_end = match_info->offsets[1];
609 if (match_info->pos > match_info->string_len)
611 /* we have reached the end of the string */
612 match_info->pos = -1;
613 match_info->matches = PCRE_ERROR_NOMATCH;
617 match_info->matches = pcre_exec (match_info->regex->pcre_re,
618 match_info->regex->extra,
620 match_info->string_len,
622 match_info->regex->match_opts | match_info->match_opts,
624 match_info->n_offsets);
625 if (IS_PCRE_ERROR (match_info->matches))
627 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
628 _("Error while matching regular expression %s: %s"),
629 match_info->regex->pattern, match_error (match_info->matches));
633 /* avoid infinite loops if the pattern is an empty string or something
635 if (match_info->pos == match_info->offsets[1])
637 if (match_info->pos > match_info->string_len)
639 /* we have reached the end of the string */
640 match_info->pos = -1;
641 match_info->matches = PCRE_ERROR_NOMATCH;
645 match_info->pos = NEXT_CHAR (match_info->regex,
646 &match_info->string[match_info->pos]) -
651 match_info->pos = match_info->offsets[1];
654 /* it's possible to get two identical matches when we are matching
655 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
656 * the string is "RegExTest" we have:
657 * - search at position 0: match from 0 to 0
658 * - search at position 1: match from 3 to 3
659 * - search at position 3: match from 3 to 3 (duplicate)
660 * - search at position 4: match from 5 to 5
661 * - search at position 5: match from 5 to 5 (duplicate)
662 * - search at position 6: no match -> stop
663 * so we have to ignore the duplicates.
664 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
665 if (match_info->matches >= 0 &&
666 prev_match_start == match_info->offsets[0] &&
667 prev_match_end == match_info->offsets[1])
669 /* ignore this match and search the next one */
670 return g_match_info_next (match_info, error);
673 return match_info->matches >= 0;
677 * g_match_info_matches:
678 * @match_info: a #GMatchInfo structure
680 * Returns whether the previous match operation succeeded.
682 * Returns: %TRUE if the previous match operation succeeded,
688 g_match_info_matches (const GMatchInfo *match_info)
690 g_return_val_if_fail (match_info != NULL, FALSE);
692 return match_info->matches >= 0;
696 * g_match_info_get_match_count:
697 * @match_info: a #GMatchInfo structure
699 * Retrieves the number of matched substrings (including substring 0,
700 * that is the whole matched text), so 1 is returned if the pattern
701 * has no substrings in it and 0 is returned if the match failed.
703 * If the last match was obtained using the DFA algorithm, that is
704 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
705 * count is not that of the number of capturing parentheses but that of
706 * the number of matched substrings.
708 * Returns: Number of matched substrings, or -1 if an error occurred
713 g_match_info_get_match_count (const GMatchInfo *match_info)
715 g_return_val_if_fail (match_info, -1);
717 if (match_info->matches == PCRE_ERROR_NOMATCH)
720 else if (match_info->matches < PCRE_ERROR_NOMATCH)
725 return match_info->matches;
729 * g_match_info_is_partial_match:
730 * @match_info: a #GMatchInfo structure
732 * Usually if the string passed to g_regex_match*() matches as far as
733 * it goes, but is too short to match the entire pattern, %FALSE is
734 * returned. There are circumstances where it might be helpful to
735 * distinguish this case from other cases in which there is no match.
737 * Consider, for example, an application where a human is required to
738 * type in data for a field with specific formatting requirements. An
739 * example might be a date in the form ddmmmyy, defined by the pattern
740 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
741 * If the application sees the user’s keystrokes one by one, and can
742 * check that what has been typed so far is potentially valid, it is
743 * able to raise an error as soon as a mistake is made.
745 * GRegex supports the concept of partial matching by means of the
746 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
747 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
748 * for a complete match, %FALSE otherwise. But, when these functions
749 * return %FALSE, you can check if the match was partial calling
750 * g_match_info_is_partial_match().
752 * When using partial matching you cannot use g_match_info_fetch*().
754 * Because of the way certain internal optimizations are implemented
755 * the partial matching algorithm cannot be used with all patterns.
756 * So repeated single characters such as "a{2,4}" and repeated single
757 * meta-sequences such as "\d+" are not permitted if the maximum number
758 * of occurrences is greater than one. Optional items such as "\d?"
759 * (where the maximum is one) are permitted. Quantifiers with any values
760 * are permitted after parentheses, so the invalid examples above can be
761 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
762 * for a pattern that does not conform to the restrictions, matching
763 * functions return an error.
765 * Returns: %TRUE if the match was partial, %FALSE otherwise
770 g_match_info_is_partial_match (const GMatchInfo *match_info)
772 g_return_val_if_fail (match_info != NULL, FALSE);
774 return match_info->matches == PCRE_ERROR_PARTIAL;
778 * g_match_info_expand_references:
779 * @match_info: (allow-none): a #GMatchInfo or %NULL
780 * @string_to_expand: the string to expand
781 * @error: location to store the error occurring, or %NULL to ignore errors
783 * Returns a new string containing the text in @string_to_expand with
784 * references and escape sequences expanded. References refer to the last
785 * match done with @string against @regex and have the same syntax used by
788 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
789 * passed to g_regex_new().
791 * The backreferences are extracted from the string passed to the match
792 * function, so you cannot call this function after freeing the string.
794 * @match_info may be %NULL in which case @string_to_expand must not
795 * contain references. For instance "foo\n" does not refer to an actual
796 * pattern and '\n' merely will be replaced with \n character,
797 * while to expand "\0" (whole match) one needs the result of a match.
798 * Use g_regex_check_replacement() to find out whether @string_to_expand
799 * contains references.
801 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
806 g_match_info_expand_references (const GMatchInfo *match_info,
807 const gchar *string_to_expand,
812 GError *tmp_error = NULL;
814 g_return_val_if_fail (string_to_expand != NULL, NULL);
815 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
817 list = split_replacement (string_to_expand, &tmp_error);
818 if (tmp_error != NULL)
820 g_propagate_error (error, tmp_error);
824 if (!match_info && interpolation_list_needs_match (list))
826 g_critical ("String '%s' contains references to the match, can't "
827 "expand references without GMatchInfo object",
832 result = g_string_sized_new (strlen (string_to_expand));
833 interpolate_replacement (match_info, result, list);
835 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
837 return g_string_free (result, FALSE);
841 * g_match_info_fetch:
842 * @match_info: #GMatchInfo structure
843 * @match_num: number of the sub expression
845 * Retrieves the text matching the @match_num<!-- -->'th capturing
846 * parentheses. 0 is the full text of the match, 1 is the first paren
847 * set, 2 the second, and so on.
849 * If @match_num is a valid sub pattern but it didn't match anything
850 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
851 * string is returned.
