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 \\");
270 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
271 *errmsg = _("numbers out of order in {} quantifier");
273 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
274 *errmsg = _("number too big in {} quantifier");
276 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
277 *errmsg = _("missing terminating ] for character class");
279 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
280 *errmsg = _("invalid escape sequence in character class");
282 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
283 *errmsg = _("range out of order in character class");
285 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
286 *errmsg = _("nothing to repeat");
288 case 111: /* internal error: unexpected repeat */
289 *errcode = G_REGEX_ERROR_INTERNAL;
290 *errmsg = _("unexpected repeat");
292 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
293 *errmsg = _("unrecognized character after (? or (?-");
295 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
296 *errmsg = _("POSIX named classes are supported only within a class");
298 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
299 *errmsg = _("missing terminating )");
301 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
302 *errmsg = _("reference to non-existent subpattern");
304 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
305 *errmsg = _("missing ) after comment");
307 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
308 *errmsg = _("regular expression is too large");
310 case G_REGEX_ERROR_MEMORY_ERROR:
311 *errmsg = _("failed to get memory");
313 case 122: /* unmatched parentheses */
314 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
315 *errmsg = _(") without opening (");
317 case 123: /* internal error: code overflow */
318 *errcode = G_REGEX_ERROR_INTERNAL;
319 *errmsg = _("code overflow");
321 case 124: /* "unrecognized character after (?<\0 */
322 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
323 *errmsg = _("unrecognized character after (?<");
325 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
326 *errmsg = _("lookbehind assertion is not fixed length");
328 case G_REGEX_ERROR_MALFORMED_CONDITION:
329 *errmsg = _("malformed number or name after (?(");
331 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
332 *errmsg = _("conditional group contains more than two branches");
334 case G_REGEX_ERROR_ASSERTION_EXPECTED:
335 *errmsg = _("assertion expected after (?(");
338 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
339 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
340 * sequences here, '(?-54' would be an example for the second group.
342 *errmsg = _("(?R or (?[+-]digits must be followed by )");
344 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
345 *errmsg = _("unknown POSIX class name");
347 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
348 *errmsg = _("POSIX collating elements are not supported");
350 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
351 *errmsg = _("character value in \\x{...} sequence is too large");
353 case G_REGEX_ERROR_INVALID_CONDITION:
354 *errmsg = _("invalid condition (?(0)");
356 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
357 *errmsg = _("\\C not allowed in lookbehind assertion");
359 case 137: /* PCRE does not support \\L, \\l, \\N{name}, \\U, or \\u\0 */
360 /* A number of Perl escapes are not handled by PCRE.
361 * Therefore it explicitly raises ERR37.
363 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
364 *errmsg = _("escapes \\L, \\l, \\N{name}, \\U, and \\u are not supported");
366 case G_REGEX_ERROR_INFINITE_LOOP:
367 *errmsg = _("recursive call could loop indefinitely");
369 case 141: /* unrecognized character after (?P\0 */
370 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
371 *errmsg = _("unrecognized character after (?P");
373 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
374 *errmsg = _("missing terminator in subpattern name");
376 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
377 *errmsg = _("two named subpatterns have the same name");
379 case G_REGEX_ERROR_MALFORMED_PROPERTY:
380 *errmsg = _("malformed \\P or \\p sequence");
382 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
383 *errmsg = _("unknown property name after \\P or \\p");
385 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
386 *errmsg = _("subpattern name is too long (maximum 32 characters)");
388 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
389 *errmsg = _("too many named subpatterns (maximum 10,000)");
391 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
392 *errmsg = _("octal value is greater than \\377");
394 case 152: /* internal error: overran compiling workspace */
395 *errcode = G_REGEX_ERROR_INTERNAL;
396 *errmsg = _("overran compiling workspace");
398 case 153: /* internal error: previously-checked referenced subpattern not found */
399 *errcode = G_REGEX_ERROR_INTERNAL;
400 *errmsg = _("previously-checked referenced subpattern not found");
402 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
403 *errmsg = _("DEFINE group contains more than one branch");
405 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
406 *errmsg = _("inconsistent NEWLINE options");
408 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
409 *errmsg = _("\\g is not followed by a braced, angle-bracketed, or quoted name or "
410 "number, or by a plain number");
412 case G_REGEX_ERROR_INVALID_RELATIVE_REFERENCE:
413 *errmsg = _("a numbered reference must not be zero");
415 case G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_FORBIDDEN:
416 *errmsg = _("an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)");
418 case G_REGEX_ERROR_UNKNOWN_BACKTRACKING_CONTROL_VERB:
419 *errmsg = _("(*VERB) not recognized");
421 case G_REGEX_ERROR_NUMBER_TOO_BIG:
422 *errmsg = _("number is too bug");
424 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME:
425 *errmsg = _("missing subpattern name after (?&");
427 case G_REGEX_ERROR_MISSING_DIGIT:
428 *errmsg = _("digit expected after (?+");
430 case G_REGEX_ERROR_EXTRA_SUBPATTERN_NAME:
431 *errmsg = _("different names for subpatterns of the same number are not allowed");
433 case G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_REQUIRED:
434 *errmsg = _("(*MARK) must have an argument");
436 case G_REGEX_ERROR_INVALID_CONTROL_CHAR:
437 *errmsg = _( "\\c must be followed by an ASCII character");
439 case G_REGEX_ERROR_MISSING_NAME:
440 *errmsg = _("\\k is not followed by a braced, angle-bracketed, or quoted name");
442 case G_REGEX_ERROR_NOT_SUPPORTED_IN_CLASS:
443 *errmsg = _("\\N is not supported in a class");
445 case G_REGEX_ERROR_TOO_MANY_FORWARD_REFERENCES:
446 *errmsg = _("too many forward references");
448 case G_REGEX_ERROR_NAME_TOO_LONG:
449 *errmsg = _("name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)");
452 case 116: /* erroffset passed as NULL */
453 /* This should not happen as we never pass a NULL erroffset */
454 g_warning ("erroffset passed as NULL");
455 *errcode = G_REGEX_ERROR_COMPILE;
457 case 117: /* unknown option bit(s) set */
458 /* This should not happen as we check options before passing them
459 * to pcre_compile2() */
460 g_warning ("unknown option bit(s) set");
461 *errcode = G_REGEX_ERROR_COMPILE;
463 case 132: /* this version of PCRE is compiled without UTF support */
464 case 144: /* invalid UTF-8 string */
465 case 145: /* support for \\P, \\p, and \\X has not been compiled */
466 case 167: /* this version of PCRE is not compiled with Unicode property support */
467 case 173: /* disallowed Unicode code point (>= 0xd800 && <= 0xdfff) */
468 case 174: /* invalid UTF-16 string */
469 /* These errors should not happen as we are using an UTF-8 and UCP-enabled PCRE
470 * and we do not check if strings are valid */
471 case 164: /* ] is an invalid data character in JavaScript compatibility mode */
472 /* This should not happen as we don't use PCRE_JAVASCRIPT_COMPAT */
473 g_warning ("%s", *errmsg);
474 *errcode = G_REGEX_ERROR_COMPILE;
476 case 170: /* internal error: unknown opcode in find_fixedlength() */
477 *errcode = G_REGEX_ERROR_INTERNAL;
481 *errcode = G_REGEX_ERROR_COMPILE;
488 match_info_new (const GRegex *regex,
495 GMatchInfo *match_info;
498 string_len = strlen (string);
500 match_info = g_new0 (GMatchInfo, 1);
501 match_info->ref_count = 1;
502 match_info->regex = g_regex_ref ((GRegex *)regex);
503 match_info->string = string;
504 match_info->string_len = string_len;
505 match_info->matches = PCRE_ERROR_NOMATCH;
506 match_info->pos = start_position;
507 match_info->match_opts = match_options;
511 /* These values should be enough for most cases, if they are not
512 * enough g_regex_match_all_full() will expand them. */
513 match_info->n_offsets = 24;
514 match_info->n_workspace = 100;
515 match_info->workspace = g_new (gint, match_info->n_workspace);
520 pcre_fullinfo (regex->pcre_re, regex->extra,
521 PCRE_INFO_CAPTURECOUNT, &capture_count);
522 match_info->n_offsets = (capture_count + 1) * 3;
525 match_info->offsets = g_new0 (gint, match_info->n_offsets);
526 /* Set an invalid position for the previous match. */
527 match_info->offsets[0] = -1;
528 match_info->offsets[1] = -1;
534 * g_match_info_get_regex:
535 * @match_info: a #GMatchInfo
537 * Returns #GRegex object used in @match_info. It belongs to Glib
538 * and must not be freed. Use g_regex_ref() if you need to keep it
539 * after you free @match_info object.
