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");
239 case PCRE_ERROR_RECURSELOOP:
240 return _("recursion loop");
244 return _("unknown error");
248 translate_compile_error (gint *errcode, const gchar **errmsg)
250 /* Compile errors are created adding 100 to the error code returned
252 * If errcode is known we put the translatable error message in
253 * erromsg. If errcode is unknown we put the generic
254 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
255 * untranslated error message returned by PCRE.
256 * Note that there can be more PCRE errors with the same GRegexError
257 * and that some PCRE errors are useless for us.
263 case G_REGEX_ERROR_STRAY_BACKSLASH:
264 *errmsg = _("\\ at end of pattern");
266 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
267 *errmsg = _("\\c at end of pattern");
269 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
270 *errmsg = _("unrecognized character follows \\");
272 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
273 *errmsg = _("numbers out of order in {} quantifier");
275 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
276 *errmsg = _("number too big in {} quantifier");
278 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
279 *errmsg = _("missing terminating ] for character class");
281 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
282 *errmsg = _("invalid escape sequence in character class");
284 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
285 *errmsg = _("range out of order in character class");
287 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
288 *errmsg = _("nothing to repeat");
290 case 111: /* internal error: unexpected repeat */
291 *errcode = G_REGEX_ERROR_INTERNAL;
292 *errmsg = _("unexpected repeat");
294 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
295 *errmsg = _("unrecognized character after (? or (?-");
297 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
298 *errmsg = _("POSIX named classes are supported only within a class");
300 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
301 *errmsg = _("missing terminating )");
303 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
304 *errmsg = _("reference to non-existent subpattern");
306 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
307 *errmsg = _("missing ) after comment");
309 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
310 *errmsg = _("regular expression is too large");
312 case G_REGEX_ERROR_MEMORY_ERROR:
313 *errmsg = _("failed to get memory");
315 case 122: /* unmatched parentheses */
316 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
317 *errmsg = _(") without opening (");
319 case 123: /* internal error: code overflow */
320 *errcode = G_REGEX_ERROR_INTERNAL;
321 *errmsg = _("code overflow");
323 case 124: /* "unrecognized character after (?<\0 */
324 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
325 *errmsg = _("unrecognized character after (?<");
327 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
328 *errmsg = _("lookbehind assertion is not fixed length");
330 case G_REGEX_ERROR_MALFORMED_CONDITION:
331 *errmsg = _("malformed number or name after (?(");
333 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
334 *errmsg = _("conditional group contains more than two branches");
336 case G_REGEX_ERROR_ASSERTION_EXPECTED:
337 *errmsg = _("assertion expected after (?(");
340 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
341 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
342 * sequences here, '(?-54' would be an example for the second group.
344 *errmsg = _("(?R or (?[+-]digits must be followed by )");
346 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
347 *errmsg = _("unknown POSIX class name");
349 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
350 *errmsg = _("POSIX collating elements are not supported");
352 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
353 *errmsg = _("character value in \\x{...} sequence is too large");
355 case G_REGEX_ERROR_INVALID_CONDITION:
356 *errmsg = _("invalid condition (?(0)");
358 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
359 *errmsg = _("\\C not allowed in lookbehind assertion");
361 case 137: /* PCRE does not support \\L, \\l, \\N{name}, \\U, or \\u\0 */
362 /* A number of Perl escapes are not handled by PCRE.
363 * Therefore it explicitly raises ERR37.
365 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
366 *errmsg = _("escapes \\L, \\l, \\N{name}, \\U, and \\u are not supported");
368 case G_REGEX_ERROR_INFINITE_LOOP:
369 *errmsg = _("recursive call could loop indefinitely");
371 case 141: /* unrecognized character after (?P\0 */
372 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
373 *errmsg = _("unrecognized character after (?P");
375 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
376 *errmsg = _("missing terminator in subpattern name");
378 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
379 *errmsg = _("two named subpatterns have the same name");
381 case G_REGEX_ERROR_MALFORMED_PROPERTY:
382 *errmsg = _("malformed \\P or \\p sequence");
384 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
385 *errmsg = _("unknown property name after \\P or \\p");
387 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
388 *errmsg = _("subpattern name is too long (maximum 32 characters)");
390 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
391 *errmsg = _("too many named subpatterns (maximum 10,000)");
393 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
394 *errmsg = _("octal value is greater than \\377");
396 case 152: /* internal error: overran compiling workspace */
397 *errcode = G_REGEX_ERROR_INTERNAL;
398 *errmsg = _("overran compiling workspace");
400 case 153: /* internal error: previously-checked referenced subpattern not found */
401 *errcode = G_REGEX_ERROR_INTERNAL;
402 *errmsg = _("previously-checked referenced subpattern not found");
404 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
405 *errmsg = _("DEFINE group contains more than one branch");
407 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
408 *errmsg = _("inconsistent NEWLINE options");
410 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
411 *errmsg = _("\\g is not followed by a braced, angle-bracketed, or quoted name or "
412 "number, or by a plain number");
414 case G_REGEX_ERROR_INVALID_RELATIVE_REFERENCE:
415 *errmsg = _("a numbered reference must not be zero");
417 case G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_FORBIDDEN:
418 *errmsg = _("an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)");
420 case G_REGEX_ERROR_UNKNOWN_BACKTRACKING_CONTROL_VERB:
421 *errmsg = _("(*VERB) not recognized");
423 case G_REGEX_ERROR_NUMBER_TOO_BIG:
424 *errmsg = _("number is too bug");
426 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME:
427 *errmsg = _("missing subpattern name after (?&");
429 case G_REGEX_ERROR_MISSING_DIGIT:
430 *errmsg = _("digit expected after (?+");
432 case G_REGEX_ERROR_EXTRA_SUBPATTERN_NAME:
433 *errmsg = _("different names for subpatterns of the same number are not allowed");
435 case G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_REQUIRED:
436 *errmsg = _("(*MARK) must have an argument");
438 case G_REGEX_ERROR_INVALID_CONTROL_CHAR:
439 *errmsg = _( "\\c must be followed by an ASCII character");
441 case G_REGEX_ERROR_MISSING_NAME:
442 *errmsg = _("\\k is not followed by a braced, angle-bracketed, or quoted name");
444 case G_REGEX_ERROR_NOT_SUPPORTED_IN_CLASS:
445 *errmsg = _("\\N is not supported in a class");
447 case G_REGEX_ERROR_TOO_MANY_FORWARD_REFERENCES:
448 *errmsg = _("too many forward references");
450 case G_REGEX_ERROR_NAME_TOO_LONG:
451 *errmsg = _("name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)");
454 case 116: /* erroffset passed as NULL */
455 /* This should not happen as we never pass a NULL erroffset */
456 g_warning ("erroffset passed as NULL");
457 *errcode = G_REGEX_ERROR_COMPILE;
459 case 117: /* unknown option bit(s) set */
460 /* This should not happen as we check options before passing them
461 * to pcre_compile2() */
462 g_warning ("unknown option bit(s) set");
463 *errcode = G_REGEX_ERROR_COMPILE;
465 case 132: /* this version of PCRE is compiled without UTF support */
466 case 144: /* invalid UTF-8 string */
467 case 145: /* support for \\P, \\p, and \\X has not been compiled */
468 case 167: /* this version of PCRE is not compiled with Unicode property support */
469 case 173: /* disallowed Unicode code point (>= 0xd800 && <= 0xdfff) */
470 case 174: /* invalid UTF-16 string */
471 /* These errors should not happen as we are using an UTF-8 and UCP-enabled PCRE
472 * and we do not check if strings are valid */
473 case 164: /* ] is an invalid data character in JavaScript compatibility mode */
474 /* This should not happen as we don't use PCRE_JAVASCRIPT_COMPAT */
475 g_warning ("%s", *errmsg);
476 *errcode = G_REGEX_ERROR_COMPILE;
478 case 170: /* internal error: unknown opcode in find_fixedlength() */
479 *errcode = G_REGEX_ERROR_INTERNAL;
483 *errcode = G_REGEX_ERROR_COMPILE;
490 match_info_new (const GRegex *regex,
497 GMatchInfo *match_info;
500 string_len = strlen (string);
502 match_info = g_new0 (GMatchInfo, 1);
503 match_info->ref_count = 1;
504 match_info->regex = g_regex_ref ((GRegex *)regex);
505 match_info->string = string;
506 match_info->string_len = string_len;
507 match_info->matches = PCRE_ERROR_NOMATCH;
508 match_info->pos = start_position;
509 match_info->match_opts = match_options;
513 /* These values should be enough for most cases, if they are not
514 * enough g_regex_match_all_full() will expand them. */
515 match_info->n_offsets = 24;
516 match_info->n_workspace = 100;
517 match_info->workspace = g_new (gint, match_info->n_workspace);
522 pcre_fullinfo (regex->pcre_re, regex->extra,
523 PCRE_INFO_CAPTURECOUNT, &capture_count);
524 match_info->n_offsets = (capture_count + 1) * 3;
527 match_info->offsets = g_new0 (gint, match_info->n_offsets);
528 /* Set an invalid position for the previous match. */
529 match_info->offsets[0] = -1;
530 match_info->offsets[1] = -1;
536 * g_match_info_get_regex:
537 * @match_info: a #GMatchInfo
539 * Returns #GRegex object used in @match_info. It belongs to Glib
540 * and must not be freed. Use g_regex_ref() if you need to keep it
541 * after you free @match_info object.