853 * If the match was obtained using the DFA algorithm, that is using
854 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
855 * string is not that of a set of parentheses but that of a matched
856 * substring. Substrings are matched in reverse order of length, so
857 * 0 is the longest match.
859 * The string is fetched from the string passed to the match function,
860 * so you cannot call this function after freeing the string.
862 * Returns: (allow-none): The matched substring, or %NULL if an error
863 * occurred. You have to free the string yourself
868 g_match_info_fetch (const GMatchInfo *match_info,
871 /* we cannot use pcre_get_substring() because it allocates the
872 * string using pcre_malloc(). */
876 g_return_val_if_fail (match_info != NULL, NULL);
877 g_return_val_if_fail (match_num >= 0, NULL);
879 /* match_num does not exist or it didn't matched, i.e. matching "b"
880 * against "(a)?b" then group 0 is empty. */
881 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
883 else if (start == -1)
884 match = g_strdup ("");
886 match = g_strndup (&match_info->string[start], end - start);
892 * g_match_info_fetch_pos:
893 * @match_info: #GMatchInfo structure
894 * @match_num: number of the sub expression
895 * @start_pos: (out) (allow-none): pointer to location where to store
896 * the start position, or %NULL
897 * @end_pos: (out) (allow-none): pointer to location where to store
898 * the end position, or %NULL
900 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
901 * parentheses. 0 is the full text of the match, 1 is the first
902 * paren set, 2 the second, and so on.
904 * If @match_num is a valid sub pattern but it didn't match anything
905 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
906 * and @end_pos are set to -1 and %TRUE is returned.
908 * If the match was obtained using the DFA algorithm, that is using
909 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
910 * position is not that of a set of parentheses but that of a matched
911 * substring. Substrings are matched in reverse order of length, so
912 * 0 is the longest match.
914 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
915 * the position cannot be fetched, @start_pos and @end_pos are left
921 g_match_info_fetch_pos (const GMatchInfo *match_info,
926 g_return_val_if_fail (match_info != NULL, FALSE);
927 g_return_val_if_fail (match_num >= 0, FALSE);
929 /* make sure the sub expression number they're requesting is less than
930 * the total number of sub expressions that were matched. */
931 if (match_num >= match_info->matches)
934 if (start_pos != NULL)
935 *start_pos = match_info->offsets[2 * match_num];
938 *end_pos = match_info->offsets[2 * match_num + 1];
944 * Returns number of first matched subpattern with name @name.
945 * There may be more than one in case when DUPNAMES is used,
946 * and not all subpatterns with that name match;
947 * pcre_get_stringnumber() does not work in that case.
950 get_matched_substring_number (const GMatchInfo *match_info,
957 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
958 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
960 /* This code is copied from pcre_get.c: get_first_set() */
961 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
969 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
971 gint n = (entry[0] << 8) + entry[1];
972 if (match_info->offsets[n*2] >= 0)
976 return (first[0] << 8) + first[1];
980 * g_match_info_fetch_named:
981 * @match_info: #GMatchInfo structure
982 * @name: name of the subexpression
984 * Retrieves the text matching the capturing parentheses named @name.
986 * If @name is a valid sub pattern name but it didn't match anything
987 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
988 * then an empty string is returned.
990 * The string is fetched from the string passed to the match function,
991 * so you cannot call this function after freeing the string.
993 * Returns: (allow-none): The matched substring, or %NULL if an error
994 * occurred. You have to free the string yourself
999 g_match_info_fetch_named (const GMatchInfo *match_info,
1002 /* we cannot use pcre_get_named_substring() because it allocates the
1003 * string using pcre_malloc(). */
1006 g_return_val_if_fail (match_info != NULL, NULL);
1007 g_return_val_if_fail (name != NULL, NULL);
1009 num = get_matched_substring_number (match_info, name);
1013 return g_match_info_fetch (match_info, num);
1017 * g_match_info_fetch_named_pos:
1018 * @match_info: #GMatchInfo structure
1019 * @name: name of the subexpression
1020 * @start_pos: (out) (allow-none): pointer to location where to store
1021 * the start position, or %NULL
1022 * @end_pos: (out) (allow-none): pointer to location where to store
1023 * the end position, or %NULL
1025 * Retrieves the position in bytes of the capturing parentheses named @name.
1027 * If @name is a valid sub pattern name but it didn't match anything
1028 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1029 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
1031 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
1032 * If the position cannot be fetched, @start_pos and @end_pos
1033 * are left unchanged.
1038 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
1045 g_return_val_if_fail (match_info != NULL, FALSE);
1046 g_return_val_if_fail (name != NULL, FALSE);
1048 num = get_matched_substring_number (match_info, name);
1052 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1056 * g_match_info_fetch_all:
1057 * @match_info: a #GMatchInfo structure
1059 * Bundles up pointers to each of the matching substrings from a match
1060 * and stores them in an array of gchar pointers. The first element in
1061 * the returned array is the match number 0, i.e. the entire matched
1064 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1065 * "b" against "(a)?b") then an empty string is inserted.
1067 * If the last match was obtained using the DFA algorithm, that is using
1068 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1069 * strings are not that matched by sets of parentheses but that of the
1070 * matched substring. Substrings are matched in reverse order of length,
1071 * so the first one is the longest match.
1073 * The strings are fetched from the string passed to the match function,
1074 * so you cannot call this function after freeing the string.
1076 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1077 * It must be freed using g_strfreev(). If the previous match failed
1083 g_match_info_fetch_all (const GMatchInfo *match_info)
1085 /* we cannot use pcre_get_substring_list() because the returned value
1086 * isn't suitable for g_strfreev(). */
1090 g_return_val_if_fail (match_info != NULL, NULL);
1092 if (match_info->matches < 0)
1095 result = g_new (gchar *, match_info->matches + 1);
1096 for (i = 0; i < match_info->matches; i++)
1097 result[i] = g_match_info_fetch (match_info, i);
1107 g_regex_error_quark (void)
1109 static GQuark error_quark = 0;
1111 if (error_quark == 0)
1112 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1121 * Increases reference count of @regex by 1.
1128 g_regex_ref (GRegex *regex)
1130 g_return_val_if_fail (regex != NULL, NULL);
1131 g_atomic_int_inc (®ex->ref_count);
1139 * Decreases reference count of @regex by 1. When reference count drops
1140 * to zero, it frees all the memory associated with the regex structure.
1145 g_regex_unref (GRegex *regex)
1147 g_return_if_fail (regex != NULL);
1149 if (g_atomic_int_dec_and_test (®ex->ref_count))
1151 g_free (regex->pattern);
1152 if (regex->pcre_re != NULL)
1153 pcre_free (regex->pcre_re);
1154 if (regex->extra != NULL)
1155 pcre_free (regex->extra);
1162 * @pattern: the regular expression
1163 * @compile_options: compile options for the regular expression, or 0
1164 * @match_options: match options for the regular expression, or 0
1165 * @error: return location for a #GError
1167 * Compiles the regular expression to an internal form, and does
1168 * the initial setup of the #GRegex structure.