541 * Returns: #GRegex object used in @match_info
546 g_match_info_get_regex (const GMatchInfo *match_info)
548 g_return_val_if_fail (match_info != NULL, NULL);
549 return match_info->regex;
553 * g_match_info_get_string:
554 * @match_info: a #GMatchInfo
556 * Returns the string searched with @match_info. This is the
557 * string passed to g_regex_match() or g_regex_replace() so
558 * you may not free it before calling this function.
560 * Returns: the string searched with @match_info
565 g_match_info_get_string (const GMatchInfo *match_info)
567 g_return_val_if_fail (match_info != NULL, NULL);
568 return match_info->string;
573 * @match_info: a #GMatchInfo
575 * Increases reference count of @match_info by 1.
577 * Returns: @match_info
582 g_match_info_ref (GMatchInfo *match_info)
584 g_return_val_if_fail (match_info != NULL, NULL);
585 g_atomic_int_inc (&match_info->ref_count);
590 * g_match_info_unref:
591 * @match_info: a #GMatchInfo
593 * Decreases reference count of @match_info by 1. When reference count drops
594 * to zero, it frees all the memory associated with the match_info structure.
599 g_match_info_unref (GMatchInfo *match_info)
601 if (g_atomic_int_dec_and_test (&match_info->ref_count))
603 g_regex_unref (match_info->regex);
604 g_free (match_info->offsets);
605 g_free (match_info->workspace);
612 * @match_info: (allow-none): a #GMatchInfo, or %NULL
614 * If @match_info is not %NULL, calls g_match_info_unref(); otherwise does
620 g_match_info_free (GMatchInfo *match_info)
622 if (match_info == NULL)
625 g_match_info_unref (match_info);
630 * @match_info: a #GMatchInfo structure
631 * @error: location to store the error occurring, or %NULL to ignore errors
633 * Scans for the next match using the same parameters of the previous
634 * call to g_regex_match_full() or g_regex_match() that returned
637 * The match is done on the string passed to the match function, so you
638 * cannot free it before calling this function.
640 * Returns: %TRUE is the string matched, %FALSE otherwise
645 g_match_info_next (GMatchInfo *match_info,
648 gint prev_match_start;
651 g_return_val_if_fail (match_info != NULL, FALSE);
652 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
653 g_return_val_if_fail (match_info->pos >= 0, FALSE);
655 prev_match_start = match_info->offsets[0];
656 prev_match_end = match_info->offsets[1];
658 if (match_info->pos > match_info->string_len)
660 /* we have reached the end of the string */
661 match_info->pos = -1;
662 match_info->matches = PCRE_ERROR_NOMATCH;
666 match_info->matches = pcre_exec (match_info->regex->pcre_re,
667 match_info->regex->extra,
669 match_info->string_len,
671 match_info->regex->match_opts | match_info->match_opts,
673 match_info->n_offsets);
674 if (IS_PCRE_ERROR (match_info->matches))
676 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
677 _("Error while matching regular expression %s: %s"),
678 match_info->regex->pattern, match_error (match_info->matches));
682 /* avoid infinite loops if the pattern is an empty string or something
684 if (match_info->pos == match_info->offsets[1])
686 if (match_info->pos > match_info->string_len)
688 /* we have reached the end of the string */
689 match_info->pos = -1;
690 match_info->matches = PCRE_ERROR_NOMATCH;
694 match_info->pos = NEXT_CHAR (match_info->regex,
695 &match_info->string[match_info->pos]) -
700 match_info->pos = match_info->offsets[1];
703 /* it's possible to get two identical matches when we are matching
704 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
705 * the string is "RegExTest" we have:
706 * - search at position 0: match from 0 to 0
707 * - search at position 1: match from 3 to 3
708 * - search at position 3: match from 3 to 3 (duplicate)
709 * - search at position 4: match from 5 to 5
710 * - search at position 5: match from 5 to 5 (duplicate)
711 * - search at position 6: no match -> stop
712 * so we have to ignore the duplicates.
713 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
714 if (match_info->matches >= 0 &&
715 prev_match_start == match_info->offsets[0] &&
716 prev_match_end == match_info->offsets[1])
718 /* ignore this match and search the next one */
719 return g_match_info_next (match_info, error);
722 return match_info->matches >= 0;
726 * g_match_info_matches:
727 * @match_info: a #GMatchInfo structure
729 * Returns whether the previous match operation succeeded.
731 * Returns: %TRUE if the previous match operation succeeded,
737 g_match_info_matches (const GMatchInfo *match_info)
739 g_return_val_if_fail (match_info != NULL, FALSE);
741 return match_info->matches >= 0;
745 * g_match_info_get_match_count:
746 * @match_info: a #GMatchInfo structure
748 * Retrieves the number of matched substrings (including substring 0,
749 * that is the whole matched text), so 1 is returned if the pattern
750 * has no substrings in it and 0 is returned if the match failed.
752 * If the last match was obtained using the DFA algorithm, that is
753 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
754 * count is not that of the number of capturing parentheses but that of
755 * the number of matched substrings.
757 * Returns: Number of matched substrings, or -1 if an error occurred
762 g_match_info_get_match_count (const GMatchInfo *match_info)
764 g_return_val_if_fail (match_info, -1);
766 if (match_info->matches == PCRE_ERROR_NOMATCH)
769 else if (match_info->matches < PCRE_ERROR_NOMATCH)
774 return match_info->matches;
778 * g_match_info_is_partial_match:
779 * @match_info: a #GMatchInfo structure
781 * Usually if the string passed to g_regex_match*() matches as far as
782 * it goes, but is too short to match the entire pattern, %FALSE is
783 * returned. There are circumstances where it might be helpful to
784 * distinguish this case from other cases in which there is no match.
786 * Consider, for example, an application where a human is required to
787 * type in data for a field with specific formatting requirements. An
788 * example might be a date in the form ddmmmyy, defined by the pattern
789 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
790 * If the application sees the user’s keystrokes one by one, and can
791 * check that what has been typed so far is potentially valid, it is
792 * able to raise an error as soon as a mistake is made.
794 * GRegex supports the concept of partial matching by means of the
795 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
796 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
797 * for a complete match, %FALSE otherwise. But, when these functions
798 * return %FALSE, you can check if the match was partial calling
799 * g_match_info_is_partial_match().
801 * When using partial matching you cannot use g_match_info_fetch*().
803 * Because of the way certain internal optimizations are implemented
804 * the partial matching algorithm cannot be used with all patterns.
805 * So repeated single characters such as "a{2,4}" and repeated single
806 * meta-sequences such as "\d+" are not permitted if the maximum number
807 * of occurrences is greater than one. Optional items such as "\d?"
808 * (where the maximum is one) are permitted. Quantifiers with any values
809 * are permitted after parentheses, so the invalid examples above can be
810 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
811 * for a pattern that does not conform to the restrictions, matching
812 * functions return an error.
814 * Returns: %TRUE if the match was partial, %FALSE otherwise
819 g_match_info_is_partial_match (const GMatchInfo *match_info)
821 g_return_val_if_fail (match_info != NULL, FALSE);
823 return match_info->matches == PCRE_ERROR_PARTIAL;
827 * g_match_info_expand_references:
828 * @match_info: (allow-none): a #GMatchInfo or %NULL
829 * @string_to_expand: the string to expand
830 * @error: location to store the error occurring, or %NULL to ignore errors
832 * Returns a new string containing the text in @string_to_expand with
833 * references and escape sequences expanded. References refer to the last
834 * match done with @string against @regex and have the same syntax used by
837 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
838 * passed to g_regex_new().