543 * Returns: #GRegex object used in @match_info
548 g_match_info_get_regex (const GMatchInfo *match_info)
550 g_return_val_if_fail (match_info != NULL, NULL);
551 return match_info->regex;
555 * g_match_info_get_string:
556 * @match_info: a #GMatchInfo
558 * Returns the string searched with @match_info. This is the
559 * string passed to g_regex_match() or g_regex_replace() so
560 * you may not free it before calling this function.
562 * Returns: the string searched with @match_info
567 g_match_info_get_string (const GMatchInfo *match_info)
569 g_return_val_if_fail (match_info != NULL, NULL);
570 return match_info->string;
575 * @match_info: a #GMatchInfo
577 * Increases reference count of @match_info by 1.
579 * Returns: @match_info
584 g_match_info_ref (GMatchInfo *match_info)
586 g_return_val_if_fail (match_info != NULL, NULL);
587 g_atomic_int_inc (&match_info->ref_count);
592 * g_match_info_unref:
593 * @match_info: a #GMatchInfo
595 * Decreases reference count of @match_info by 1. When reference count drops
596 * to zero, it frees all the memory associated with the match_info structure.
601 g_match_info_unref (GMatchInfo *match_info)
603 if (g_atomic_int_dec_and_test (&match_info->ref_count))
605 g_regex_unref (match_info->regex);
606 g_free (match_info->offsets);
607 g_free (match_info->workspace);
614 * @match_info: (allow-none): a #GMatchInfo, or %NULL
616 * If @match_info is not %NULL, calls g_match_info_unref(); otherwise does
622 g_match_info_free (GMatchInfo *match_info)
624 if (match_info == NULL)
627 g_match_info_unref (match_info);
632 * @match_info: a #GMatchInfo structure
633 * @error: location to store the error occurring, or %NULL to ignore errors
635 * Scans for the next match using the same parameters of the previous
636 * call to g_regex_match_full() or g_regex_match() that returned
639 * The match is done on the string passed to the match function, so you
640 * cannot free it before calling this function.
642 * Returns: %TRUE is the string matched, %FALSE otherwise
647 g_match_info_next (GMatchInfo *match_info,
650 gint prev_match_start;
653 g_return_val_if_fail (match_info != NULL, FALSE);
654 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
655 g_return_val_if_fail (match_info->pos >= 0, FALSE);
657 prev_match_start = match_info->offsets[0];
658 prev_match_end = match_info->offsets[1];
660 if (match_info->pos > match_info->string_len)
662 /* we have reached the end of the string */
663 match_info->pos = -1;
664 match_info->matches = PCRE_ERROR_NOMATCH;
668 match_info->matches = pcre_exec (match_info->regex->pcre_re,
669 match_info->regex->extra,
671 match_info->string_len,
673 match_info->regex->match_opts | match_info->match_opts,
675 match_info->n_offsets);
676 if (IS_PCRE_ERROR (match_info->matches))
678 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
679 _("Error while matching regular expression %s: %s"),
680 match_info->regex->pattern, match_error (match_info->matches));
684 /* avoid infinite loops if the pattern is an empty string or something
686 if (match_info->pos == match_info->offsets[1])
688 if (match_info->pos > match_info->string_len)
690 /* we have reached the end of the string */
691 match_info->pos = -1;
692 match_info->matches = PCRE_ERROR_NOMATCH;
696 match_info->pos = NEXT_CHAR (match_info->regex,
697 &match_info->string[match_info->pos]) -
702 match_info->pos = match_info->offsets[1];
705 /* it's possible to get two identical matches when we are matching
706 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
707 * the string is "RegExTest" we have:
708 * - search at position 0: match from 0 to 0
709 * - search at position 1: match from 3 to 3
710 * - search at position 3: match from 3 to 3 (duplicate)
711 * - search at position 4: match from 5 to 5
712 * - search at position 5: match from 5 to 5 (duplicate)
713 * - search at position 6: no match -> stop
714 * so we have to ignore the duplicates.
715 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
716 if (match_info->matches >= 0 &&
717 prev_match_start == match_info->offsets[0] &&
718 prev_match_end == match_info->offsets[1])
720 /* ignore this match and search the next one */
721 return g_match_info_next (match_info, error);
724 return match_info->matches >= 0;
728 * g_match_info_matches:
729 * @match_info: a #GMatchInfo structure
731 * Returns whether the previous match operation succeeded.
733 * Returns: %TRUE if the previous match operation succeeded,
739 g_match_info_matches (const GMatchInfo *match_info)
741 g_return_val_if_fail (match_info != NULL, FALSE);
743 return match_info->matches >= 0;
747 * g_match_info_get_match_count:
748 * @match_info: a #GMatchInfo structure
750 * Retrieves the number of matched substrings (including substring 0,
751 * that is the whole matched text), so 1 is returned if the pattern
752 * has no substrings in it and 0 is returned if the match failed.
754 * If the last match was obtained using the DFA algorithm, that is
755 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
756 * count is not that of the number of capturing parentheses but that of
757 * the number of matched substrings.
759 * Returns: Number of matched substrings, or -1 if an error occurred
764 g_match_info_get_match_count (const GMatchInfo *match_info)
766 g_return_val_if_fail (match_info, -1);
768 if (match_info->matches == PCRE_ERROR_NOMATCH)
771 else if (match_info->matches < PCRE_ERROR_NOMATCH)
776 return match_info->matches;
780 * g_match_info_is_partial_match:
781 * @match_info: a #GMatchInfo structure
783 * Usually if the string passed to g_regex_match*() matches as far as
784 * it goes, but is too short to match the entire pattern, %FALSE is
785 * returned. There are circumstances where it might be helpful to
786 * distinguish this case from other cases in which there is no match.
788 * Consider, for example, an application where a human is required to
789 * type in data for a field with specific formatting requirements. An
790 * example might be a date in the form ddmmmyy, defined by the pattern
791 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
792 * If the application sees the user’s keystrokes one by one, and can
793 * check that what has been typed so far is potentially valid, it is
794 * able to raise an error as soon as a mistake is made.
796 * GRegex supports the concept of partial matching by means of the
797 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
798 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
799 * for a complete match, %FALSE otherwise. But, when these functions
800 * return %FALSE, you can check if the match was partial calling
801 * g_match_info_is_partial_match().
803 * When using partial matching you cannot use g_match_info_fetch*().
805 * Because of the way certain internal optimizations are implemented
806 * the partial matching algorithm cannot be used with all patterns.
807 * So repeated single characters such as "a{2,4}" and repeated single
808 * meta-sequences such as "\d+" are not permitted if the maximum number
809 * of occurrences is greater than one. Optional items such as "\d?"
810 * (where the maximum is one) are permitted. Quantifiers with any values
811 * are permitted after parentheses, so the invalid examples above can be
812 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
813 * for a pattern that does not conform to the restrictions, matching
814 * functions return an error.