1170 * Returns: a #GRegex structure. Call g_regex_unref() when you
1176 g_regex_new (const gchar *pattern,
1177 GRegexCompileFlags compile_options,
1178 GRegexMatchFlags match_options,
1183 const gchar *errmsg;
1186 gboolean optimize = FALSE;
1187 static gsize initialised;
1188 unsigned long int pcre_compile_options;
1190 g_return_val_if_fail (pattern != NULL, NULL);
1191 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1192 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1193 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1195 if (g_once_init_enter (&initialised))
1200 pcre_config (PCRE_CONFIG_UTF8, &support);
1203 msg = N_("PCRE library is compiled without UTF8 support");
1204 g_critical ("%s", msg);
1205 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1209 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1212 msg = N_("PCRE library is compiled without UTF8 properties support");
1213 g_critical ("%s", msg);
1214 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1218 g_once_init_leave (&initialised, TRUE);
1221 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1222 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1223 if (compile_options & G_REGEX_OPTIMIZE)
1226 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1227 * instead uses UTF-8 only if required with PCRE_UTF8. */
1228 if (compile_options & G_REGEX_RAW)
1231 compile_options &= ~G_REGEX_RAW;
1236 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1237 match_options |= PCRE_NO_UTF8_CHECK;
1240 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1241 * not for the system one. */
1242 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1243 !(compile_options & G_REGEX_NEWLINE_LF))
1245 compile_options |= PCRE_NEWLINE_ANY;
1248 compile_options |= PCRE_UCP;
1250 /* compile the pattern */
1251 re = pcre_compile2 (pattern, compile_options, &errcode,
1252 &errmsg, &erroffset, NULL);
1254 /* if the compilation failed, set the error member and return
1260 /* Translate the PCRE error code to GRegexError and use a translated
1261 * error message if possible */
1262 translate_compile_error (&errcode, &errmsg);
1264 /* PCRE uses byte offsets but we want to show character offsets */
1265 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1267 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1268 _("Error while compiling regular "
1269 "expression %s at char %d: %s"),
1270 pattern, erroffset, errmsg);
1271 g_propagate_error (error, tmp_error);
1276 /* For options set at the beginning of the pattern, pcre puts them into
1277 * compile options, e.g. "(?i)foo" will make the pcre structure store
1278 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1279 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1280 compile_options = pcre_compile_options;
1282 if (!(compile_options & G_REGEX_DUPNAMES))
1284 gboolean jchanged = FALSE;
1285 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1287 compile_options |= G_REGEX_DUPNAMES;
1290 regex = g_new0 (GRegex, 1);
1291 regex->ref_count = 1;
1292 regex->pattern = g_strdup (pattern);
1293 regex->pcre_re = re;
1294 regex->compile_opts = compile_options;
1295 regex->match_opts = match_options;
1299 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1302 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1303 G_REGEX_ERROR_OPTIMIZE,
1304 _("Error while optimizing "
1305 "regular expression %s: %s"),
1308 g_propagate_error (error, tmp_error);
1310 g_regex_unref (regex);
1319 * g_regex_get_pattern:
1320 * @regex: a #GRegex structure
1322 * Gets the pattern string associated with @regex, i.e. a copy of
1323 * the string passed to g_regex_new().
1325 * Returns: the pattern of @regex
1330 g_regex_get_pattern (const GRegex *regex)
1332 g_return_val_if_fail (regex != NULL, NULL);
1334 return regex->pattern;
1338 * g_regex_get_max_backref:
1341 * Returns the number of the highest back reference
1342 * in the pattern, or 0 if the pattern does not contain
1345 * Returns: the number of the highest back reference
1350 g_regex_get_max_backref (const GRegex *regex)
1354 pcre_fullinfo (regex->pcre_re, regex->extra,
1355 PCRE_INFO_BACKREFMAX, &value);
1361 * g_regex_get_capture_count:
1364 * Returns the number of capturing subpatterns in the pattern.
1366 * Returns: the number of capturing subpatterns
1371 g_regex_get_capture_count (const GRegex *regex)
1375 pcre_fullinfo (regex->pcre_re, regex->extra,
1376 PCRE_INFO_CAPTURECOUNT, &value);
1382 * g_regex_get_compile_flags:
1385 * Returns the compile options that @regex was created with.
1387 * Returns: flags from #GRegexCompileFlags
1392 g_regex_get_compile_flags (const GRegex *regex)
1394 g_return_val_if_fail (regex != NULL, 0);
1396 return regex->compile_opts;
1400 * g_regex_get_match_flags:
1403 * Returns the match options that @regex was created with.
1405 * Returns: flags from #GRegexMatchFlags
1410 g_regex_get_match_flags (const GRegex *regex)
1412 g_return_val_if_fail (regex != NULL, 0);
1414 return regex->match_opts;
1418 * g_regex_match_simple:
1419 * @pattern: the regular expression
1420 * @string: the string to scan for matches
1421 * @compile_options: compile options for the regular expression, or 0
1422 * @match_options: match options, or 0
1424 * Scans for a match in @string for @pattern.
1426 * This function is equivalent to g_regex_match() but it does not
1427 * require to compile the pattern with g_regex_new(), avoiding some
1428 * lines of code when you need just to do a match without extracting
1429 * substrings, capture counts, and so on.
1431 * If this function is to be called on the same @pattern more than
1432 * once, it's more efficient to compile the pattern once with
1433 * g_regex_new() and then use g_regex_match().
1435 * Returns: %TRUE if the string matched, %FALSE otherwise
1440 g_regex_match_simple (const gchar *pattern,
1441 const gchar *string,
1442 GRegexCompileFlags compile_options,
1443 GRegexMatchFlags match_options)
1448 regex = g_regex_new (pattern, compile_options, 0, NULL);
1451 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1452 g_regex_unref (regex);
1458 * @regex: a #GRegex structure from g_regex_new()
1459 * @string: the string to scan for matches
1460 * @match_options: match options
1461 * @match_info: (out) (allow-none): pointer to location where to store
1462 * the #GMatchInfo, or %NULL if you do not need it
1464 * Scans for a match in string for the pattern in @regex.
1465 * The @match_options are combined with the match options specified
1466 * when the @regex structure was created, letting you have more
1467 * flexibility in reusing #GRegex structures.
1469 * A #GMatchInfo structure, used to get information on the match,
1470 * is stored in @match_info if not %NULL. Note that if @match_info
1471 * is not %NULL then it is created even if the function returns %FALSE,
1472 * i.e. you must free it regardless if regular expression actually matched.