840 * The backreferences are extracted from the string passed to the match
841 * function, so you cannot call this function after freeing the string.
843 * @match_info may be %NULL in which case @string_to_expand must not
844 * contain references. For instance "foo\n" does not refer to an actual
845 * pattern and '\n' merely will be replaced with \n character,
846 * while to expand "\0" (whole match) one needs the result of a match.
847 * Use g_regex_check_replacement() to find out whether @string_to_expand
848 * contains references.
850 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
855 g_match_info_expand_references (const GMatchInfo *match_info,
856 const gchar *string_to_expand,
861 GError *tmp_error = NULL;
863 g_return_val_if_fail (string_to_expand != NULL, NULL);
864 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
866 list = split_replacement (string_to_expand, &tmp_error);
867 if (tmp_error != NULL)
869 g_propagate_error (error, tmp_error);
873 if (!match_info && interpolation_list_needs_match (list))
875 g_critical ("String '%s' contains references to the match, can't "
876 "expand references without GMatchInfo object",
881 result = g_string_sized_new (strlen (string_to_expand));
882 interpolate_replacement (match_info, result, list);
884 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
886 return g_string_free (result, FALSE);
890 * g_match_info_fetch:
891 * @match_info: #GMatchInfo structure
892 * @match_num: number of the sub expression
894 * Retrieves the text matching the @match_num<!-- -->'th capturing
895 * parentheses. 0 is the full text of the match, 1 is the first paren
896 * set, 2 the second, and so on.
898 * If @match_num is a valid sub pattern but it didn't match anything
899 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
900 * string is returned.
902 * If the match was obtained using the DFA algorithm, that is using
903 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
904 * string is not that of a set of parentheses but that of a matched
905 * substring. Substrings are matched in reverse order of length, so
906 * 0 is the longest match.
908 * The string is fetched from the string passed to the match function,
909 * so you cannot call this function after freeing the string.
911 * Returns: (allow-none): The matched substring, or %NULL if an error
912 * occurred. You have to free the string yourself
917 g_match_info_fetch (const GMatchInfo *match_info,
920 /* we cannot use pcre_get_substring() because it allocates the
921 * string using pcre_malloc(). */
925 g_return_val_if_fail (match_info != NULL, NULL);
926 g_return_val_if_fail (match_num >= 0, NULL);
928 /* match_num does not exist or it didn't matched, i.e. matching "b"
929 * against "(a)?b" then group 0 is empty. */
930 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
932 else if (start == -1)
933 match = g_strdup ("");
935 match = g_strndup (&match_info->string[start], end - start);
941 * g_match_info_fetch_pos:
942 * @match_info: #GMatchInfo structure
943 * @match_num: number of the sub expression
944 * @start_pos: (out) (allow-none): pointer to location where to store
945 * the start position, or %NULL
946 * @end_pos: (out) (allow-none): pointer to location where to store
947 * the end position, or %NULL
949 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
950 * parentheses. 0 is the full text of the match, 1 is the first
951 * paren set, 2 the second, and so on.
953 * If @match_num is a valid sub pattern but it didn't match anything
954 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
955 * and @end_pos are set to -1 and %TRUE is returned.
957 * If the match was obtained using the DFA algorithm, that is using
958 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
959 * position is not that of a set of parentheses but that of a matched
960 * substring. Substrings are matched in reverse order of length, so
961 * 0 is the longest match.
963 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
964 * the position cannot be fetched, @start_pos and @end_pos are left
970 g_match_info_fetch_pos (const GMatchInfo *match_info,
975 g_return_val_if_fail (match_info != NULL, FALSE);
976 g_return_val_if_fail (match_num >= 0, FALSE);
978 /* make sure the sub expression number they're requesting is less than
979 * the total number of sub expressions that were matched. */
980 if (match_num >= match_info->matches)
983 if (start_pos != NULL)
984 *start_pos = match_info->offsets[2 * match_num];
987 *end_pos = match_info->offsets[2 * match_num + 1];
993 * Returns number of first matched subpattern with name @name.
994 * There may be more than one in case when DUPNAMES is used,
995 * and not all subpatterns with that name match;
996 * pcre_get_stringnumber() does not work in that case.
999 get_matched_substring_number (const GMatchInfo *match_info,
1003 gchar *first, *last;
1006 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
1007 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
1009 /* This code is copied from pcre_get.c: get_first_set() */
1010 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
1018 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
1020 gint n = (entry[0] << 8) + entry[1];
1021 if (match_info->offsets[n*2] >= 0)
1025 return (first[0] << 8) + first[1];
1029 * g_match_info_fetch_named:
1030 * @match_info: #GMatchInfo structure
1031 * @name: name of the subexpression
1033 * Retrieves the text matching the capturing parentheses named @name.
1035 * If @name is a valid sub pattern name but it didn't match anything
1036 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1037 * then an empty string is returned.
1039 * The string is fetched from the string passed to the match function,
1040 * so you cannot call this function after freeing the string.
1042 * Returns: (allow-none): The matched substring, or %NULL if an error
1043 * occurred. You have to free the string yourself
1048 g_match_info_fetch_named (const GMatchInfo *match_info,
1051 /* we cannot use pcre_get_named_substring() because it allocates the
1052 * string using pcre_malloc(). */
1055 g_return_val_if_fail (match_info != NULL, NULL);
1056 g_return_val_if_fail (name != NULL, NULL);
1058 num = get_matched_substring_number (match_info, name);
1062 return g_match_info_fetch (match_info, num);
1066 * g_match_info_fetch_named_pos:
1067 * @match_info: #GMatchInfo structure
1068 * @name: name of the subexpression
1069 * @start_pos: (out) (allow-none): pointer to location where to store
1070 * the start position, or %NULL
1071 * @end_pos: (out) (allow-none): pointer to location where to store
1072 * the end position, or %NULL
1074 * Retrieves the position in bytes of the capturing parentheses named @name.
1076 * If @name is a valid sub pattern name but it didn't match anything
1077 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1078 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
1080 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
1081 * If the position cannot be fetched, @start_pos and @end_pos
1082 * are left unchanged.
1087 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
1094 g_return_val_if_fail (match_info != NULL, FALSE);
1095 g_return_val_if_fail (name != NULL, FALSE);
1097 num = get_matched_substring_number (match_info, name);
1101 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1105 * g_match_info_fetch_all:
1106 * @match_info: a #GMatchInfo structure
1108 * Bundles up pointers to each of the matching substrings from a match
1109 * and stores them in an array of gchar pointers. The first element in
1110 * the returned array is the match number 0, i.e. the entire matched
1113 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1114 * "b" against "(a)?b") then an empty string is inserted.
1116 * If the last match was obtained using the DFA algorithm, that is using
1117 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1118 * strings are not that matched by sets of parentheses but that of the
1119 * matched substring. Substrings are matched in reverse order of length,
1120 * so the first one is the longest match.
1122 * The strings are fetched from the string passed to the match function,
1123 * so you cannot call this function after freeing the string.
1125 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1126 * It must be freed using g_strfreev(). If the previous match failed
1132 g_match_info_fetch_all (const GMatchInfo *match_info)
1134 /* we cannot use pcre_get_substring_list() because the returned value
1135 * isn't suitable for g_strfreev(). */
1139 g_return_val_if_fail (match_info != NULL, NULL);
1141 if (match_info->matches < 0)
1144 result = g_new (gchar *, match_info->matches + 1);
1145 for (i = 0; i < match_info->matches; i++)
1146 result[i] = g_match_info_fetch (match_info, i);
1156 g_regex_error_quark (void)
1158 static GQuark error_quark = 0;
1160 if (error_quark == 0)
1161 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1170 * Increases reference count of @regex by 1.
1177 g_regex_ref (GRegex *regex)
1179 g_return_val_if_fail (regex != NULL, NULL);
1180 g_atomic_int_inc (®ex->ref_count);
1188 * Decreases reference count of @regex by 1. When reference count drops
1189 * to zero, it frees all the memory associated with the regex structure.