816 * Returns: %TRUE if the match was partial, %FALSE otherwise
821 g_match_info_is_partial_match (const GMatchInfo *match_info)
823 g_return_val_if_fail (match_info != NULL, FALSE);
825 return match_info->matches == PCRE_ERROR_PARTIAL;
829 * g_match_info_expand_references:
830 * @match_info: (allow-none): a #GMatchInfo or %NULL
831 * @string_to_expand: the string to expand
832 * @error: location to store the error occurring, or %NULL to ignore errors
834 * Returns a new string containing the text in @string_to_expand with
835 * references and escape sequences expanded. References refer to the last
836 * match done with @string against @regex and have the same syntax used by
839 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
840 * passed to g_regex_new().
842 * The backreferences are extracted from the string passed to the match
843 * function, so you cannot call this function after freeing the string.
845 * @match_info may be %NULL in which case @string_to_expand must not
846 * contain references. For instance "foo\n" does not refer to an actual
847 * pattern and '\n' merely will be replaced with \n character,
848 * while to expand "\0" (whole match) one needs the result of a match.
849 * Use g_regex_check_replacement() to find out whether @string_to_expand
850 * contains references.
852 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
857 g_match_info_expand_references (const GMatchInfo *match_info,
858 const gchar *string_to_expand,
863 GError *tmp_error = NULL;
865 g_return_val_if_fail (string_to_expand != NULL, NULL);
866 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
868 list = split_replacement (string_to_expand, &tmp_error);
869 if (tmp_error != NULL)
871 g_propagate_error (error, tmp_error);
875 if (!match_info && interpolation_list_needs_match (list))
877 g_critical ("String '%s' contains references to the match, can't "
878 "expand references without GMatchInfo object",
883 result = g_string_sized_new (strlen (string_to_expand));
884 interpolate_replacement (match_info, result, list);
886 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
888 return g_string_free (result, FALSE);
892 * g_match_info_fetch:
893 * @match_info: #GMatchInfo structure
894 * @match_num: number of the sub expression
896 * Retrieves the text matching the @match_num<!-- -->'th capturing
897 * parentheses. 0 is the full text of the match, 1 is the first paren
898 * set, 2 the second, and so on.
900 * If @match_num is a valid sub pattern but it didn't match anything
901 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
902 * string is returned.
904 * If the match was obtained using the DFA algorithm, that is using
905 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
906 * string is not that of a set of parentheses but that of a matched
907 * substring. Substrings are matched in reverse order of length, so
908 * 0 is the longest match.
910 * The string is fetched from the string passed to the match function,
911 * so you cannot call this function after freeing the string.
913 * Returns: (allow-none): The matched substring, or %NULL if an error
914 * occurred. You have to free the string yourself
919 g_match_info_fetch (const GMatchInfo *match_info,
922 /* we cannot use pcre_get_substring() because it allocates the
923 * string using pcre_malloc(). */
927 g_return_val_if_fail (match_info != NULL, NULL);
928 g_return_val_if_fail (match_num >= 0, NULL);
930 /* match_num does not exist or it didn't matched, i.e. matching "b"
931 * against "(a)?b" then group 0 is empty. */
932 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
934 else if (start == -1)
935 match = g_strdup ("");
937 match = g_strndup (&match_info->string[start], end - start);
943 * g_match_info_fetch_pos:
944 * @match_info: #GMatchInfo structure
945 * @match_num: number of the sub expression
946 * @start_pos: (out) (allow-none): pointer to location where to store
947 * the start position, or %NULL
948 * @end_pos: (out) (allow-none): pointer to location where to store
949 * the end position, or %NULL
951 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
952 * parentheses. 0 is the full text of the match, 1 is the first
953 * paren set, 2 the second, and so on.
955 * If @match_num is a valid sub pattern but it didn't match anything
956 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
957 * and @end_pos are set to -1 and %TRUE is returned.
959 * If the match was obtained using the DFA algorithm, that is using
960 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
961 * position is not that of a set of parentheses but that of a matched
962 * substring. Substrings are matched in reverse order of length, so
963 * 0 is the longest match.
965 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
966 * the position cannot be fetched, @start_pos and @end_pos are left
972 g_match_info_fetch_pos (const GMatchInfo *match_info,
977 g_return_val_if_fail (match_info != NULL, FALSE);
978 g_return_val_if_fail (match_num >= 0, FALSE);
980 /* make sure the sub expression number they're requesting is less than
981 * the total number of sub expressions that were matched. */
982 if (match_num >= match_info->matches)
985 if (start_pos != NULL)
986 *start_pos = match_info->offsets[2 * match_num];
989 *end_pos = match_info->offsets[2 * match_num + 1];
995 * Returns number of first matched subpattern with name @name.
996 * There may be more than one in case when DUPNAMES is used,
997 * and not all subpatterns with that name match;
998 * pcre_get_stringnumber() does not work in that case.
1001 get_matched_substring_number (const GMatchInfo *match_info,
1005 gchar *first, *last;
1008 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
1009 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
1011 /* This code is copied from pcre_get.c: get_first_set() */
1012 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
1020 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
1022 gint n = (entry[0] << 8) + entry[1];
1023 if (match_info->offsets[n*2] >= 0)
1027 return (first[0] << 8) + first[1];
1031 * g_match_info_fetch_named:
1032 * @match_info: #GMatchInfo structure
1033 * @name: name of the subexpression
1035 * Retrieves the text matching the capturing parentheses named @name.
1037 * If @name is a valid sub pattern name but it didn't match anything
1038 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1039 * then an empty string is returned.
1041 * The string is fetched from the string passed to the match function,
1042 * so you cannot call this function after freeing the string.
1044 * Returns: (allow-none): The matched substring, or %NULL if an error
1045 * occurred. You have to free the string yourself
1050 g_match_info_fetch_named (const GMatchInfo *match_info,
1053 /* we cannot use pcre_get_named_substring() because it allocates the
1054 * string using pcre_malloc(). */
1057 g_return_val_if_fail (match_info != NULL, NULL);
1058 g_return_val_if_fail (name != NULL, NULL);
1060 num = get_matched_substring_number (match_info, name);
1064 return g_match_info_fetch (match_info, num);
1068 * g_match_info_fetch_named_pos:
1069 * @match_info: #GMatchInfo structure
1070 * @name: name of the subexpression
1071 * @start_pos: (out) (allow-none): pointer to location where to store
1072 * the start position, or %NULL
1073 * @end_pos: (out) (allow-none): pointer to location where to store
1074 * the end position, or %NULL
1076 * Retrieves the position in bytes of the capturing parentheses named @name.
1078 * If @name is a valid sub pattern name but it didn't match anything
1079 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1080 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
1082 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
1083 * If the position cannot be fetched, @start_pos and @end_pos
1084 * are left unchanged.
1089 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
1096 g_return_val_if_fail (match_info != NULL, FALSE);
1097 g_return_val_if_fail (name != NULL, FALSE);
1099 num = get_matched_substring_number (match_info, name);
1103 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1107 * g_match_info_fetch_all:
1108 * @match_info: a #GMatchInfo structure
1110 * Bundles up pointers to each of the matching substrings from a match
1111 * and stores them in an array of gchar pointers. The first element in
1112 * the returned array is the match number 0, i.e. the entire matched
1115 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1116 * "b" against "(a)?b") then an empty string is inserted.
1118 * If the last match was obtained using the DFA algorithm, that is using
1119 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1120 * strings are not that matched by sets of parentheses but that of the
1121 * matched substring. Substrings are matched in reverse order of length,
1122 * so the first one is the longest match.
1124 * The strings are fetched from the string passed to the match function,
1125 * so you cannot call this function after freeing the string.
1127 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1128 * It must be freed using g_strfreev(). If the previous match failed
1134 g_match_info_fetch_all (const GMatchInfo *match_info)
1136 /* we cannot use pcre_get_substring_list() because the returned value
1137 * isn't suitable for g_strfreev(). */
1141 g_return_val_if_fail (match_info != NULL, NULL);
1143 if (match_info->matches < 0)
1146 result = g_new (gchar *, match_info->matches + 1);
1147 for (i = 0; i < match_info->matches; i++)
1148 result[i] = g_match_info_fetch (match_info, i);
1158 g_regex_error_quark (void)
1160 static GQuark error_quark = 0;
1162 if (error_quark == 0)
1163 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1172 * Increases reference count of @regex by 1.