1474 * To retrieve all the non-overlapping matches of the pattern in
1475 * string you can use g_match_info_next().
1479 * print_uppercase_words (const gchar *string)
1481 * /* Print all uppercase-only words. */
1483 * GMatchInfo *match_info;
1485 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1486 * g_regex_match (regex, string, 0, &match_info);
1487 * while (g_match_info_matches (match_info))
1489 * gchar *word = g_match_info_fetch (match_info, 0);
1490 * g_print ("Found: %s\n", word);
1492 * g_match_info_next (match_info, NULL);
1494 * g_match_info_free (match_info);
1495 * g_regex_unref (regex);
1499 * @string is not copied and is used in #GMatchInfo internally. If
1500 * you use any #GMatchInfo method (except g_match_info_free()) after
1501 * freeing or modifying @string then the behaviour is undefined.
1503 * Returns: %TRUE is the string matched, %FALSE otherwise
1508 g_regex_match (const GRegex *regex,
1509 const gchar *string,
1510 GRegexMatchFlags match_options,
1511 GMatchInfo **match_info)
1513 return g_regex_match_full (regex, string, -1, 0, match_options,
1518 * g_regex_match_full:
1519 * @regex: a #GRegex structure from g_regex_new()
1520 * @string: (array length=string_len): the string to scan for matches
1521 * @string_len: the length of @string, or -1 if @string is nul-terminated
1522 * @start_position: starting index of the string to match
1523 * @match_options: match options
1524 * @match_info: (out) (allow-none): pointer to location where to store
1525 * the #GMatchInfo, or %NULL if you do not need it
1526 * @error: location to store the error occurring, or %NULL to ignore errors
1528 * Scans for a match in string for the pattern in @regex.
1529 * The @match_options are combined with the match options specified
1530 * when the @regex structure was created, letting you have more
1531 * flexibility in reusing #GRegex structures.
1533 * Setting @start_position differs from just passing over a shortened
1534 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1535 * that begins with any kind of lookbehind assertion, such as "\b".
1537 * A #GMatchInfo structure, used to get information on the match, is
1538 * stored in @match_info if not %NULL. Note that if @match_info is
1539 * not %NULL then it is created even if the function returns %FALSE,
1540 * i.e. you must free it regardless if regular expression actually
1543 * @string is not copied and is used in #GMatchInfo internally. If
1544 * you use any #GMatchInfo method (except g_match_info_free()) after
1545 * freeing or modifying @string then the behaviour is undefined.
1547 * To retrieve all the non-overlapping matches of the pattern in
1548 * string you can use g_match_info_next().
1552 * print_uppercase_words (const gchar *string)
1554 * /* Print all uppercase-only words. */
1556 * GMatchInfo *match_info;
1557 * GError *error = NULL;
1559 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1560 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1561 * while (g_match_info_matches (match_info))
1563 * gchar *word = g_match_info_fetch (match_info, 0);
1564 * g_print ("Found: %s\n", word);
1566 * g_match_info_next (match_info, &error);
1568 * g_match_info_free (match_info);
1569 * g_regex_unref (regex);
1570 * if (error != NULL)
1572 * g_printerr ("Error while matching: %s\n", error->message);
1573 * g_error_free (error);
1578 * Returns: %TRUE is the string matched, %FALSE otherwise
1583 g_regex_match_full (const GRegex *regex,
1584 const gchar *string,
1586 gint start_position,
1587 GRegexMatchFlags match_options,
1588 GMatchInfo **match_info,
1594 g_return_val_if_fail (regex != NULL, FALSE);
1595 g_return_val_if_fail (string != NULL, FALSE);
1596 g_return_val_if_fail (start_position >= 0, FALSE);
1597 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1598 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1600 info = match_info_new (regex, string, string_len, start_position,
1601 match_options, FALSE);
1602 match_ok = g_match_info_next (info, error);
1603 if (match_info != NULL)
1606 g_match_info_free (info);
1612 * g_regex_match_all:
1613 * @regex: a #GRegex structure from g_regex_new()
1614 * @string: the string to scan for matches
1615 * @match_options: match options
1616 * @match_info: (out) (allow-none): pointer to location where to store
1617 * the #GMatchInfo, or %NULL if you do not need it
1619 * Using the standard algorithm for regular expression matching only
1620 * the longest match in the string is retrieved. This function uses
1621 * a different algorithm so it can retrieve all the possible matches.
1622 * For more documentation see g_regex_match_all_full().
1624 * A #GMatchInfo structure, used to get information on the match, is
1625 * stored in @match_info if not %NULL. Note that if @match_info is
1626 * not %NULL then it is created even if the function returns %FALSE,
1627 * i.e. you must free it regardless if regular expression actually
1630 * @string is not copied and is used in #GMatchInfo internally. If
1631 * you use any #GMatchInfo method (except g_match_info_free()) after
1632 * freeing or modifying @string then the behaviour is undefined.
1634 * Returns: %TRUE is the string matched, %FALSE otherwise
1639 g_regex_match_all (const GRegex *regex,
1640 const gchar *string,
1641 GRegexMatchFlags match_options,
1642 GMatchInfo **match_info)
1644 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1649 * g_regex_match_all_full:
1650 * @regex: a #GRegex structure from g_regex_new()
1651 * @string: (array length=string_len): the string to scan for matches
1652 * @string_len: the length of @string, or -1 if @string is nul-terminated
1653 * @start_position: starting index of the string to match
1654 * @match_options: match options
1655 * @match_info: (out) (allow-none): pointer to location where to store
1656 * the #GMatchInfo, or %NULL if you do not need it
1657 * @error: location to store the error occurring, or %NULL to ignore errors
1659 * Using the standard algorithm for regular expression matching only
1660 * the longest match in the string is retrieved, it is not possible
1661 * to obtain all the available matches. For instance matching
1662 * "<a> <b> <c>" against the pattern "<.*>"
1663 * you get "<a> <b> <c>".
1665 * This function uses a different algorithm (called DFA, i.e. deterministic
1666 * finite automaton), so it can retrieve all the possible matches, all
1667 * starting at the same point in the string. For instance matching
1668 * "<a> <b> <c>" against the pattern "<.*>"
1669 * you would obtain three matches: "<a> <b> <c>",
1670 * "<a> <b>" and "<a>".
1672 * The number of matched strings is retrieved using
1673 * g_match_info_get_match_count(). To obtain the matched strings and
1674 * their position you can use, respectively, g_match_info_fetch() and
1675 * g_match_info_fetch_pos(). Note that the strings are returned in
1676 * reverse order of length; that is, the longest matching string is
1679 * Note that the DFA algorithm is slower than the standard one and it
1680 * is not able to capture substrings, so backreferences do not work.