1194 g_regex_unref (GRegex *regex)
1196 g_return_if_fail (regex != NULL);
1198 if (g_atomic_int_dec_and_test (®ex->ref_count))
1200 g_free (regex->pattern);
1201 if (regex->pcre_re != NULL)
1202 pcre_free (regex->pcre_re);
1203 if (regex->extra != NULL)
1204 pcre_free (regex->extra);
1211 * @pattern: the regular expression
1212 * @compile_options: compile options for the regular expression, or 0
1213 * @match_options: match options for the regular expression, or 0
1214 * @error: return location for a #GError
1216 * Compiles the regular expression to an internal form, and does
1217 * the initial setup of the #GRegex structure.
1219 * Returns: a #GRegex structure. Call g_regex_unref() when you
1225 g_regex_new (const gchar *pattern,
1226 GRegexCompileFlags compile_options,
1227 GRegexMatchFlags match_options,
1232 const gchar *errmsg;
1235 gboolean optimize = FALSE;
1236 static gsize initialised;
1237 unsigned long int pcre_compile_options;
1239 g_return_val_if_fail (pattern != NULL, NULL);
1240 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1241 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1242 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1244 if (g_once_init_enter (&initialised))
1249 pcre_config (PCRE_CONFIG_UTF8, &support);
1252 msg = N_("PCRE library is compiled without UTF8 support");
1253 g_critical ("%s", msg);
1254 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1258 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1261 msg = N_("PCRE library is compiled without UTF8 properties support");
1262 g_critical ("%s", msg);
1263 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1267 g_once_init_leave (&initialised, TRUE);
1270 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1271 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1272 if (compile_options & G_REGEX_OPTIMIZE)
1275 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1276 * instead uses UTF-8 only if required with PCRE_UTF8. */
1277 if (compile_options & G_REGEX_RAW)
1280 compile_options &= ~G_REGEX_RAW;
1285 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1286 match_options |= PCRE_NO_UTF8_CHECK;
1289 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1290 * not for the system one. */
1291 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1292 !(compile_options & G_REGEX_NEWLINE_LF))
1294 compile_options |= PCRE_NEWLINE_ANY;
1297 compile_options |= PCRE_UCP;
1299 /* compile the pattern */
1300 re = pcre_compile2 (pattern, compile_options, &errcode,
1301 &errmsg, &erroffset, NULL);
1303 /* if the compilation failed, set the error member and return
1309 /* Translate the PCRE error code to GRegexError and use a translated
1310 * error message if possible */
1311 translate_compile_error (&errcode, &errmsg);
1313 /* PCRE uses byte offsets but we want to show character offsets */
1314 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1316 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1317 _("Error while compiling regular "
1318 "expression %s at char %d: %s"),
1319 pattern, erroffset, errmsg);
1320 g_propagate_error (error, tmp_error);
1325 /* For options set at the beginning of the pattern, pcre puts them into
1326 * compile options, e.g. "(?i)foo" will make the pcre structure store
1327 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1328 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1329 compile_options = pcre_compile_options;
1331 if (!(compile_options & G_REGEX_DUPNAMES))
1333 gboolean jchanged = FALSE;
1334 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1336 compile_options |= G_REGEX_DUPNAMES;
1339 regex = g_new0 (GRegex, 1);
1340 regex->ref_count = 1;
1341 regex->pattern = g_strdup (pattern);
1342 regex->pcre_re = re;
1343 regex->compile_opts = compile_options;
1344 regex->match_opts = match_options;
1348 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1351 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1352 G_REGEX_ERROR_OPTIMIZE,
1353 _("Error while optimizing "
1354 "regular expression %s: %s"),
1357 g_propagate_error (error, tmp_error);
1359 g_regex_unref (regex);
1368 * g_regex_get_pattern:
1369 * @regex: a #GRegex structure
1371 * Gets the pattern string associated with @regex, i.e. a copy of
1372 * the string passed to g_regex_new().
1374 * Returns: the pattern of @regex
1379 g_regex_get_pattern (const GRegex *regex)
1381 g_return_val_if_fail (regex != NULL, NULL);
1383 return regex->pattern;
1387 * g_regex_get_max_backref:
1390 * Returns the number of the highest back reference
1391 * in the pattern, or 0 if the pattern does not contain
1394 * Returns: the number of the highest back reference
1399 g_regex_get_max_backref (const GRegex *regex)
1403 pcre_fullinfo (regex->pcre_re, regex->extra,
1404 PCRE_INFO_BACKREFMAX, &value);
1410 * g_regex_get_capture_count:
1413 * Returns the number of capturing subpatterns in the pattern.
1415 * Returns: the number of capturing subpatterns
1420 g_regex_get_capture_count (const GRegex *regex)
1424 pcre_fullinfo (regex->pcre_re, regex->extra,
1425 PCRE_INFO_CAPTURECOUNT, &value);
1431 * g_regex_get_compile_flags:
1434 * Returns the compile options that @regex was created with.
1436 * Returns: flags from #GRegexCompileFlags
1441 g_regex_get_compile_flags (const GRegex *regex)
1443 g_return_val_if_fail (regex != NULL, 0);
1445 return regex->compile_opts;
1449 * g_regex_get_match_flags:
1452 * Returns the match options that @regex was created with.
1454 * Returns: flags from #GRegexMatchFlags
1459 g_regex_get_match_flags (const GRegex *regex)
1461 g_return_val_if_fail (regex != NULL, 0);
1463 return regex->match_opts;
1467 * g_regex_match_simple:
1468 * @pattern: the regular expression
1469 * @string: the string to scan for matches
1470 * @compile_options: compile options for the regular expression, or 0
1471 * @match_options: match options, or 0
1473 * Scans for a match in @string for @pattern.
1475 * This function is equivalent to g_regex_match() but it does not
1476 * require to compile the pattern with g_regex_new(), avoiding some
1477 * lines of code when you need just to do a match without extracting
1478 * substrings, capture counts, and so on.
1480 * If this function is to be called on the same @pattern more than
1481 * once, it's more efficient to compile the pattern once with
1482 * g_regex_new() and then use g_regex_match().
1484 * Returns: %TRUE if the string matched, %FALSE otherwise
1489 g_regex_match_simple (const gchar *pattern,
1490 const gchar *string,
1491 GRegexCompileFlags compile_options,
1492 GRegexMatchFlags match_options)
1497 regex = g_regex_new (pattern, compile_options, 0, NULL);
1500 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1501 g_regex_unref (regex);
1507 * @regex: a #GRegex structure from g_regex_new()
1508 * @string: the string to scan for matches
1509 * @match_options: match options
1510 * @match_info: (out) (allow-none): pointer to location where to store
1511 * the #GMatchInfo, or %NULL if you do not need it
1513 * Scans for a match in string for the pattern in @regex.
1514 * The @match_options are combined with the match options specified
1515 * when the @regex structure was created, letting you have more
1516 * flexibility in reusing #GRegex structures.
1518 * A #GMatchInfo structure, used to get information on the match,
1519 * is stored in @match_info if not %NULL. Note that if @match_info
1520 * is not %NULL then it is created even if the function returns %FALSE,
1521 * i.e. you must free it regardless if regular expression actually matched.
1523 * To retrieve all the non-overlapping matches of the pattern in
1524 * string you can use g_match_info_next().
1528 * print_uppercase_words (const gchar *string)
1530 * /* Print all uppercase-only words. */
1532 * GMatchInfo *match_info;
1534 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1535 * g_regex_match (regex, string, 0, &match_info);
1536 * while (g_match_info_matches (match_info))
1538 * gchar *word = g_match_info_fetch (match_info, 0);
1539 * g_print ("Found: %s\n", word);
1541 * g_match_info_next (match_info, NULL);
1543 * g_match_info_free (match_info);
1544 * g_regex_unref (regex);
1548 * @string is not copied and is used in #GMatchInfo internally. If
1549 * you use any #GMatchInfo method (except g_match_info_free()) after
1550 * freeing or modifying @string then the behaviour is undefined.