1179 g_regex_ref (GRegex *regex)
1181 g_return_val_if_fail (regex != NULL, NULL);
1182 g_atomic_int_inc (®ex->ref_count);
1190 * Decreases reference count of @regex by 1. When reference count drops
1191 * to zero, it frees all the memory associated with the regex structure.
1196 g_regex_unref (GRegex *regex)
1198 g_return_if_fail (regex != NULL);
1200 if (g_atomic_int_dec_and_test (®ex->ref_count))
1202 g_free (regex->pattern);
1203 if (regex->pcre_re != NULL)
1204 pcre_free (regex->pcre_re);
1205 if (regex->extra != NULL)
1206 pcre_free (regex->extra);
1213 * @pattern: the regular expression
1214 * @compile_options: compile options for the regular expression, or 0
1215 * @match_options: match options for the regular expression, or 0
1216 * @error: return location for a #GError
1218 * Compiles the regular expression to an internal form, and does
1219 * the initial setup of the #GRegex structure.
1221 * Returns: a #GRegex structure. Call g_regex_unref() when you
1227 g_regex_new (const gchar *pattern,
1228 GRegexCompileFlags compile_options,
1229 GRegexMatchFlags match_options,
1234 const gchar *errmsg;
1237 gboolean optimize = FALSE;
1238 static gsize initialised;
1239 unsigned long int pcre_compile_options;
1241 g_return_val_if_fail (pattern != NULL, NULL);
1242 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1243 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1244 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1246 if (g_once_init_enter (&initialised))
1251 pcre_config (PCRE_CONFIG_UTF8, &support);
1254 msg = N_("PCRE library is compiled without UTF8 support");
1255 g_critical ("%s", msg);
1256 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1260 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1263 msg = N_("PCRE library is compiled without UTF8 properties support");
1264 g_critical ("%s", msg);
1265 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1269 g_once_init_leave (&initialised, TRUE);
1272 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1273 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1274 if (compile_options & G_REGEX_OPTIMIZE)
1277 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1278 * instead uses UTF-8 only if required with PCRE_UTF8. */
1279 if (compile_options & G_REGEX_RAW)
1282 compile_options &= ~G_REGEX_RAW;
1287 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1288 match_options |= PCRE_NO_UTF8_CHECK;
1291 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1292 * not for the system one. */
1293 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1294 !(compile_options & G_REGEX_NEWLINE_LF))
1296 compile_options |= PCRE_NEWLINE_ANY;
1299 compile_options |= PCRE_UCP;
1301 /* compile the pattern */
1302 re = pcre_compile2 (pattern, compile_options, &errcode,
1303 &errmsg, &erroffset, NULL);
1305 /* if the compilation failed, set the error member and return
1311 /* Translate the PCRE error code to GRegexError and use a translated
1312 * error message if possible */
1313 translate_compile_error (&errcode, &errmsg);
1315 /* PCRE uses byte offsets but we want to show character offsets */
1316 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1318 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1319 _("Error while compiling regular "
1320 "expression %s at char %d: %s"),
1321 pattern, erroffset, errmsg);
1322 g_propagate_error (error, tmp_error);
1327 /* For options set at the beginning of the pattern, pcre puts them into
1328 * compile options, e.g. "(?i)foo" will make the pcre structure store
1329 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1330 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1331 compile_options = pcre_compile_options;
1333 if (!(compile_options & G_REGEX_DUPNAMES))
1335 gboolean jchanged = FALSE;
1336 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1338 compile_options |= G_REGEX_DUPNAMES;
1341 regex = g_new0 (GRegex, 1);
1342 regex->ref_count = 1;
1343 regex->pattern = g_strdup (pattern);
1344 regex->pcre_re = re;
1345 regex->compile_opts = compile_options;
1346 regex->match_opts = match_options;
1350 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1353 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1354 G_REGEX_ERROR_OPTIMIZE,
1355 _("Error while optimizing "
1356 "regular expression %s: %s"),
1359 g_propagate_error (error, tmp_error);
1361 g_regex_unref (regex);
1370 * g_regex_get_pattern:
1371 * @regex: a #GRegex structure
1373 * Gets the pattern string associated with @regex, i.e. a copy of
1374 * the string passed to g_regex_new().
1376 * Returns: the pattern of @regex
1381 g_regex_get_pattern (const GRegex *regex)
1383 g_return_val_if_fail (regex != NULL, NULL);
1385 return regex->pattern;
1389 * g_regex_get_max_backref:
1392 * Returns the number of the highest back reference
1393 * in the pattern, or 0 if the pattern does not contain
1396 * Returns: the number of the highest back reference
1401 g_regex_get_max_backref (const GRegex *regex)
1405 pcre_fullinfo (regex->pcre_re, regex->extra,
1406 PCRE_INFO_BACKREFMAX, &value);
1412 * g_regex_get_capture_count:
1415 * Returns the number of capturing subpatterns in the pattern.
1417 * Returns: the number of capturing subpatterns
1422 g_regex_get_capture_count (const GRegex *regex)
1426 pcre_fullinfo (regex->pcre_re, regex->extra,
1427 PCRE_INFO_CAPTURECOUNT, &value);
1433 * g_regex_get_compile_flags:
1436 * Returns the compile options that @regex was created with.
1438 * Returns: flags from #GRegexCompileFlags
1443 g_regex_get_compile_flags (const GRegex *regex)
1445 g_return_val_if_fail (regex != NULL, 0);
1447 return regex->compile_opts;
1451 * g_regex_get_match_flags:
1454 * Returns the match options that @regex was created with.
1456 * Returns: flags from #GRegexMatchFlags
1461 g_regex_get_match_flags (const GRegex *regex)
1463 g_return_val_if_fail (regex != NULL, 0);
1465 return regex->match_opts;
1469 * g_regex_match_simple:
1470 * @pattern: the regular expression
1471 * @string: the string to scan for matches
1472 * @compile_options: compile options for the regular expression, or 0
1473 * @match_options: match options, or 0
1475 * Scans for a match in @string for @pattern.
1477 * This function is equivalent to g_regex_match() but it does not
1478 * require to compile the pattern with g_regex_new(), avoiding some
1479 * lines of code when you need just to do a match without extracting
1480 * substrings, capture counts, and so on.
1482 * If this function is to be called on the same @pattern more than
1483 * once, it's more efficient to compile the pattern once with
1484 * g_regex_new() and then use g_regex_match().
1486 * Returns: %TRUE if the string matched, %FALSE otherwise
1491 g_regex_match_simple (const gchar *pattern,
1492 const gchar *string,
1493 GRegexCompileFlags compile_options,
1494 GRegexMatchFlags match_options)
1499 regex = g_regex_new (pattern, compile_options, 0, NULL);
1502 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1503 g_regex_unref (regex);
1509 * @regex: a #GRegex structure from g_regex_new()
1510 * @string: the string to scan for matches
1511 * @match_options: match options
1512 * @match_info: (out) (allow-none): pointer to location where to store
1513 * the #GMatchInfo, or %NULL if you do not need it
1515 * Scans for a match in string for the pattern in @regex.
1516 * The @match_options are combined with the match options specified
1517 * when the @regex structure was created, letting you have more
1518 * flexibility in reusing #GRegex structures.
1520 * A #GMatchInfo structure, used to get information on the match,
1521 * is stored in @match_info if not %NULL. Note that if @match_info
1522 * is not %NULL then it is created even if the function returns %FALSE,
1523 * i.e. you must free it regardless if regular expression actually matched.
1525 * To retrieve all the non-overlapping matches of the pattern in
1526 * string you can use g_match_info_next().
1530 * print_uppercase_words (const gchar *string)
1532 * /* Print all uppercase-only words. */
1534 * GMatchInfo *match_info;
1536 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1537 * g_regex_match (regex, string, 0, &match_info);
1538 * while (g_match_info_matches (match_info))
1540 * gchar *word = g_match_info_fetch (match_info, 0);
1541 * g_print ("Found: %s\n", word);
1543 * g_match_info_next (match_info, NULL);
1545 * g_match_info_free (match_info);
1546 * g_regex_unref (regex);
1550 * @string is not copied and is used in #GMatchInfo internally. If
1551 * you use any #GMatchInfo method (except g_match_info_free()) after
1552 * freeing or modifying @string then the behaviour is undefined.