1682 * Setting @start_position differs from just passing over a shortened
1683 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1684 * that begins with any kind of lookbehind assertion, such as "\b".
1686 * A #GMatchInfo structure, used to get information on the match, is
1687 * stored in @match_info if not %NULL. Note that if @match_info is
1688 * not %NULL then it is created even if the function returns %FALSE,
1689 * i.e. you must free it regardless if regular expression actually
1692 * @string is not copied and is used in #GMatchInfo internally. If
1693 * you use any #GMatchInfo method (except g_match_info_free()) after
1694 * freeing or modifying @string then the behaviour is undefined.
1696 * Returns: %TRUE is the string matched, %FALSE otherwise
1701 g_regex_match_all_full (const GRegex *regex,
1702 const gchar *string,
1704 gint start_position,
1705 GRegexMatchFlags match_options,
1706 GMatchInfo **match_info,
1712 g_return_val_if_fail (regex != NULL, FALSE);
1713 g_return_val_if_fail (string != NULL, FALSE);
1714 g_return_val_if_fail (start_position >= 0, FALSE);
1715 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1716 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1718 info = match_info_new (regex, string, string_len, start_position,
1719 match_options, TRUE);
1725 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1726 info->string, info->string_len,
1728 regex->match_opts | match_options,
1729 info->offsets, info->n_offsets,
1730 info->workspace, info->n_workspace);
1731 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1733 /* info->workspace is too small. */
1734 info->n_workspace *= 2;
1735 info->workspace = g_realloc (info->workspace,
1736 info->n_workspace * sizeof (gint));
1739 else if (info->matches == 0)
1741 /* info->offsets is too small. */
1742 info->n_offsets *= 2;
1743 info->offsets = g_realloc (info->offsets,
1744 info->n_offsets * sizeof (gint));
1747 else if (IS_PCRE_ERROR (info->matches))
1749 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1750 _("Error while matching regular expression %s: %s"),
1751 regex->pattern, match_error (info->matches));
1755 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1758 if (match_info != NULL)
1761 g_match_info_free (info);
1763 return info->matches >= 0;
1767 * g_regex_get_string_number:
1768 * @regex: #GRegex structure
1769 * @name: name of the subexpression
1771 * Retrieves the number of the subexpression named @name.
1773 * Returns: The number of the subexpression or -1 if @name
1779 g_regex_get_string_number (const GRegex *regex,
1784 g_return_val_if_fail (regex != NULL, -1);
1785 g_return_val_if_fail (name != NULL, -1);
1787 num = pcre_get_stringnumber (regex->pcre_re, name);
1788 if (num == PCRE_ERROR_NOSUBSTRING)
1795 * g_regex_split_simple:
1796 * @pattern: the regular expression
1797 * @string: the string to scan for matches
1798 * @compile_options: compile options for the regular expression, or 0
1799 * @match_options: match options, or 0
1801 * Breaks the string on the pattern, and returns an array of
1802 * the tokens. If the pattern contains capturing parentheses,
1803 * then the text for each of the substrings will also be returned.
1804 * If the pattern does not match anywhere in the string, then the
1805 * whole string is returned as the first token.
1807 * This function is equivalent to g_regex_split() but it does
1808 * not require to compile the pattern with g_regex_new(), avoiding
1809 * some lines of code when you need just to do a split without
1810 * extracting substrings, capture counts, and so on.
1812 * If this function is to be called on the same @pattern more than
1813 * once, it's more efficient to compile the pattern once with
1814 * g_regex_new() and then use g_regex_split().
1816 * As a special case, the result of splitting the empty string ""
1817 * is an empty vector, not a vector containing a single string.
1818 * The reason for this special case is that being able to represent
1819 * a empty vector is typically more useful than consistent handling
1820 * of empty elements. If you do need to represent empty elements,
1821 * you'll need to check for the empty string before calling this
1824 * A pattern that can match empty strings splits @string into
1825 * separate characters wherever it matches the empty string between
1826 * characters. For example splitting "ab c" using as a separator
1827 * "\s*", you will get "a", "b" and "c".
1829 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1834 g_regex_split_simple (const gchar *pattern,
1835 const gchar *string,
1836 GRegexCompileFlags compile_options,
1837 GRegexMatchFlags match_options)
1842 regex = g_regex_new (pattern, compile_options, 0, NULL);
1846 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1847 g_regex_unref (regex);
1853 * @regex: a #GRegex structure
1854 * @string: the string to split with the pattern
1855 * @match_options: match time option flags
1857 * Breaks the string on the pattern, and returns an array of the tokens.
1858 * If the pattern contains capturing parentheses, then the text for each
1859 * of the substrings will also be returned. If the pattern does not match
1860 * anywhere in the string, then the whole string is returned as the first
1863 * As a special case, the result of splitting the empty string "" is an
1864 * empty vector, not a vector containing a single string. The reason for
1865 * this special case is that being able to represent a empty vector is
1866 * typically more useful than consistent handling of empty elements. If
1867 * you do need to represent empty elements, you'll need to check for the
1868 * empty string before calling this function.
1870 * A pattern that can match empty strings splits @string into separate
1871 * characters wherever it matches the empty string between characters.
1872 * For example splitting "ab c" using as a separator "\s*", you will get
1875 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1880 g_regex_split (const GRegex *regex,
1881 const gchar *string,
1882 GRegexMatchFlags match_options)
1884 return g_regex_split_full (regex, string, -1, 0,
1885 match_options, 0, NULL);
1889 * g_regex_split_full:
1890 * @regex: a #GRegex structure
1891 * @string: (array length=string_len): the string to split with the pattern
1892 * @string_len: the length of @string, or -1 if @string is nul-terminated
1893 * @start_position: starting index of the string to match
1894 * @match_options: match time option flags
1895 * @max_tokens: the maximum number of tokens to split @string into.
1896 * If this is less than 1, the string is split completely
1897 * @error: return location for a #GError
1899 * Breaks the string on the pattern, and returns an array of the tokens.
1900 * If the pattern contains capturing parentheses, then the text for each
1901 * of the substrings will also be returned. If the pattern does not match
1902 * anywhere in the string, then the whole string is returned as the first
1905 * As a special case, the result of splitting the empty string "" is an
1906 * empty vector, not a vector containing a single string. The reason for
1907 * this special case is that being able to represent a empty vector is
1908 * typically more useful than consistent handling of empty elements. If
1909 * you do need to represent empty elements, you'll need to check for the
1910 * empty string before calling this function.
1912 * A pattern that can match empty strings splits @string into separate
1913 * characters wherever it matches the empty string between characters.
1914 * For example splitting "ab c" using as a separator "\s*", you will get
1917 * Setting @start_position differs from just passing over a shortened
1918 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1919 * that begins with any kind of lookbehind assertion, such as "\b".