1552 * Returns: %TRUE is the string matched, %FALSE otherwise
1557 g_regex_match (const GRegex *regex,
1558 const gchar *string,
1559 GRegexMatchFlags match_options,
1560 GMatchInfo **match_info)
1562 return g_regex_match_full (regex, string, -1, 0, match_options,
1567 * g_regex_match_full:
1568 * @regex: a #GRegex structure from g_regex_new()
1569 * @string: (array length=string_len): the string to scan for matches
1570 * @string_len: the length of @string, or -1 if @string is nul-terminated
1571 * @start_position: starting index of the string to match
1572 * @match_options: match options
1573 * @match_info: (out) (allow-none): pointer to location where to store
1574 * the #GMatchInfo, or %NULL if you do not need it
1575 * @error: location to store the error occurring, or %NULL to ignore errors
1577 * Scans for a match in string for the pattern in @regex.
1578 * The @match_options are combined with the match options specified
1579 * when the @regex structure was created, letting you have more
1580 * flexibility in reusing #GRegex structures.
1582 * Setting @start_position differs from just passing over a shortened
1583 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1584 * that begins with any kind of lookbehind assertion, such as "\b".
1586 * A #GMatchInfo structure, used to get information on the match, is
1587 * stored in @match_info if not %NULL. Note that if @match_info is
1588 * not %NULL then it is created even if the function returns %FALSE,
1589 * i.e. you must free it regardless if regular expression actually
1592 * @string is not copied and is used in #GMatchInfo internally. If
1593 * you use any #GMatchInfo method (except g_match_info_free()) after
1594 * freeing or modifying @string then the behaviour is undefined.
1596 * To retrieve all the non-overlapping matches of the pattern in
1597 * string you can use g_match_info_next().
1601 * print_uppercase_words (const gchar *string)
1603 * /* Print all uppercase-only words. */
1605 * GMatchInfo *match_info;
1606 * GError *error = NULL;
1608 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1609 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1610 * while (g_match_info_matches (match_info))
1612 * gchar *word = g_match_info_fetch (match_info, 0);
1613 * g_print ("Found: %s\n", word);
1615 * g_match_info_next (match_info, &error);
1617 * g_match_info_free (match_info);
1618 * g_regex_unref (regex);
1619 * if (error != NULL)
1621 * g_printerr ("Error while matching: %s\n", error->message);
1622 * g_error_free (error);
1627 * Returns: %TRUE is the string matched, %FALSE otherwise
1632 g_regex_match_full (const GRegex *regex,
1633 const gchar *string,
1635 gint start_position,
1636 GRegexMatchFlags match_options,
1637 GMatchInfo **match_info,
1643 g_return_val_if_fail (regex != NULL, FALSE);
1644 g_return_val_if_fail (string != NULL, FALSE);
1645 g_return_val_if_fail (start_position >= 0, FALSE);
1646 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1647 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1649 info = match_info_new (regex, string, string_len, start_position,
1650 match_options, FALSE);
1651 match_ok = g_match_info_next (info, error);
1652 if (match_info != NULL)
1655 g_match_info_free (info);
1661 * g_regex_match_all:
1662 * @regex: a #GRegex structure from g_regex_new()
1663 * @string: the string to scan for matches
1664 * @match_options: match options
1665 * @match_info: (out) (allow-none): pointer to location where to store
1666 * the #GMatchInfo, or %NULL if you do not need it
1668 * Using the standard algorithm for regular expression matching only
1669 * the longest match in the string is retrieved. This function uses
1670 * a different algorithm so it can retrieve all the possible matches.
1671 * For more documentation see g_regex_match_all_full().
1673 * A #GMatchInfo structure, used to get information on the match, is
1674 * stored in @match_info if not %NULL. Note that if @match_info is
1675 * not %NULL then it is created even if the function returns %FALSE,
1676 * i.e. you must free it regardless if regular expression actually
1679 * @string is not copied and is used in #GMatchInfo internally. If
1680 * you use any #GMatchInfo method (except g_match_info_free()) after
1681 * freeing or modifying @string then the behaviour is undefined.
1683 * Returns: %TRUE is the string matched, %FALSE otherwise
1688 g_regex_match_all (const GRegex *regex,
1689 const gchar *string,
1690 GRegexMatchFlags match_options,
1691 GMatchInfo **match_info)
1693 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1698 * g_regex_match_all_full:
1699 * @regex: a #GRegex structure from g_regex_new()
1700 * @string: (array length=string_len): the string to scan for matches
1701 * @string_len: the length of @string, or -1 if @string is nul-terminated
1702 * @start_position: starting index of the string to match
1703 * @match_options: match options
1704 * @match_info: (out) (allow-none): pointer to location where to store
1705 * the #GMatchInfo, or %NULL if you do not need it
1706 * @error: location to store the error occurring, or %NULL to ignore errors
1708 * Using the standard algorithm for regular expression matching only
1709 * the longest match in the string is retrieved, it is not possible
1710 * to obtain all the available matches. For instance matching
1711 * "<a> <b> <c>" against the pattern "<.*>"
1712 * you get "<a> <b> <c>".
1714 * This function uses a different algorithm (called DFA, i.e. deterministic
1715 * finite automaton), so it can retrieve all the possible matches, all
1716 * starting at the same point in the string. For instance matching
1717 * "<a> <b> <c>" against the pattern "<.*>"
1718 * you would obtain three matches: "<a> <b> <c>",
1719 * "<a> <b>" and "<a>".
1721 * The number of matched strings is retrieved using
1722 * g_match_info_get_match_count(). To obtain the matched strings and
1723 * their position you can use, respectively, g_match_info_fetch() and
1724 * g_match_info_fetch_pos(). Note that the strings are returned in
1725 * reverse order of length; that is, the longest matching string is
1728 * Note that the DFA algorithm is slower than the standard one and it
1729 * is not able to capture substrings, so backreferences do not work.
1731 * Setting @start_position differs from just passing over a shortened
1732 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1733 * that begins with any kind of lookbehind assertion, such as "\b".
1735 * A #GMatchInfo structure, used to get information on the match, is
1736 * stored in @match_info if not %NULL. Note that if @match_info is
1737 * not %NULL then it is created even if the function returns %FALSE,
1738 * i.e. you must free it regardless if regular expression actually
1741 * @string is not copied and is used in #GMatchInfo internally. If
1742 * you use any #GMatchInfo method (except g_match_info_free()) after
1743 * freeing or modifying @string then the behaviour is undefined.
1745 * Returns: %TRUE is the string matched, %FALSE otherwise
1750 g_regex_match_all_full (const GRegex *regex,
1751 const gchar *string,
1753 gint start_position,
1754 GRegexMatchFlags match_options,
1755 GMatchInfo **match_info,
1761 g_return_val_if_fail (regex != NULL, FALSE);
1762 g_return_val_if_fail (string != NULL, FALSE);
1763 g_return_val_if_fail (start_position >= 0, FALSE);
1764 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1765 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1767 info = match_info_new (regex, string, string_len, start_position,
1768 match_options, TRUE);
1774 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1775 info->string, info->string_len,
1777 regex->match_opts | match_options,
1778 info->offsets, info->n_offsets,
1779 info->workspace, info->n_workspace);
1780 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1782 /* info->workspace is too small. */
1783 info->n_workspace *= 2;
1784 info->workspace = g_realloc (info->workspace,
1785 info->n_workspace * sizeof (gint));
1788 else if (info->matches == 0)
1790 /* info->offsets is too small. */
1791 info->n_offsets *= 2;
1792 info->offsets = g_realloc (info->offsets,
1793 info->n_offsets * sizeof (gint));
1796 else if (IS_PCRE_ERROR (info->matches))
1798 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1799 _("Error while matching regular expression %s: %s"),
1800 regex->pattern, match_error (info->matches));
1804 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1807 if (match_info != NULL)
1810 g_match_info_free (info);
1812 return info->matches >= 0;
1816 * g_regex_get_string_number:
1817 * @regex: #GRegex structure
1818 * @name: name of the subexpression
1820 * Retrieves the number of the subexpression named @name.