1554 * Returns: %TRUE is the string matched, %FALSE otherwise
1559 g_regex_match (const GRegex *regex,
1560 const gchar *string,
1561 GRegexMatchFlags match_options,
1562 GMatchInfo **match_info)
1564 return g_regex_match_full (regex, string, -1, 0, match_options,
1569 * g_regex_match_full:
1570 * @regex: a #GRegex structure from g_regex_new()
1571 * @string: (array length=string_len): the string to scan for matches
1572 * @string_len: the length of @string, or -1 if @string is nul-terminated
1573 * @start_position: starting index of the string to match
1574 * @match_options: match options
1575 * @match_info: (out) (allow-none): pointer to location where to store
1576 * the #GMatchInfo, or %NULL if you do not need it
1577 * @error: location to store the error occurring, or %NULL to ignore errors
1579 * Scans for a match in string for the pattern in @regex.
1580 * The @match_options are combined with the match options specified
1581 * when the @regex structure was created, letting you have more
1582 * flexibility in reusing #GRegex structures.
1584 * Setting @start_position differs from just passing over a shortened
1585 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1586 * that begins with any kind of lookbehind assertion, such as "\b".
1588 * A #GMatchInfo structure, used to get information on the match, is
1589 * stored in @match_info if not %NULL. Note that if @match_info is
1590 * not %NULL then it is created even if the function returns %FALSE,
1591 * i.e. you must free it regardless if regular expression actually
1594 * @string is not copied and is used in #GMatchInfo internally. If
1595 * you use any #GMatchInfo method (except g_match_info_free()) after
1596 * freeing or modifying @string then the behaviour is undefined.
1598 * To retrieve all the non-overlapping matches of the pattern in
1599 * string you can use g_match_info_next().
1603 * print_uppercase_words (const gchar *string)
1605 * /* Print all uppercase-only words. */
1607 * GMatchInfo *match_info;
1608 * GError *error = NULL;
1610 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1611 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1612 * while (g_match_info_matches (match_info))
1614 * gchar *word = g_match_info_fetch (match_info, 0);
1615 * g_print ("Found: %s\n", word);
1617 * g_match_info_next (match_info, &error);
1619 * g_match_info_free (match_info);
1620 * g_regex_unref (regex);
1621 * if (error != NULL)
1623 * g_printerr ("Error while matching: %s\n", error->message);
1624 * g_error_free (error);
1629 * Returns: %TRUE is the string matched, %FALSE otherwise
1634 g_regex_match_full (const GRegex *regex,
1635 const gchar *string,
1637 gint start_position,
1638 GRegexMatchFlags match_options,
1639 GMatchInfo **match_info,
1645 g_return_val_if_fail (regex != NULL, FALSE);
1646 g_return_val_if_fail (string != NULL, FALSE);
1647 g_return_val_if_fail (start_position >= 0, FALSE);
1648 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1649 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1651 info = match_info_new (regex, string, string_len, start_position,
1652 match_options, FALSE);
1653 match_ok = g_match_info_next (info, error);
1654 if (match_info != NULL)
1657 g_match_info_free (info);
1663 * g_regex_match_all:
1664 * @regex: a #GRegex structure from g_regex_new()
1665 * @string: the string to scan for matches
1666 * @match_options: match options
1667 * @match_info: (out) (allow-none): pointer to location where to store
1668 * the #GMatchInfo, or %NULL if you do not need it
1670 * Using the standard algorithm for regular expression matching only
1671 * the longest match in the string is retrieved. This function uses
1672 * a different algorithm so it can retrieve all the possible matches.
1673 * For more documentation see g_regex_match_all_full().
1675 * A #GMatchInfo structure, used to get information on the match, is
1676 * stored in @match_info if not %NULL. Note that if @match_info is
1677 * not %NULL then it is created even if the function returns %FALSE,
1678 * i.e. you must free it regardless if regular expression actually
1681 * @string is not copied and is used in #GMatchInfo internally. If
1682 * you use any #GMatchInfo method (except g_match_info_free()) after
1683 * freeing or modifying @string then the behaviour is undefined.
1685 * Returns: %TRUE is the string matched, %FALSE otherwise
1690 g_regex_match_all (const GRegex *regex,
1691 const gchar *string,
1692 GRegexMatchFlags match_options,
1693 GMatchInfo **match_info)
1695 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1700 * g_regex_match_all_full:
1701 * @regex: a #GRegex structure from g_regex_new()
1702 * @string: (array length=string_len): the string to scan for matches
1703 * @string_len: the length of @string, or -1 if @string is nul-terminated
1704 * @start_position: starting index of the string to match
1705 * @match_options: match options
1706 * @match_info: (out) (allow-none): pointer to location where to store
1707 * the #GMatchInfo, or %NULL if you do not need it
1708 * @error: location to store the error occurring, or %NULL to ignore errors
1710 * Using the standard algorithm for regular expression matching only
1711 * the longest match in the string is retrieved, it is not possible
1712 * to obtain all the available matches. For instance matching
1713 * "<a> <b> <c>" against the pattern "<.*>"
1714 * you get "<a> <b> <c>".
1716 * This function uses a different algorithm (called DFA, i.e. deterministic
1717 * finite automaton), so it can retrieve all the possible matches, all
1718 * starting at the same point in the string. For instance matching
1719 * "<a> <b> <c>" against the pattern "<.*>"
1720 * you would obtain three matches: "<a> <b> <c>",
1721 * "<a> <b>" and "<a>".
1723 * The number of matched strings is retrieved using
1724 * g_match_info_get_match_count(). To obtain the matched strings and
1725 * their position you can use, respectively, g_match_info_fetch() and
1726 * g_match_info_fetch_pos(). Note that the strings are returned in
1727 * reverse order of length; that is, the longest matching string is
1730 * Note that the DFA algorithm is slower than the standard one and it
1731 * is not able to capture substrings, so backreferences do not work.
1733 * Setting @start_position differs from just passing over a shortened
1734 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1735 * that begins with any kind of lookbehind assertion, such as "\b".
1737 * A #GMatchInfo structure, used to get information on the match, is
1738 * stored in @match_info if not %NULL. Note that if @match_info is
1739 * not %NULL then it is created even if the function returns %FALSE,
1740 * i.e. you must free it regardless if regular expression actually
1743 * @string is not copied and is used in #GMatchInfo internally. If
1744 * you use any #GMatchInfo method (except g_match_info_free()) after
1745 * freeing or modifying @string then the behaviour is undefined.
1747 * Returns: %TRUE is the string matched, %FALSE otherwise
1752 g_regex_match_all_full (const GRegex *regex,
1753 const gchar *string,
1755 gint start_position,
1756 GRegexMatchFlags match_options,
1757 GMatchInfo **match_info,
1763 g_return_val_if_fail (regex != NULL, FALSE);
1764 g_return_val_if_fail (string != NULL, FALSE);
1765 g_return_val_if_fail (start_position >= 0, FALSE);
1766 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1767 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1769 info = match_info_new (regex, string, string_len, start_position,
1770 match_options, TRUE);
1776 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1777 info->string, info->string_len,
1779 regex->match_opts | match_options,
1780 info->offsets, info->n_offsets,
1781 info->workspace, info->n_workspace);
1782 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1784 /* info->workspace is too small. */
1785 info->n_workspace *= 2;
1786 info->workspace = g_realloc (info->workspace,
1787 info->n_workspace * sizeof (gint));
1790 else if (info->matches == 0)
1792 /* info->offsets is too small. */
1793 info->n_offsets *= 2;
1794 info->offsets = g_realloc (info->offsets,
1795 info->n_offsets * sizeof (gint));
1798 else if (IS_PCRE_ERROR (info->matches))
1800 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1801 _("Error while matching regular expression %s: %s"),
1802 regex->pattern, match_error (info->matches));
1806 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1809 if (match_info != NULL)
1812 g_match_info_free (info);
1814 return info->matches >= 0;
1818 * g_regex_get_string_number:
1819 * @regex: #GRegex structure
1820 * @name: name of the subexpression
1822 * Retrieves the number of the subexpression named @name.