1921 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1926 g_regex_split_full (const GRegex *regex,
1927 const gchar *string,
1929 gint start_position,
1930 GRegexMatchFlags match_options,
1934 GError *tmp_error = NULL;
1935 GMatchInfo *match_info;
1940 /* position of the last separator. */
1941 gint last_separator_end;
1942 /* was the last match 0 bytes long? */
1943 gboolean last_match_is_empty;
1944 /* the returned array of char **s */
1945 gchar **string_list;
1947 g_return_val_if_fail (regex != NULL, NULL);
1948 g_return_val_if_fail (string != NULL, NULL);
1949 g_return_val_if_fail (start_position >= 0, NULL);
1950 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1951 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1953 if (max_tokens <= 0)
1954 max_tokens = G_MAXINT;
1957 string_len = strlen (string);
1959 /* zero-length string */
1960 if (string_len - start_position == 0)
1961 return g_new0 (gchar *, 1);
1963 if (max_tokens == 1)
1965 string_list = g_new0 (gchar *, 2);
1966 string_list[0] = g_strndup (&string[start_position],
1967 string_len - start_position);
1973 last_separator_end = start_position;
1974 last_match_is_empty = FALSE;
1976 match_ok = g_regex_match_full (regex, string, string_len, start_position,
1977 match_options, &match_info, &tmp_error);
1979 while (tmp_error == NULL)
1983 last_match_is_empty =
1984 (match_info->offsets[0] == match_info->offsets[1]);
1986 /* we need to skip empty separators at the same position of the end
1987 * of another separator. e.g. the string is "a b" and the separator
1988 * is " *", so from 1 to 2 we have a match and at position 2 we have
1989 * an empty match. */
1990 if (last_separator_end != match_info->offsets[1])
1995 token = g_strndup (string + last_separator_end,
1996 match_info->offsets[0] - last_separator_end);
1997 list = g_list_prepend (list, token);
2000 /* if there were substrings, these need to be added to
2002 match_count = g_match_info_get_match_count (match_info);
2003 if (match_count > 1)
2005 for (i = 1; i < match_count; i++)
2006 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
2012 /* if there was no match, copy to end of string. */
2013 if (!last_match_is_empty)
2015 gchar *token = g_strndup (string + last_separator_end,
2016 match_info->string_len - last_separator_end);
2017 list = g_list_prepend (list, token);
2019 /* no more tokens, end the loop. */
2023 /* -1 to leave room for the last part. */
2024 if (token_count >= max_tokens - 1)
2026 /* we have reached the maximum number of tokens, so we copy
2027 * the remaining part of the string. */
2028 if (last_match_is_empty)
2030 /* the last match was empty, so we have moved one char
2031 * after the real position to avoid empty matches at the
2033 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
2035 /* the if is needed in the case we have terminated the available
2036 * tokens, but we are at the end of the string, so there are no
2037 * characters left to copy. */
2038 if (string_len > match_info->pos)
2040 gchar *token = g_strndup (string + match_info->pos,
2041 string_len - match_info->pos);
2042 list = g_list_prepend (list, token);
2048 last_separator_end = match_info->pos;
2049 if (last_match_is_empty)
2050 /* if the last match was empty, g_match_info_next() has moved
2051 * forward to avoid infinite loops, but we still need to copy that
2053 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
2055 match_ok = g_match_info_next (match_info, &tmp_error);
2057 g_match_info_free (match_info);
2058 if (tmp_error != NULL)
2060 g_propagate_error (error, tmp_error);
2061 g_list_free_full (list, g_free);
2062 match_info->pos = -1;
2066 string_list = g_new (gchar *, g_list_length (list) + 1);
2068 for (last = g_list_last (list); last; last = g_list_previous (last))
2069 string_list[i++] = last->data;
2070 string_list[i] = NULL;
2079 REPL_TYPE_CHARACTER,
2080 REPL_TYPE_SYMBOLIC_REFERENCE,
2081 REPL_TYPE_NUMERIC_REFERENCE,
2082 REPL_TYPE_CHANGE_CASE
2087 CHANGE_CASE_NONE = 1 << 0,
2088 CHANGE_CASE_UPPER = 1 << 1,
2089 CHANGE_CASE_LOWER = 1 << 2,
2090 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2091 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2092 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2093 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2094 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2097 struct _InterpolationData
2103 ChangeCase change_case;
2107 free_interpolation_data (InterpolationData *data)
2109 g_free (data->text);
2113 static const gchar *
2114 expand_escape (const gchar *replacement,
2116 InterpolationData *data,
2121 const gchar *error_detail;
2123 GError *tmp_error = NULL;
2131 data->type = REPL_TYPE_CHARACTER;
2136 data->type = REPL_TYPE_CHARACTER;
2141 data->type = REPL_TYPE_CHARACTER;
2146 data->type = REPL_TYPE_CHARACTER;
2151 data->type = REPL_TYPE_CHARACTER;
2156 data->type = REPL_TYPE_CHARACTER;
2161 data->type = REPL_TYPE_CHARACTER;
2166 data->type = REPL_TYPE_CHARACTER;
2176 h = g_ascii_xdigit_value (*p);
2179 error_detail = _("hexadecimal digit or '}' expected");
2190 for (i = 0; i < 2; i++)
2192 h = g_ascii_xdigit_value (*p);
2195 error_detail = _("hexadecimal digit expected");
2202 data->type = REPL_TYPE_STRING;
2203 data->text = g_new0 (gchar, 8);
2204 g_unichar_to_utf8 (x, data->text);
2208 data->type = REPL_TYPE_CHANGE_CASE;
2209 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2213 data->type = REPL_TYPE_CHANGE_CASE;
2214 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2218 data->type = REPL_TYPE_CHANGE_CASE;
2219 data->change_case = CHANGE_CASE_LOWER;
2223 data->type = REPL_TYPE_CHANGE_CASE;
2224 data->change_case = CHANGE_CASE_UPPER;
2228 data->type = REPL_TYPE_CHANGE_CASE;
2229 data->change_case = CHANGE_CASE_NONE;
2235 error_detail = _("missing '<' in symbolic reference");
2244 error_detail = _("unfinished symbolic reference");
2251 error_detail = _("zero-length symbolic reference");
2254 if (g_ascii_isdigit (*q))
2259 h = g_ascii_digit_value (*q);
2262 error_detail = _("digit expected");
2271 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2278 if (!g_ascii_isalnum (*r))
2280 error_detail = _("illegal symbolic reference");
2287 data->text = g_strndup (q, p - q);
2288 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2293 /* if \0 is followed by a number is an octal number representing a
2294 * character, else it is a numeric reference. */
2295 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2298 p = g_utf8_next_char (p);
2311 for (i = 0; i < 3; i++)
2313 h = g_ascii_digit_value (*p);
2323 if (i == 2 && base == 10)
2329 if (base == 8 || i == 3)
2331 data->type = REPL_TYPE_STRING;
2332 data->text = g_new0 (gchar, 8);
2333 g_unichar_to_utf8 (x, data->text);
2337 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2342 error_detail = _("stray final '\\'");