1822 * Returns: The number of the subexpression or -1 if @name
1828 g_regex_get_string_number (const GRegex *regex,
1833 g_return_val_if_fail (regex != NULL, -1);
1834 g_return_val_if_fail (name != NULL, -1);
1836 num = pcre_get_stringnumber (regex->pcre_re, name);
1837 if (num == PCRE_ERROR_NOSUBSTRING)
1844 * g_regex_split_simple:
1845 * @pattern: the regular expression
1846 * @string: the string to scan for matches
1847 * @compile_options: compile options for the regular expression, or 0
1848 * @match_options: match options, or 0
1850 * Breaks the string on the pattern, and returns an array of
1851 * the tokens. If the pattern contains capturing parentheses,
1852 * then the text for each of the substrings will also be returned.
1853 * If the pattern does not match anywhere in the string, then the
1854 * whole string is returned as the first token.
1856 * This function is equivalent to g_regex_split() but it does
1857 * not require to compile the pattern with g_regex_new(), avoiding
1858 * some lines of code when you need just to do a split without
1859 * extracting substrings, capture counts, and so on.
1861 * If this function is to be called on the same @pattern more than
1862 * once, it's more efficient to compile the pattern once with
1863 * g_regex_new() and then use g_regex_split().
1865 * As a special case, the result of splitting the empty string ""
1866 * is an empty vector, not a vector containing a single string.
1867 * The reason for this special case is that being able to represent
1868 * a empty vector is typically more useful than consistent handling
1869 * of empty elements. If you do need to represent empty elements,
1870 * you'll need to check for the empty string before calling this
1873 * A pattern that can match empty strings splits @string into
1874 * separate characters wherever it matches the empty string between
1875 * characters. For example splitting "ab c" using as a separator
1876 * "\s*", you will get "a", "b" and "c".
1878 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1883 g_regex_split_simple (const gchar *pattern,
1884 const gchar *string,
1885 GRegexCompileFlags compile_options,
1886 GRegexMatchFlags match_options)
1891 regex = g_regex_new (pattern, compile_options, 0, NULL);
1895 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1896 g_regex_unref (regex);
1902 * @regex: a #GRegex structure
1903 * @string: the string to split with the pattern
1904 * @match_options: match time option flags
1906 * Breaks the string on the pattern, and returns an array of the tokens.
1907 * If the pattern contains capturing parentheses, then the text for each
1908 * of the substrings will also be returned. If the pattern does not match
1909 * anywhere in the string, then the whole string is returned as the first
1912 * As a special case, the result of splitting the empty string "" is an
1913 * empty vector, not a vector containing a single string. The reason for
1914 * this special case is that being able to represent a empty vector is
1915 * typically more useful than consistent handling of empty elements. If
1916 * you do need to represent empty elements, you'll need to check for the
1917 * empty string before calling this function.
1919 * A pattern that can match empty strings splits @string into separate
1920 * characters wherever it matches the empty string between characters.
1921 * For example splitting "ab c" using as a separator "\s*", you will get
1924 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1929 g_regex_split (const GRegex *regex,
1930 const gchar *string,
1931 GRegexMatchFlags match_options)
1933 return g_regex_split_full (regex, string, -1, 0,
1934 match_options, 0, NULL);
1938 * g_regex_split_full:
1939 * @regex: a #GRegex structure
1940 * @string: (array length=string_len): the string to split with the pattern
1941 * @string_len: the length of @string, or -1 if @string is nul-terminated
1942 * @start_position: starting index of the string to match
1943 * @match_options: match time option flags
1944 * @max_tokens: the maximum number of tokens to split @string into.
1945 * If this is less than 1, the string is split completely
1946 * @error: return location for a #GError
1948 * Breaks the string on the pattern, and returns an array of the tokens.
1949 * If the pattern contains capturing parentheses, then the text for each
1950 * of the substrings will also be returned. If the pattern does not match
1951 * anywhere in the string, then the whole string is returned as the first
1954 * As a special case, the result of splitting the empty string "" is an
1955 * empty vector, not a vector containing a single string. The reason for
1956 * this special case is that being able to represent a empty vector is
1957 * typically more useful than consistent handling of empty elements. If
1958 * you do need to represent empty elements, you'll need to check for the
1959 * empty string before calling this function.
1961 * A pattern that can match empty strings splits @string into separate
1962 * characters wherever it matches the empty string between characters.
1963 * For example splitting "ab c" using as a separator "\s*", you will get
1966 * Setting @start_position differs from just passing over a shortened
1967 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1968 * that begins with any kind of lookbehind assertion, such as "\b".
1970 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1975 g_regex_split_full (const GRegex *regex,
1976 const gchar *string,
1978 gint start_position,
1979 GRegexMatchFlags match_options,
1983 GError *tmp_error = NULL;
1984 GMatchInfo *match_info;
1989 /* position of the last separator. */
1990 gint last_separator_end;
1991 /* was the last match 0 bytes long? */
1992 gboolean last_match_is_empty;
1993 /* the returned array of char **s */
1994 gchar **string_list;
1996 g_return_val_if_fail (regex != NULL, NULL);
1997 g_return_val_if_fail (string != NULL, NULL);
1998 g_return_val_if_fail (start_position >= 0, NULL);
1999 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2000 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2002 if (max_tokens <= 0)
2003 max_tokens = G_MAXINT;
2006 string_len = strlen (string);
2008 /* zero-length string */
2009 if (string_len - start_position == 0)
2010 return g_new0 (gchar *, 1);
2012 if (max_tokens == 1)
2014 string_list = g_new0 (gchar *, 2);
2015 string_list[0] = g_strndup (&string[start_position],
2016 string_len - start_position);
2022 last_separator_end = start_position;
2023 last_match_is_empty = FALSE;
2025 match_ok = g_regex_match_full (regex, string, string_len, start_position,
2026 match_options, &match_info, &tmp_error);
2028 while (tmp_error == NULL)
2032 last_match_is_empty =
2033 (match_info->offsets[0] == match_info->offsets[1]);
2035 /* we need to skip empty separators at the same position of the end
2036 * of another separator. e.g. the string is "a b" and the separator
2037 * is " *", so from 1 to 2 we have a match and at position 2 we have
2038 * an empty match. */
2039 if (last_separator_end != match_info->offsets[1])
2044 token = g_strndup (string + last_separator_end,
2045 match_info->offsets[0] - last_separator_end);
2046 list = g_list_prepend (list, token);
2049 /* if there were substrings, these need to be added to
2051 match_count = g_match_info_get_match_count (match_info);
2052 if (match_count > 1)
2054 for (i = 1; i < match_count; i++)
2055 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
2061 /* if there was no match, copy to end of string. */
2062 if (!last_match_is_empty)
2064 gchar *token = g_strndup (string + last_separator_end,
2065 match_info->string_len - last_separator_end);
2066 list = g_list_prepend (list, token);
2068 /* no more tokens, end the loop. */
2072 /* -1 to leave room for the last part. */
2073 if (token_count >= max_tokens - 1)
2075 /* we have reached the maximum number of tokens, so we copy
2076 * the remaining part of the string. */
2077 if (last_match_is_empty)
2079 /* the last match was empty, so we have moved one char
2080 * after the real position to avoid empty matches at the
2082 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
2084 /* the if is needed in the case we have terminated the available
2085 * tokens, but we are at the end of the string, so there are no
2086 * characters left to copy. */
2087 if (string_len > match_info->pos)
2089 gchar *token = g_strndup (string + match_info->pos,
2090 string_len - match_info->pos);
2091 list = g_list_prepend (list, token);