1824 * Returns: The number of the subexpression or -1 if @name
1830 g_regex_get_string_number (const GRegex *regex,
1835 g_return_val_if_fail (regex != NULL, -1);
1836 g_return_val_if_fail (name != NULL, -1);
1838 num = pcre_get_stringnumber (regex->pcre_re, name);
1839 if (num == PCRE_ERROR_NOSUBSTRING)
1846 * g_regex_split_simple:
1847 * @pattern: the regular expression
1848 * @string: the string to scan for matches
1849 * @compile_options: compile options for the regular expression, or 0
1850 * @match_options: match options, or 0
1852 * Breaks the string on the pattern, and returns an array of
1853 * the tokens. If the pattern contains capturing parentheses,
1854 * then the text for each of the substrings will also be returned.
1855 * If the pattern does not match anywhere in the string, then the
1856 * whole string is returned as the first token.
1858 * This function is equivalent to g_regex_split() but it does
1859 * not require to compile the pattern with g_regex_new(), avoiding
1860 * some lines of code when you need just to do a split without
1861 * extracting substrings, capture counts, and so on.
1863 * If this function is to be called on the same @pattern more than
1864 * once, it's more efficient to compile the pattern once with
1865 * g_regex_new() and then use g_regex_split().
1867 * As a special case, the result of splitting the empty string ""
1868 * is an empty vector, not a vector containing a single string.
1869 * The reason for this special case is that being able to represent
1870 * a empty vector is typically more useful than consistent handling
1871 * of empty elements. If you do need to represent empty elements,
1872 * you'll need to check for the empty string before calling this
1875 * A pattern that can match empty strings splits @string into
1876 * separate characters wherever it matches the empty string between
1877 * characters. For example splitting "ab c" using as a separator
1878 * "\s*", you will get "a", "b" and "c".
1880 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1885 g_regex_split_simple (const gchar *pattern,
1886 const gchar *string,
1887 GRegexCompileFlags compile_options,
1888 GRegexMatchFlags match_options)
1893 regex = g_regex_new (pattern, compile_options, 0, NULL);
1897 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1898 g_regex_unref (regex);
1904 * @regex: a #GRegex structure
1905 * @string: the string to split with the pattern
1906 * @match_options: match time option flags
1908 * Breaks the string on the pattern, and returns an array of the tokens.
1909 * If the pattern contains capturing parentheses, then the text for each
1910 * of the substrings will also be returned. If the pattern does not match
1911 * anywhere in the string, then the whole string is returned as the first
1914 * As a special case, the result of splitting the empty string "" is an
1915 * empty vector, not a vector containing a single string. The reason for
1916 * this special case is that being able to represent a empty vector is
1917 * typically more useful than consistent handling of empty elements. If
1918 * you do need to represent empty elements, you'll need to check for the
1919 * empty string before calling this function.
1921 * A pattern that can match empty strings splits @string into separate
1922 * characters wherever it matches the empty string between characters.
1923 * For example splitting "ab c" using as a separator "\s*", you will get
1926 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1931 g_regex_split (const GRegex *regex,
1932 const gchar *string,
1933 GRegexMatchFlags match_options)
1935 return g_regex_split_full (regex, string, -1, 0,
1936 match_options, 0, NULL);
1940 * g_regex_split_full:
1941 * @regex: a #GRegex structure
1942 * @string: (array length=string_len): the string to split with the pattern
1943 * @string_len: the length of @string, or -1 if @string is nul-terminated
1944 * @start_position: starting index of the string to match
1945 * @match_options: match time option flags
1946 * @max_tokens: the maximum number of tokens to split @string into.
1947 * If this is less than 1, the string is split completely
1948 * @error: return location for a #GError
1950 * Breaks the string on the pattern, and returns an array of the tokens.
1951 * If the pattern contains capturing parentheses, then the text for each
1952 * of the substrings will also be returned. If the pattern does not match
1953 * anywhere in the string, then the whole string is returned as the first
1956 * As a special case, the result of splitting the empty string "" is an
1957 * empty vector, not a vector containing a single string. The reason for
1958 * this special case is that being able to represent a empty vector is
1959 * typically more useful than consistent handling of empty elements. If
1960 * you do need to represent empty elements, you'll need to check for the
1961 * empty string before calling this function.
1963 * A pattern that can match empty strings splits @string into separate
1964 * characters wherever it matches the empty string between characters.
1965 * For example splitting "ab c" using as a separator "\s*", you will get
1968 * Setting @start_position differs from just passing over a shortened
1969 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1970 * that begins with any kind of lookbehind assertion, such as "\b".
1972 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1977 g_regex_split_full (const GRegex *regex,
1978 const gchar *string,
1980 gint start_position,
1981 GRegexMatchFlags match_options,
1985 GError *tmp_error = NULL;
1986 GMatchInfo *match_info;
1991 /* position of the last separator. */
1992 gint last_separator_end;
1993 /* was the last match 0 bytes long? */
1994 gboolean last_match_is_empty;
1995 /* the returned array of char **s */
1996 gchar **string_list;
1998 g_return_val_if_fail (regex != NULL, NULL);
1999 g_return_val_if_fail (string != NULL, NULL);
2000 g_return_val_if_fail (start_position >= 0, NULL);
2001 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2002 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2004 if (max_tokens <= 0)
2005 max_tokens = G_MAXINT;
2008 string_len = strlen (string);
2010 /* zero-length string */
2011 if (string_len - start_position == 0)
2012 return g_new0 (gchar *, 1);
2014 if (max_tokens == 1)
2016 string_list = g_new0 (gchar *, 2);
2017 string_list[0] = g_strndup (&string[start_position],
2018 string_len - start_position);
2024 last_separator_end = start_position;
2025 last_match_is_empty = FALSE;
2027 match_ok = g_regex_match_full (regex, string, string_len, start_position,
2028 match_options, &match_info, &tmp_error);
2030 while (tmp_error == NULL)
2034 last_match_is_empty =
2035 (match_info->offsets[0] == match_info->offsets[1]);
2037 /* we need to skip empty separators at the same position of the end
2038 * of another separator. e.g. the string is "a b" and the separator
2039 * is " *", so from 1 to 2 we have a match and at position 2 we have
2040 * an empty match. */
2041 if (last_separator_end != match_info->offsets[1])
2046 token = g_strndup (string + last_separator_end,
2047 match_info->offsets[0] - last_separator_end);
2048 list = g_list_prepend (list, token);
2051 /* if there were substrings, these need to be added to
2053 match_count = g_match_info_get_match_count (match_info);
2054 if (match_count > 1)
2056 for (i = 1; i < match_count; i++)
2057 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
2063 /* if there was no match, copy to end of string. */
2064 if (!last_match_is_empty)
2066 gchar *token = g_strndup (string + last_separator_end,
2067 match_info->string_len - last_separator_end);
2068 list = g_list_prepend (list, token);
2070 /* no more tokens, end the loop. */
2074 /* -1 to leave room for the last part. */
2075 if (token_count >= max_tokens - 1)
2077 /* we have reached the maximum number of tokens, so we copy
2078 * the remaining part of the string. */
2079 if (last_match_is_empty)
2081 /* the last match was empty, so we have moved one char
2082 * after the real position to avoid empty matches at the
2084 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
2086 /* the if is needed in the case we have terminated the available
2087 * tokens, but we are at the end of the string, so there are no
2088 * characters left to copy. */
2089 if (string_len > match_info->pos)
2091 gchar *token = g_strndup (string + match_info->pos,
2092 string_len - match_info->pos);
2093 list = g_list_prepend (list, token);