2346 error_detail = _("unknown escape sequence");
2353 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2354 tmp_error = g_error_new (G_REGEX_ERROR,
2355 G_REGEX_ERROR_REPLACE,
2356 _("Error while parsing replacement "
2357 "text \"%s\" at char %lu: %s"),
2359 (gulong)(p - replacement),
2361 g_propagate_error (error, tmp_error);
2367 split_replacement (const gchar *replacement,
2371 InterpolationData *data;
2372 const gchar *p, *start;
2374 start = p = replacement;
2379 data = g_new0 (InterpolationData, 1);
2380 start = p = expand_escape (replacement, p, data, error);
2383 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2384 free_interpolation_data (data);
2388 list = g_list_prepend (list, data);
2393 if (*p == '\\' || *p == '\0')
2397 data = g_new0 (InterpolationData, 1);
2398 data->text = g_strndup (start, p - start);
2399 data->type = REPL_TYPE_STRING;
2400 list = g_list_prepend (list, data);
2406 return g_list_reverse (list);
2409 /* Change the case of c based on change_case. */
2410 #define CHANGE_CASE(c, change_case) \
2411 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2412 g_unichar_tolower (c) : \
2413 g_unichar_toupper (c))
2416 string_append (GString *string,
2418 ChangeCase *change_case)
2422 if (text[0] == '\0')
2425 if (*change_case == CHANGE_CASE_NONE)
2427 g_string_append (string, text);
2429 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2431 c = g_utf8_get_char (text);
2432 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2433 g_string_append (string, g_utf8_next_char (text));
2434 *change_case = CHANGE_CASE_NONE;
2438 while (*text != '\0')
2440 c = g_utf8_get_char (text);
2441 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2442 text = g_utf8_next_char (text);
2448 interpolate_replacement (const GMatchInfo *match_info,
2453 InterpolationData *idata;
2455 ChangeCase change_case = CHANGE_CASE_NONE;
2457 for (list = data; list; list = list->next)
2460 switch (idata->type)
2462 case REPL_TYPE_STRING:
2463 string_append (result, idata->text, &change_case);
2465 case REPL_TYPE_CHARACTER:
2466 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2467 if (change_case & CHANGE_CASE_SINGLE_MASK)
2468 change_case = CHANGE_CASE_NONE;
2470 case REPL_TYPE_NUMERIC_REFERENCE:
2471 match = g_match_info_fetch (match_info, idata->num);
2474 string_append (result, match, &change_case);
2478 case REPL_TYPE_SYMBOLIC_REFERENCE:
2479 match = g_match_info_fetch_named (match_info, idata->text);
2482 string_append (result, match, &change_case);
2486 case REPL_TYPE_CHANGE_CASE:
2487 change_case = idata->change_case;
2495 /* whether actual match_info is needed for replacement, i.e.
2496 * whether there are references
2499 interpolation_list_needs_match (GList *list)
2501 while (list != NULL)
2503 InterpolationData *data = list->data;
2505 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2506 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2519 * @regex: a #GRegex structure
2520 * @string: (array length=string_len): the string to perform matches against
2521 * @string_len: the length of @string, or -1 if @string is nul-terminated
2522 * @start_position: starting index of the string to match
2523 * @replacement: text to replace each match with
2524 * @match_options: options for the match
2525 * @error: location to store the error occurring, or %NULL to ignore errors
2527 * Replaces all occurrences of the pattern in @regex with the
2528 * replacement text. Backreferences of the form '\number' or
2529 * '\g<number>' in the replacement text are interpolated by the
2530 * number-th captured subexpression of the match, '\g<name>' refers
2531 * to the captured subexpression with the given name. '\0' refers to the
2532 * complete match, but '\0' followed by a number is the octal representation
2533 * of a character. To include a literal '\' in the replacement, write '\\'.
2534 * There are also escapes that changes the case of the following text:
2537 * <varlistentry><term>\l</term>
2539 * <para>Convert to lower case the next character</para>
2542 * <varlistentry><term>\u</term>
2544 * <para>Convert to upper case the next character</para>
2547 * <varlistentry><term>\L</term>
2549 * <para>Convert to lower case till \E</para>
2552 * <varlistentry><term>\U</term>
2554 * <para>Convert to upper case till \E</para>
2557 * <varlistentry><term>\E</term>
2559 * <para>End case modification</para>
2564 * If you do not need to use backreferences use g_regex_replace_literal().
2566 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2567 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2568 * you can use g_regex_replace_literal().
2570 * Setting @start_position differs from just passing over a shortened
2571 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2572 * begins with any kind of lookbehind assertion, such as "\b".
2574 * Returns: a newly allocated string containing the replacements
2579 g_regex_replace (const GRegex *regex,
2580 const gchar *string,
2582 gint start_position,
2583 const gchar *replacement,
2584 GRegexMatchFlags match_options,
2589 GError *tmp_error = NULL;
2591 g_return_val_if_fail (regex != NULL, NULL);
2592 g_return_val_if_fail (string != NULL, NULL);
2593 g_return_val_if_fail (start_position >= 0, NULL);
2594 g_return_val_if_fail (replacement != NULL, NULL);
2595 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2596 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2598 list = split_replacement (replacement, &tmp_error);
2599 if (tmp_error != NULL)
2601 g_propagate_error (error, tmp_error);
2605 result = g_regex_replace_eval (regex,
2606 string, string_len, start_position,
2608 interpolate_replacement,
2611 if (tmp_error != NULL)
2612 g_propagate_error (error, tmp_error);
2614 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2620 literal_replacement (const GMatchInfo *match_info,
2624 g_string_append (result, data);
2629 * g_regex_replace_literal:
2630 * @regex: a #GRegex structure
2631 * @string: (array length=string_len): the string to perform matches against
2632 * @string_len: the length of @string, or -1 if @string is nul-terminated
2633 * @start_position: starting index of the string to match
2634 * @replacement: text to replace each match with
2635 * @match_options: options for the match
2636 * @error: location to store the error occurring, or %NULL to ignore errors
2638 * Replaces all occurrences of the pattern in @regex with the
2639 * replacement text. @replacement is replaced literally, to
2640 * include backreferences use g_regex_replace().
2642 * Setting @start_position differs from just passing over a
2643 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2644 * case of a pattern that begins with any kind of lookbehind
2645 * assertion, such as "\b".