2097 last_separator_end = match_info->pos;
2098 if (last_match_is_empty)
2099 /* if the last match was empty, g_match_info_next() has moved
2100 * forward to avoid infinite loops, but we still need to copy that
2102 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
2104 match_ok = g_match_info_next (match_info, &tmp_error);
2106 g_match_info_free (match_info);
2107 if (tmp_error != NULL)
2109 g_propagate_error (error, tmp_error);
2110 g_list_free_full (list, g_free);
2111 match_info->pos = -1;
2115 string_list = g_new (gchar *, g_list_length (list) + 1);
2117 for (last = g_list_last (list); last; last = g_list_previous (last))
2118 string_list[i++] = last->data;
2119 string_list[i] = NULL;
2128 REPL_TYPE_CHARACTER,
2129 REPL_TYPE_SYMBOLIC_REFERENCE,
2130 REPL_TYPE_NUMERIC_REFERENCE,
2131 REPL_TYPE_CHANGE_CASE
2136 CHANGE_CASE_NONE = 1 << 0,
2137 CHANGE_CASE_UPPER = 1 << 1,
2138 CHANGE_CASE_LOWER = 1 << 2,
2139 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2140 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2141 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2142 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2143 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2146 struct _InterpolationData
2152 ChangeCase change_case;
2156 free_interpolation_data (InterpolationData *data)
2158 g_free (data->text);
2162 static const gchar *
2163 expand_escape (const gchar *replacement,
2165 InterpolationData *data,
2170 const gchar *error_detail;
2172 GError *tmp_error = NULL;
2180 data->type = REPL_TYPE_CHARACTER;
2185 data->type = REPL_TYPE_CHARACTER;
2190 data->type = REPL_TYPE_CHARACTER;
2195 data->type = REPL_TYPE_CHARACTER;
2200 data->type = REPL_TYPE_CHARACTER;
2205 data->type = REPL_TYPE_CHARACTER;
2210 data->type = REPL_TYPE_CHARACTER;
2215 data->type = REPL_TYPE_CHARACTER;
2225 h = g_ascii_xdigit_value (*p);
2228 error_detail = _("hexadecimal digit or '}' expected");
2239 for (i = 0; i < 2; i++)
2241 h = g_ascii_xdigit_value (*p);
2244 error_detail = _("hexadecimal digit expected");
2251 data->type = REPL_TYPE_STRING;
2252 data->text = g_new0 (gchar, 8);
2253 g_unichar_to_utf8 (x, data->text);
2257 data->type = REPL_TYPE_CHANGE_CASE;
2258 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2262 data->type = REPL_TYPE_CHANGE_CASE;
2263 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2267 data->type = REPL_TYPE_CHANGE_CASE;
2268 data->change_case = CHANGE_CASE_LOWER;
2272 data->type = REPL_TYPE_CHANGE_CASE;
2273 data->change_case = CHANGE_CASE_UPPER;
2277 data->type = REPL_TYPE_CHANGE_CASE;
2278 data->change_case = CHANGE_CASE_NONE;
2284 error_detail = _("missing '<' in symbolic reference");
2293 error_detail = _("unfinished symbolic reference");
2300 error_detail = _("zero-length symbolic reference");
2303 if (g_ascii_isdigit (*q))
2308 h = g_ascii_digit_value (*q);
2311 error_detail = _("digit expected");
2320 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2327 if (!g_ascii_isalnum (*r))
2329 error_detail = _("illegal symbolic reference");
2336 data->text = g_strndup (q, p - q);
2337 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2342 /* if \0 is followed by a number is an octal number representing a
2343 * character, else it is a numeric reference. */
2344 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2347 p = g_utf8_next_char (p);
2360 for (i = 0; i < 3; i++)
2362 h = g_ascii_digit_value (*p);
2372 if (i == 2 && base == 10)
2378 if (base == 8 || i == 3)
2380 data->type = REPL_TYPE_STRING;
2381 data->text = g_new0 (gchar, 8);
2382 g_unichar_to_utf8 (x, data->text);
2386 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2391 error_detail = _("stray final '\\'");
2395 error_detail = _("unknown escape sequence");
2402 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2403 tmp_error = g_error_new (G_REGEX_ERROR,
2404 G_REGEX_ERROR_REPLACE,
2405 _("Error while parsing replacement "
2406 "text \"%s\" at char %lu: %s"),
2408 (gulong)(p - replacement),
2410 g_propagate_error (error, tmp_error);
2416 split_replacement (const gchar *replacement,
2420 InterpolationData *data;
2421 const gchar *p, *start;
2423 start = p = replacement;
2428 data = g_new0 (InterpolationData, 1);
2429 start = p = expand_escape (replacement, p, data, error);
2432 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2433 free_interpolation_data (data);
2437 list = g_list_prepend (list, data);
2442 if (*p == '\\' || *p == '\0')
2446 data = g_new0 (InterpolationData, 1);
2447 data->text = g_strndup (start, p - start);
2448 data->type = REPL_TYPE_STRING;
2449 list = g_list_prepend (list, data);
2455 return g_list_reverse (list);
2458 /* Change the case of c based on change_case. */
2459 #define CHANGE_CASE(c, change_case) \
2460 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2461 g_unichar_tolower (c) : \
2462 g_unichar_toupper (c))
2465 string_append (GString *string,
2467 ChangeCase *change_case)
2471 if (text[0] == '\0')
2474 if (*change_case == CHANGE_CASE_NONE)
2476 g_string_append (string, text);
2478 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2480 c = g_utf8_get_char (text);
2481 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2482 g_string_append (string, g_utf8_next_char (text));
2483 *change_case = CHANGE_CASE_NONE;
2487 while (*text != '\0')
2489 c = g_utf8_get_char (text);
2490 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2491 text = g_utf8_next_char (text);
2497 interpolate_replacement (const GMatchInfo *match_info,
2502 InterpolationData *idata;
2504 ChangeCase change_case = CHANGE_CASE_NONE;
2506 for (list = data; list; list = list->next)
2509 switch (idata->type)
2511 case REPL_TYPE_STRING:
2512 string_append (result, idata->text, &change_case);
2514 case REPL_TYPE_CHARACTER:
2515 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2516 if (change_case & CHANGE_CASE_SINGLE_MASK)
2517 change_case = CHANGE_CASE_NONE;
2519 case REPL_TYPE_NUMERIC_REFERENCE:
2520 match = g_match_info_fetch (match_info, idata->num);
2523 string_append (result, match, &change_case);
2527 case REPL_TYPE_SYMBOLIC_REFERENCE:
2528 match = g_match_info_fetch_named (match_info, idata->text);
2531 string_append (result, match, &change_case);
2535 case REPL_TYPE_CHANGE_CASE:
2536 change_case = idata->change_case;
2544 /* whether actual match_info is needed for replacement, i.e.
2545 * whether there are references
2548 interpolation_list_needs_match (GList *list)
2550 while (list != NULL)
2552 InterpolationData *data = list->data;
2554 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2555 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2568 * @regex: a #GRegex structure
2569 * @string: (array length=string_len): the string to perform matches against
2570 * @string_len: the length of @string, or -1 if @string is nul-terminated
2571 * @start_position: starting index of the string to match
2572 * @replacement: text to replace each match with
2573 * @match_options: options for the match
2574 * @error: location to store the error occurring, or %NULL to ignore errors
2576 * Replaces all occurrences of the pattern in @regex with the
2577 * replacement text. Backreferences of the form '\number' or
2578 * '\g<number>' in the replacement text are interpolated by the
2579 * number-th captured subexpression of the match, '\g<name>' refers
2580 * to the captured subexpression with the given name. '\0' refers to the
2581 * complete match, but '\0' followed by a number is the octal representation
2582 * of a character. To include a literal '\' in the replacement, write '\\'.
2583 * There are also escapes that changes the case of the following text:
2586 * <varlistentry><term>\l</term>
2588 * <para>Convert to lower case the next character</para>
2591 * <varlistentry><term>\u</term>
2593 * <para>Convert to upper case the next character</para>
2596 * <varlistentry><term>\L</term>
2598 * <para>Convert to lower case till \E</para>
2601 * <varlistentry><term>\U</term>
2603 * <para>Convert to upper case till \E</para>
2606 * <varlistentry><term>\E</term>
2608 * <para>End case modification</para>
2613 * If you do not need to use backreferences use g_regex_replace_literal().
2615 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2616 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2617 * you can use g_regex_replace_literal().
2619 * Setting @start_position differs from just passing over a shortened
2620 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2621 * begins with any kind of lookbehind assertion, such as "\b".