2099 last_separator_end = match_info->pos;
2100 if (last_match_is_empty)
2101 /* if the last match was empty, g_match_info_next() has moved
2102 * forward to avoid infinite loops, but we still need to copy that
2104 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
2106 match_ok = g_match_info_next (match_info, &tmp_error);
2108 g_match_info_free (match_info);
2109 if (tmp_error != NULL)
2111 g_propagate_error (error, tmp_error);
2112 g_list_free_full (list, g_free);
2113 match_info->pos = -1;
2117 string_list = g_new (gchar *, g_list_length (list) + 1);
2119 for (last = g_list_last (list); last; last = g_list_previous (last))
2120 string_list[i++] = last->data;
2121 string_list[i] = NULL;
2130 REPL_TYPE_CHARACTER,
2131 REPL_TYPE_SYMBOLIC_REFERENCE,
2132 REPL_TYPE_NUMERIC_REFERENCE,
2133 REPL_TYPE_CHANGE_CASE
2138 CHANGE_CASE_NONE = 1 << 0,
2139 CHANGE_CASE_UPPER = 1 << 1,
2140 CHANGE_CASE_LOWER = 1 << 2,
2141 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2142 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2143 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2144 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2145 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2148 struct _InterpolationData
2154 ChangeCase change_case;
2158 free_interpolation_data (InterpolationData *data)
2160 g_free (data->text);
2164 static const gchar *
2165 expand_escape (const gchar *replacement,
2167 InterpolationData *data,
2172 const gchar *error_detail;
2174 GError *tmp_error = NULL;
2182 data->type = REPL_TYPE_CHARACTER;
2187 data->type = REPL_TYPE_CHARACTER;
2192 data->type = REPL_TYPE_CHARACTER;
2197 data->type = REPL_TYPE_CHARACTER;
2202 data->type = REPL_TYPE_CHARACTER;
2207 data->type = REPL_TYPE_CHARACTER;
2212 data->type = REPL_TYPE_CHARACTER;
2217 data->type = REPL_TYPE_CHARACTER;
2227 h = g_ascii_xdigit_value (*p);
2230 error_detail = _("hexadecimal digit or '}' expected");
2241 for (i = 0; i < 2; i++)
2243 h = g_ascii_xdigit_value (*p);
2246 error_detail = _("hexadecimal digit expected");
2253 data->type = REPL_TYPE_STRING;
2254 data->text = g_new0 (gchar, 8);
2255 g_unichar_to_utf8 (x, data->text);
2259 data->type = REPL_TYPE_CHANGE_CASE;
2260 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2264 data->type = REPL_TYPE_CHANGE_CASE;
2265 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2269 data->type = REPL_TYPE_CHANGE_CASE;
2270 data->change_case = CHANGE_CASE_LOWER;
2274 data->type = REPL_TYPE_CHANGE_CASE;
2275 data->change_case = CHANGE_CASE_UPPER;
2279 data->type = REPL_TYPE_CHANGE_CASE;
2280 data->change_case = CHANGE_CASE_NONE;
2286 error_detail = _("missing '<' in symbolic reference");
2295 error_detail = _("unfinished symbolic reference");
2302 error_detail = _("zero-length symbolic reference");
2305 if (g_ascii_isdigit (*q))
2310 h = g_ascii_digit_value (*q);
2313 error_detail = _("digit expected");
2322 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2329 if (!g_ascii_isalnum (*r))
2331 error_detail = _("illegal symbolic reference");
2338 data->text = g_strndup (q, p - q);
2339 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2344 /* if \0 is followed by a number is an octal number representing a
2345 * character, else it is a numeric reference. */
2346 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2349 p = g_utf8_next_char (p);
2362 for (i = 0; i < 3; i++)
2364 h = g_ascii_digit_value (*p);
2374 if (i == 2 && base == 10)
2380 if (base == 8 || i == 3)
2382 data->type = REPL_TYPE_STRING;
2383 data->text = g_new0 (gchar, 8);
2384 g_unichar_to_utf8 (x, data->text);
2388 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2393 error_detail = _("stray final '\\'");
2397 error_detail = _("unknown escape sequence");
2404 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2405 tmp_error = g_error_new (G_REGEX_ERROR,
2406 G_REGEX_ERROR_REPLACE,
2407 _("Error while parsing replacement "
2408 "text \"%s\" at char %lu: %s"),
2410 (gulong)(p - replacement),
2412 g_propagate_error (error, tmp_error);
2418 split_replacement (const gchar *replacement,
2422 InterpolationData *data;
2423 const gchar *p, *start;
2425 start = p = replacement;
2430 data = g_new0 (InterpolationData, 1);
2431 start = p = expand_escape (replacement, p, data, error);
2434 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2435 free_interpolation_data (data);
2439 list = g_list_prepend (list, data);
2444 if (*p == '\\' || *p == '\0')
2448 data = g_new0 (InterpolationData, 1);
2449 data->text = g_strndup (start, p - start);
2450 data->type = REPL_TYPE_STRING;
2451 list = g_list_prepend (list, data);
2457 return g_list_reverse (list);
2460 /* Change the case of c based on change_case. */
2461 #define CHANGE_CASE(c, change_case) \
2462 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2463 g_unichar_tolower (c) : \
2464 g_unichar_toupper (c))
2467 string_append (GString *string,
2469 ChangeCase *change_case)
2473 if (text[0] == '\0')
2476 if (*change_case == CHANGE_CASE_NONE)
2478 g_string_append (string, text);
2480 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2482 c = g_utf8_get_char (text);
2483 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2484 g_string_append (string, g_utf8_next_char (text));
2485 *change_case = CHANGE_CASE_NONE;
2489 while (*text != '\0')
2491 c = g_utf8_get_char (text);
2492 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2493 text = g_utf8_next_char (text);
2499 interpolate_replacement (const GMatchInfo *match_info,
2504 InterpolationData *idata;
2506 ChangeCase change_case = CHANGE_CASE_NONE;
2508 for (list = data; list; list = list->next)
2511 switch (idata->type)
2513 case REPL_TYPE_STRING:
2514 string_append (result, idata->text, &change_case);
2516 case REPL_TYPE_CHARACTER:
2517 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2518 if (change_case & CHANGE_CASE_SINGLE_MASK)
2519 change_case = CHANGE_CASE_NONE;
2521 case REPL_TYPE_NUMERIC_REFERENCE:
2522 match = g_match_info_fetch (match_info, idata->num);
2525 string_append (result, match, &change_case);
2529 case REPL_TYPE_SYMBOLIC_REFERENCE:
2530 match = g_match_info_fetch_named (match_info, idata->text);
2533 string_append (result, match, &change_case);
2537 case REPL_TYPE_CHANGE_CASE:
2538 change_case = idata->change_case;
2546 /* whether actual match_info is needed for replacement, i.e.
2547 * whether there are references
2550 interpolation_list_needs_match (GList *list)
2552 while (list != NULL)
2554 InterpolationData *data = list->data;
2556 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2557 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2570 * @regex: a #GRegex structure
2571 * @string: (array length=string_len): the string to perform matches against
2572 * @string_len: the length of @string, or -1 if @string is nul-terminated
2573 * @start_position: starting index of the string to match
2574 * @replacement: text to replace each match with
2575 * @match_options: options for the match
2576 * @error: location to store the error occurring, or %NULL to ignore errors
2578 * Replaces all occurrences of the pattern in @regex with the
2579 * replacement text. Backreferences of the form '\number' or
2580 * '\g<number>' in the replacement text are interpolated by the
2581 * number-th captured subexpression of the match, '\g<name>' refers
2582 * to the captured subexpression with the given name. '\0' refers to the
2583 * complete match, but '\0' followed by a number is the octal representation
2584 * of a character. To include a literal '\' in the replacement, write '\\'.
2585 * There are also escapes that changes the case of the following text:
2588 * <varlistentry><term>\l</term>
2590 * <para>Convert to lower case the next character</para>
2593 * <varlistentry><term>\u</term>
2595 * <para>Convert to upper case the next character</para>
2598 * <varlistentry><term>\L</term>
2600 * <para>Convert to lower case till \E</para>
2603 * <varlistentry><term>\U</term>
2605 * <para>Convert to upper case till \E</para>
2608 * <varlistentry><term>\E</term>
2610 * <para>End case modification</para>
2615 * If you do not need to use backreferences use g_regex_replace_literal().
2617 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2618 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2619 * you can use g_regex_replace_literal().
2621 * Setting @start_position differs from just passing over a shortened
2622 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2623 * begins with any kind of lookbehind assertion, such as "\b".