2647 * Returns: a newly allocated string containing the replacements
2652 g_regex_replace_literal (const GRegex *regex,
2653 const gchar *string,
2655 gint start_position,
2656 const gchar *replacement,
2657 GRegexMatchFlags match_options,
2660 g_return_val_if_fail (replacement != NULL, NULL);
2661 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2663 return g_regex_replace_eval (regex,
2664 string, string_len, start_position,
2666 literal_replacement,
2667 (gpointer)replacement,
2672 * g_regex_replace_eval:
2673 * @regex: a #GRegex structure from g_regex_new()
2674 * @string: (array length=string_len): string to perform matches against
2675 * @string_len: the length of @string, or -1 if @string is nul-terminated
2676 * @start_position: starting index of the string to match
2677 * @match_options: options for the match
2678 * @eval: a function to call for each match
2679 * @user_data: user data to pass to the function
2680 * @error: location to store the error occurring, or %NULL to ignore errors
2682 * Replaces occurrences of the pattern in regex with the output of
2683 * @eval for that occurrence.
2685 * Setting @start_position differs from just passing over a shortened
2686 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2687 * that begins with any kind of lookbehind assertion, such as "\b".
2689 * The following example uses g_regex_replace_eval() to replace multiple
2693 * eval_cb (const GMatchInfo *info,
2700 * match = g_match_info_fetch (info, 0);
2701 * r = g_hash_table_lookup ((GHashTable *)data, match);
2702 * g_string_append (res, r);
2714 * h = g_hash_table_new (g_str_hash, g_str_equal);
2716 * g_hash_table_insert (h, "1", "ONE");
2717 * g_hash_table_insert (h, "2", "TWO");
2718 * g_hash_table_insert (h, "3", "THREE");
2719 * g_hash_table_insert (h, "4", "FOUR");
2721 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2722 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2723 * g_hash_table_destroy (h);
2728 * Returns: a newly allocated string containing the replacements
2733 g_regex_replace_eval (const GRegex *regex,
2734 const gchar *string,
2736 gint start_position,
2737 GRegexMatchFlags match_options,
2738 GRegexEvalCallback eval,
2742 GMatchInfo *match_info;
2745 gboolean done = FALSE;
2746 GError *tmp_error = NULL;
2748 g_return_val_if_fail (regex != NULL, NULL);
2749 g_return_val_if_fail (string != NULL, NULL);
2750 g_return_val_if_fail (start_position >= 0, NULL);
2751 g_return_val_if_fail (eval != NULL, NULL);
2752 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2755 string_len = strlen (string);
2757 result = g_string_sized_new (string_len);
2759 /* run down the string making matches. */
2760 g_regex_match_full (regex, string, string_len, start_position,
2761 match_options, &match_info, &tmp_error);
2762 while (!done && g_match_info_matches (match_info))
2764 g_string_append_len (result,
2766 match_info->offsets[0] - str_pos);
2767 done = (*eval) (match_info, result, user_data);
2768 str_pos = match_info->offsets[1];
2769 g_match_info_next (match_info, &tmp_error);
2771 g_match_info_free (match_info);
2772 if (tmp_error != NULL)
2774 g_propagate_error (error, tmp_error);
2775 g_string_free (result, TRUE);
2779 g_string_append_len (result, string + str_pos, string_len - str_pos);
2780 return g_string_free (result, FALSE);
2784 * g_regex_check_replacement:
2785 * @replacement: the replacement string
2786 * @has_references: (out) (allow-none): location to store information about
2787 * references in @replacement or %NULL
2788 * @error: location to store error
2790 * Checks whether @replacement is a valid replacement string
2791 * (see g_regex_replace()), i.e. that all escape sequences in
2794 * If @has_references is not %NULL then @replacement is checked
2795 * for pattern references. For instance, replacement text 'foo\n'
2796 * does not contain references and may be evaluated without information
2797 * about actual match, but '\0\1' (whole match followed by first
2798 * subpattern) requires valid #GMatchInfo object.
2800 * Returns: whether @replacement is a valid replacement string
2805 g_regex_check_replacement (const gchar *replacement,
2806 gboolean *has_references,
2812 list = split_replacement (replacement, &tmp);
2816 g_propagate_error (error, tmp);
2821 *has_references = interpolation_list_needs_match (list);
2823 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2829 * g_regex_escape_nul:
2830 * @string: the string to escape
2831 * @length: the length of @string
2833 * Escapes the nul characters in @string to "\x00". It can be used
2834 * to compile a regex with embedded nul characters.
2836 * For completeness, @length can be -1 for a nul-terminated string.
2837 * In this case the output string will be of course equal to @string.
2839 * Returns: a newly-allocated escaped string
2844 g_regex_escape_nul (const gchar *string,
2848 const gchar *p, *piece_start, *end;
2851 g_return_val_if_fail (string != NULL, NULL);
2854 return g_strdup (string);
2856 end = string + length;
2857 p = piece_start = string;
2858 escaped = g_string_sized_new (length + 1);
2866 if (p != piece_start)
2868 /* copy the previous piece. */
2869 g_string_append_len (escaped, piece_start, p - piece_start);
2871 if ((backslashes & 1) == 0)
2872 g_string_append_c (escaped, '\\');
2873 g_string_append_c (escaped, 'x');
2874 g_string_append_c (escaped, '0');
2875 g_string_append_c (escaped, '0');
2885 p = g_utf8_next_char (p);
2890 if (piece_start < end)
2891 g_string_append_len (escaped, piece_start, end - piece_start);
2893 return g_string_free (escaped, FALSE);
2897 * g_regex_escape_string:
2898 * @string: (array length=length): the string to escape
2899 * @length: the length of @string, or -1 if @string is nul-terminated
2901 * Escapes the special characters used for regular expressions
2902 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2903 * function is useful to dynamically generate regular expressions.
2905 * @string can contain nul characters that are replaced with "\0",
2906 * in this case remember to specify the correct length of @string
2909 * Returns: a newly-allocated escaped string
2914 g_regex_escape_string (const gchar *string,
2918 const char *p, *piece_start, *end;
2920 g_return_val_if_fail (string != NULL, NULL);
2923 length = strlen (string);
2925 end = string + length;
2926 p = piece_start = string;
2927 escaped = g_string_sized_new (length + 1);
2948 if (p != piece_start)
2949 /* copy the previous piece. */
2950 g_string_append_len (escaped, piece_start, p - piece_start);
2951 g_string_append_c (escaped, '\\');
2953 g_string_append_c (escaped, '0');
2955 g_string_append_c (escaped, *p);
2959 p = g_utf8_next_char (p);
2964 if (piece_start < end)
2965 g_string_append_len (escaped, piece_start, end - piece_start);
2967 return g_string_free (escaped, FALSE);