2623 * Returns: a newly allocated string containing the replacements
2628 g_regex_replace (const GRegex *regex,
2629 const gchar *string,
2631 gint start_position,
2632 const gchar *replacement,
2633 GRegexMatchFlags match_options,
2638 GError *tmp_error = NULL;
2640 g_return_val_if_fail (regex != NULL, NULL);
2641 g_return_val_if_fail (string != NULL, NULL);
2642 g_return_val_if_fail (start_position >= 0, NULL);
2643 g_return_val_if_fail (replacement != NULL, NULL);
2644 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2645 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2647 list = split_replacement (replacement, &tmp_error);
2648 if (tmp_error != NULL)
2650 g_propagate_error (error, tmp_error);
2654 result = g_regex_replace_eval (regex,
2655 string, string_len, start_position,
2657 interpolate_replacement,
2660 if (tmp_error != NULL)
2661 g_propagate_error (error, tmp_error);
2663 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2669 literal_replacement (const GMatchInfo *match_info,
2673 g_string_append (result, data);
2678 * g_regex_replace_literal:
2679 * @regex: a #GRegex structure
2680 * @string: (array length=string_len): the string to perform matches against
2681 * @string_len: the length of @string, or -1 if @string is nul-terminated
2682 * @start_position: starting index of the string to match
2683 * @replacement: text to replace each match with
2684 * @match_options: options for the match
2685 * @error: location to store the error occurring, or %NULL to ignore errors
2687 * Replaces all occurrences of the pattern in @regex with the
2688 * replacement text. @replacement is replaced literally, to
2689 * include backreferences use g_regex_replace().
2691 * Setting @start_position differs from just passing over a
2692 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2693 * case of a pattern that begins with any kind of lookbehind
2694 * assertion, such as "\b".
2696 * Returns: a newly allocated string containing the replacements
2701 g_regex_replace_literal (const GRegex *regex,
2702 const gchar *string,
2704 gint start_position,
2705 const gchar *replacement,
2706 GRegexMatchFlags match_options,
2709 g_return_val_if_fail (replacement != NULL, NULL);
2710 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2712 return g_regex_replace_eval (regex,
2713 string, string_len, start_position,
2715 literal_replacement,
2716 (gpointer)replacement,
2721 * g_regex_replace_eval:
2722 * @regex: a #GRegex structure from g_regex_new()
2723 * @string: (array length=string_len): string to perform matches against
2724 * @string_len: the length of @string, or -1 if @string is nul-terminated
2725 * @start_position: starting index of the string to match
2726 * @match_options: options for the match
2727 * @eval: a function to call for each match
2728 * @user_data: user data to pass to the function
2729 * @error: location to store the error occurring, or %NULL to ignore errors
2731 * Replaces occurrences of the pattern in regex with the output of
2732 * @eval for that occurrence.
2734 * Setting @start_position differs from just passing over a shortened
2735 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2736 * that begins with any kind of lookbehind assertion, such as "\b".
2738 * The following example uses g_regex_replace_eval() to replace multiple
2742 * eval_cb (const GMatchInfo *info,
2749 * match = g_match_info_fetch (info, 0);
2750 * r = g_hash_table_lookup ((GHashTable *)data, match);
2751 * g_string_append (res, r);
2763 * h = g_hash_table_new (g_str_hash, g_str_equal);
2765 * g_hash_table_insert (h, "1", "ONE");
2766 * g_hash_table_insert (h, "2", "TWO");
2767 * g_hash_table_insert (h, "3", "THREE");
2768 * g_hash_table_insert (h, "4", "FOUR");
2770 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2771 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2772 * g_hash_table_destroy (h);
2777 * Returns: a newly allocated string containing the replacements
2782 g_regex_replace_eval (const GRegex *regex,
2783 const gchar *string,
2785 gint start_position,
2786 GRegexMatchFlags match_options,
2787 GRegexEvalCallback eval,
2791 GMatchInfo *match_info;
2794 gboolean done = FALSE;
2795 GError *tmp_error = NULL;
2797 g_return_val_if_fail (regex != NULL, NULL);
2798 g_return_val_if_fail (string != NULL, NULL);
2799 g_return_val_if_fail (start_position >= 0, NULL);
2800 g_return_val_if_fail (eval != NULL, NULL);
2801 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2804 string_len = strlen (string);
2806 result = g_string_sized_new (string_len);
2808 /* run down the string making matches. */
2809 g_regex_match_full (regex, string, string_len, start_position,
2810 match_options, &match_info, &tmp_error);
2811 while (!done && g_match_info_matches (match_info))
2813 g_string_append_len (result,
2815 match_info->offsets[0] - str_pos);
2816 done = (*eval) (match_info, result, user_data);
2817 str_pos = match_info->offsets[1];
2818 g_match_info_next (match_info, &tmp_error);
2820 g_match_info_free (match_info);
2821 if (tmp_error != NULL)
2823 g_propagate_error (error, tmp_error);
2824 g_string_free (result, TRUE);
2828 g_string_append_len (result, string + str_pos, string_len - str_pos);
2829 return g_string_free (result, FALSE);
2833 * g_regex_check_replacement:
2834 * @replacement: the replacement string
2835 * @has_references: (out) (allow-none): location to store information about
2836 * references in @replacement or %NULL
2837 * @error: location to store error
2839 * Checks whether @replacement is a valid replacement string
2840 * (see g_regex_replace()), i.e. that all escape sequences in
2843 * If @has_references is not %NULL then @replacement is checked
2844 * for pattern references. For instance, replacement text 'foo\n'
2845 * does not contain references and may be evaluated without information
2846 * about actual match, but '\0\1' (whole match followed by first
2847 * subpattern) requires valid #GMatchInfo object.
2849 * Returns: whether @replacement is a valid replacement string
2854 g_regex_check_replacement (const gchar *replacement,
2855 gboolean *has_references,
2861 list = split_replacement (replacement, &tmp);
2865 g_propagate_error (error, tmp);
2870 *has_references = interpolation_list_needs_match (list);
2872 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2878 * g_regex_escape_nul:
2879 * @string: the string to escape
2880 * @length: the length of @string
2882 * Escapes the nul characters in @string to "\x00". It can be used
2883 * to compile a regex with embedded nul characters.
2885 * For completeness, @length can be -1 for a nul-terminated string.
2886 * In this case the output string will be of course equal to @string.
2888 * Returns: a newly-allocated escaped string
2893 g_regex_escape_nul (const gchar *string,
2897 const gchar *p, *piece_start, *end;
2900 g_return_val_if_fail (string != NULL, NULL);
2903 return g_strdup (string);
2905 end = string + length;
2906 p = piece_start = string;
2907 escaped = g_string_sized_new (length + 1);
2915 if (p != piece_start)
2917 /* copy the previous piece. */
2918 g_string_append_len (escaped, piece_start, p - piece_start);
2920 if ((backslashes & 1) == 0)
2921 g_string_append_c (escaped, '\\');
2922 g_string_append_c (escaped, 'x');
2923 g_string_append_c (escaped, '0');
2924 g_string_append_c (escaped, '0');
2934 p = g_utf8_next_char (p);
2939 if (piece_start < end)
2940 g_string_append_len (escaped, piece_start, end - piece_start);
2942 return g_string_free (escaped, FALSE);
2946 * g_regex_escape_string:
2947 * @string: (array length=length): the string to escape
2948 * @length: the length of @string, or -1 if @string is nul-terminated
2950 * Escapes the special characters used for regular expressions
2951 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2952 * function is useful to dynamically generate regular expressions.
2954 * @string can contain nul characters that are replaced with "\0",
2955 * in this case remember to specify the correct length of @string
2958 * Returns: a newly-allocated escaped string
2963 g_regex_escape_string (const gchar *string,
2967 const char *p, *piece_start, *end;
2969 g_return_val_if_fail (string != NULL, NULL);
2972 length = strlen (string);
2974 end = string + length;
2975 p = piece_start = string;
2976 escaped = g_string_sized_new (length + 1);
2997 if (p != piece_start)
2998 /* copy the previous piece. */
2999 g_string_append_len (escaped, piece_start, p - piece_start);
3000 g_string_append_c (escaped, '\\');
3002 g_string_append_c (escaped, '0');
3004 g_string_append_c (escaped, *p);
3008 p = g_utf8_next_char (p);
3013 if (piece_start < end)
3014 g_string_append_len (escaped, piece_start, end - piece_start);
3016 return g_string_free (escaped, FALSE);