2625 * Returns: a newly allocated string containing the replacements
2630 g_regex_replace (const GRegex *regex,
2631 const gchar *string,
2633 gint start_position,
2634 const gchar *replacement,
2635 GRegexMatchFlags match_options,
2640 GError *tmp_error = NULL;
2642 g_return_val_if_fail (regex != NULL, NULL);
2643 g_return_val_if_fail (string != NULL, NULL);
2644 g_return_val_if_fail (start_position >= 0, NULL);
2645 g_return_val_if_fail (replacement != NULL, NULL);
2646 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2647 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2649 list = split_replacement (replacement, &tmp_error);
2650 if (tmp_error != NULL)
2652 g_propagate_error (error, tmp_error);
2656 result = g_regex_replace_eval (regex,
2657 string, string_len, start_position,
2659 interpolate_replacement,
2662 if (tmp_error != NULL)
2663 g_propagate_error (error, tmp_error);
2665 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2671 literal_replacement (const GMatchInfo *match_info,
2675 g_string_append (result, data);
2680 * g_regex_replace_literal:
2681 * @regex: a #GRegex structure
2682 * @string: (array length=string_len): the string to perform matches against
2683 * @string_len: the length of @string, or -1 if @string is nul-terminated
2684 * @start_position: starting index of the string to match
2685 * @replacement: text to replace each match with
2686 * @match_options: options for the match
2687 * @error: location to store the error occurring, or %NULL to ignore errors
2689 * Replaces all occurrences of the pattern in @regex with the
2690 * replacement text. @replacement is replaced literally, to
2691 * include backreferences use g_regex_replace().
2693 * Setting @start_position differs from just passing over a
2694 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2695 * case of a pattern that begins with any kind of lookbehind
2696 * assertion, such as "\b".
2698 * Returns: a newly allocated string containing the replacements
2703 g_regex_replace_literal (const GRegex *regex,
2704 const gchar *string,
2706 gint start_position,
2707 const gchar *replacement,
2708 GRegexMatchFlags match_options,
2711 g_return_val_if_fail (replacement != NULL, NULL);
2712 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2714 return g_regex_replace_eval (regex,
2715 string, string_len, start_position,
2717 literal_replacement,
2718 (gpointer)replacement,
2723 * g_regex_replace_eval:
2724 * @regex: a #GRegex structure from g_regex_new()
2725 * @string: (array length=string_len): string to perform matches against
2726 * @string_len: the length of @string, or -1 if @string is nul-terminated
2727 * @start_position: starting index of the string to match
2728 * @match_options: options for the match
2729 * @eval: a function to call for each match
2730 * @user_data: user data to pass to the function
2731 * @error: location to store the error occurring, or %NULL to ignore errors
2733 * Replaces occurrences of the pattern in regex with the output of
2734 * @eval for that occurrence.
2736 * Setting @start_position differs from just passing over a shortened
2737 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2738 * that begins with any kind of lookbehind assertion, such as "\b".
2740 * The following example uses g_regex_replace_eval() to replace multiple
2744 * eval_cb (const GMatchInfo *info,
2751 * match = g_match_info_fetch (info, 0);
2752 * r = g_hash_table_lookup ((GHashTable *)data, match);
2753 * g_string_append (res, r);
2765 * h = g_hash_table_new (g_str_hash, g_str_equal);
2767 * g_hash_table_insert (h, "1", "ONE");
2768 * g_hash_table_insert (h, "2", "TWO");
2769 * g_hash_table_insert (h, "3", "THREE");
2770 * g_hash_table_insert (h, "4", "FOUR");
2772 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2773 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2774 * g_hash_table_destroy (h);
2779 * Returns: a newly allocated string containing the replacements
2784 g_regex_replace_eval (const GRegex *regex,
2785 const gchar *string,
2787 gint start_position,
2788 GRegexMatchFlags match_options,
2789 GRegexEvalCallback eval,
2793 GMatchInfo *match_info;
2796 gboolean done = FALSE;
2797 GError *tmp_error = NULL;
2799 g_return_val_if_fail (regex != NULL, NULL);
2800 g_return_val_if_fail (string != NULL, NULL);
2801 g_return_val_if_fail (start_position >= 0, NULL);
2802 g_return_val_if_fail (eval != NULL, NULL);
2803 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2806 string_len = strlen (string);
2808 result = g_string_sized_new (string_len);
2810 /* run down the string making matches. */
2811 g_regex_match_full (regex, string, string_len, start_position,
2812 match_options, &match_info, &tmp_error);
2813 while (!done && g_match_info_matches (match_info))
2815 g_string_append_len (result,
2817 match_info->offsets[0] - str_pos);
2818 done = (*eval) (match_info, result, user_data);
2819 str_pos = match_info->offsets[1];
2820 g_match_info_next (match_info, &tmp_error);
2822 g_match_info_free (match_info);
2823 if (tmp_error != NULL)
2825 g_propagate_error (error, tmp_error);
2826 g_string_free (result, TRUE);
2830 g_string_append_len (result, string + str_pos, string_len - str_pos);
2831 return g_string_free (result, FALSE);
2835 * g_regex_check_replacement:
2836 * @replacement: the replacement string
2837 * @has_references: (out) (allow-none): location to store information about
2838 * references in @replacement or %NULL
2839 * @error: location to store error
2841 * Checks whether @replacement is a valid replacement string
2842 * (see g_regex_replace()), i.e. that all escape sequences in
2845 * If @has_references is not %NULL then @replacement is checked
2846 * for pattern references. For instance, replacement text 'foo\n'
2847 * does not contain references and may be evaluated without information
2848 * about actual match, but '\0\1' (whole match followed by first
2849 * subpattern) requires valid #GMatchInfo object.
2851 * Returns: whether @replacement is a valid replacement string
2856 g_regex_check_replacement (const gchar *replacement,
2857 gboolean *has_references,
2863 list = split_replacement (replacement, &tmp);
2867 g_propagate_error (error, tmp);
2872 *has_references = interpolation_list_needs_match (list);
2874 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2880 * g_regex_escape_nul:
2881 * @string: the string to escape
2882 * @length: the length of @string
2884 * Escapes the nul characters in @string to "\x00". It can be used
2885 * to compile a regex with embedded nul characters.
2887 * For completeness, @length can be -1 for a nul-terminated string.
2888 * In this case the output string will be of course equal to @string.
2890 * Returns: a newly-allocated escaped string
2895 g_regex_escape_nul (const gchar *string,
2899 const gchar *p, *piece_start, *end;
2902 g_return_val_if_fail (string != NULL, NULL);
2905 return g_strdup (string);
2907 end = string + length;
2908 p = piece_start = string;
2909 escaped = g_string_sized_new (length + 1);
2917 if (p != piece_start)
2919 /* copy the previous piece. */
2920 g_string_append_len (escaped, piece_start, p - piece_start);
2922 if ((backslashes & 1) == 0)
2923 g_string_append_c (escaped, '\\');
2924 g_string_append_c (escaped, 'x');
2925 g_string_append_c (escaped, '0');
2926 g_string_append_c (escaped, '0');
2936 p = g_utf8_next_char (p);
2941 if (piece_start < end)
2942 g_string_append_len (escaped, piece_start, end - piece_start);
2944 return g_string_free (escaped, FALSE);
2948 * g_regex_escape_string:
2949 * @string: (array length=length): the string to escape
2950 * @length: the length of @string, or -1 if @string is nul-terminated
2952 * Escapes the special characters used for regular expressions
2953 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2954 * function is useful to dynamically generate regular expressions.
2956 * @string can contain nul characters that are replaced with "\0",
2957 * in this case remember to specify the correct length of @string
2960 * Returns: a newly-allocated escaped string
2965 g_regex_escape_string (const gchar *string,
2969 const char *p, *piece_start, *end;
2971 g_return_val_if_fail (string != NULL, NULL);
2974 length = strlen (string);
2976 end = string + length;
2977 p = piece_start = string;
2978 escaped = g_string_sized_new (length + 1);
2999 if (p != piece_start)
3000 /* copy the previous piece. */
3001 g_string_append_len (escaped, piece_start, p - piece_start);
3002 g_string_append_c (escaped, '\\');
3004 g_string_append_c (escaped, '0');
3006 g_string_append_c (escaped, *p);
3010 p = g_utf8_next_char (p);
3015 if (piece_start < end)
3016 g_string_append_len (escaped, piece_start, end - piece_start);
3018 return g_string_free (escaped, FALSE);