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
30 #ifdef USE_SYSTEM_PCRE
33 #include "pcre/pcre.h"
36 /* PCRE 7.3 does not contain the definition of PCRE_ERROR_NULLWSLIMIT */
37 #ifndef PCRE_ERROR_NULLWSLIMIT
38 #define PCRE_ERROR_NULLWSLIMIT (-22)
43 /* Mask of all the possible values for GRegexCompileFlags. */
44 #define G_REGEX_COMPILE_MASK (G_REGEX_CASELESS | \
49 G_REGEX_DOLLAR_ENDONLY | \
52 G_REGEX_NO_AUTO_CAPTURE | \
55 G_REGEX_NEWLINE_CR | \
56 G_REGEX_NEWLINE_LF | \
59 /* Mask of all the possible values for GRegexMatchFlags. */
60 #define G_REGEX_MATCH_MASK (G_REGEX_MATCH_ANCHORED | \
61 G_REGEX_MATCH_NOTBOL | \
62 G_REGEX_MATCH_NOTEOL | \
63 G_REGEX_MATCH_NOTEMPTY | \
64 G_REGEX_MATCH_PARTIAL | \
65 G_REGEX_MATCH_NEWLINE_CR | \
66 G_REGEX_MATCH_NEWLINE_LF | \
67 G_REGEX_MATCH_NEWLINE_CRLF | \
68 G_REGEX_MATCH_NEWLINE_ANY)
70 /* if the string is in UTF-8 use g_utf8_ functions, else use
72 #define NEXT_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
73 g_utf8_next_char (s) : \
75 #define PREV_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
76 g_utf8_prev_char (s) : \
81 GRegex *regex; /* the regex */
82 GRegexMatchFlags match_opts; /* options used at match time on the regex */
83 gint matches; /* number of matching sub patterns */
84 gint pos; /* position in the string where last match left off */
85 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
86 gint n_offsets; /* number of offsets */
87 gint *workspace; /* workspace for pcre_dfa_exec() */
88 gint n_workspace; /* number of workspace elements */
89 const gchar *string; /* string passed to the match function */
90 gssize string_len; /* length of string */
95 volatile gint ref_count; /* the ref count for the immutable part */
96 gchar *pattern; /* the pattern */
97 pcre *pcre_re; /* compiled form of the pattern */
98 GRegexCompileFlags compile_opts; /* options used at compile time on the pattern */
99 GRegexMatchFlags match_opts; /* options used at match time on the regex */
100 pcre_extra *extra; /* data stored when G_REGEX_OPTIMIZE is used */
103 /* TRUE if ret is an error code, FALSE otherwise. */
104 #define IS_PCRE_ERROR(ret) ((ret) < PCRE_ERROR_NOMATCH && (ret) != PCRE_ERROR_PARTIAL)
106 typedef struct _InterpolationData InterpolationData;
107 static gboolean interpolation_list_needs_match (GList *list);
108 static gboolean interpolate_replacement (const GMatchInfo *match_info,
111 static GList *split_replacement (const gchar *replacement,
113 static void free_interpolation_data (InterpolationData *data);
117 match_error (gint errcode)
121 case PCRE_ERROR_NOMATCH:
124 case PCRE_ERROR_NULL:
125 /* NULL argument, this should not happen in GRegex */
126 g_warning ("A NULL argument was passed to PCRE");
128 case PCRE_ERROR_BADOPTION:
129 return "bad options";
130 case PCRE_ERROR_BADMAGIC:
131 return _("corrupted object");
132 case PCRE_ERROR_UNKNOWN_OPCODE:
133 return N_("internal error or corrupted object");
134 case PCRE_ERROR_NOMEMORY:
135 return _("out of memory");
136 case PCRE_ERROR_NOSUBSTRING:
137 /* not used by pcre_exec() */
139 case PCRE_ERROR_MATCHLIMIT:
140 return _("backtracking limit reached");
141 case PCRE_ERROR_CALLOUT:
142 /* callouts are not implemented */
144 case PCRE_ERROR_BADUTF8:
145 case PCRE_ERROR_BADUTF8_OFFSET:
146 /* we do not check if strings are valid */
148 case PCRE_ERROR_PARTIAL:
151 case PCRE_ERROR_BADPARTIAL:
152 return _("the pattern contains items not supported for partial matching");
153 case PCRE_ERROR_INTERNAL:
154 return _("internal error");
155 case PCRE_ERROR_BADCOUNT:
156 /* negative ovecsize, this should not happen in GRegex */
157 g_warning ("A negative ovecsize was passed to PCRE");
159 case PCRE_ERROR_DFA_UITEM:
160 return _("the pattern contains items not supported for partial matching");
161 case PCRE_ERROR_DFA_UCOND:
162 return _("back references as conditions are not supported for partial matching");
163 case PCRE_ERROR_DFA_UMLIMIT:
164 /* the match_field field is not used in GRegex */
166 case PCRE_ERROR_DFA_WSSIZE:
167 /* handled expanding the workspace */
169 case PCRE_ERROR_DFA_RECURSE:
170 case PCRE_ERROR_RECURSIONLIMIT:
171 return _("recursion limit reached");
172 case PCRE_ERROR_NULLWSLIMIT:
173 return _("workspace limit for empty substrings reached");
174 case PCRE_ERROR_BADNEWLINE:
175 return _("invalid combination of newline flags");
179 return _("unknown error");
186 match_info_new (const GRegex *regex,
193 GMatchInfo *match_info;
196 string_len = strlen (string);
198 match_info = g_new0 (GMatchInfo, 1);
199 match_info->regex = g_regex_ref ((GRegex *)regex);
200 match_info->string = string;
201 match_info->string_len = string_len;
202 match_info->matches = PCRE_ERROR_NOMATCH;
203 match_info->pos = start_position;
204 match_info->match_opts = match_options;
208 /* These values should be enough for most cases, if they are not
209 * enough g_regex_match_all_full() will expand them. */
210 match_info->n_offsets = 24;
211 match_info->n_workspace = 100;
212 match_info->workspace = g_new (gint, match_info->n_workspace);
217 pcre_fullinfo (regex->pcre_re, regex->extra,
218 PCRE_INFO_CAPTURECOUNT, &capture_count);
219 match_info->n_offsets = (capture_count + 1) * 3;
221 match_info->offsets = g_new0 (gint, match_info->n_offsets);
227 * g_match_info_get_regex:
228 * @match_info: a #GMatchInfo
230 * Returns #GRegex object used in @match_info. It belongs to Glib
231 * and must not be freed. Use g_regex_ref() if you need to keep it
232 * after you free @match_info object.
234 * Returns: #GRegex object used in @match_info
239 g_match_info_get_regex (const GMatchInfo *match_info)
241 g_return_val_if_fail (match_info != NULL, NULL);
242 return match_info->regex;
246 * g_match_info_get_string:
247 * @match_info: a #GMatchInfo
249 * Returns the string searched with @match_info. This is the
250 * string passed to g_regex_match() or g_regex_replace() so
251 * you may not free it before calling this function.
253 * Returns: the string searched with @match_info
258 g_match_info_get_string (const GMatchInfo *match_info)
260 g_return_val_if_fail (match_info != NULL, NULL);
261 return match_info->string;
266 * @match_info: a #GMatchInfo
268 * Frees all the memory associated with the #GMatchInfo structure.
273 g_match_info_free (GMatchInfo *match_info)
277 g_regex_unref (match_info->regex);
278 g_free (match_info->offsets);
279 g_free (match_info->workspace);
286 * @match_info: a #GMatchInfo structure
287 * @error: location to store the error occuring, or %NULL to ignore errors
289 * Scans for the next match using the same parameters of the previous
290 * call to g_regex_match_full() or g_regex_match() that returned
293 * The match is done on the string passed to the match function, so you
294 * cannot free it before calling this function.
296 * Returns: %TRUE is the string matched, %FALSE otherwise
301 g_match_info_next (GMatchInfo *match_info,
306 g_return_val_if_fail (match_info != NULL, FALSE);
307 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
308 g_return_val_if_fail (match_info->pos >= 0, FALSE);
310 opts = match_info->regex->match_opts | match_info->match_opts;
312 match_info->matches = pcre_exec (match_info->regex->pcre_re,
313 match_info->regex->extra,
315 match_info->string_len,
317 match_info->regex->match_opts |
318 match_info->match_opts,
320 match_info->n_offsets);
321 if (IS_PCRE_ERROR (match_info->matches))
323 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
324 _("Error while matching regular expression %s: %s"),
325 match_info->regex->pattern, match_error (match_info->matches));
329 /* avoid infinite loops if the pattern is an empty string or something
331 if (match_info->pos == match_info->offsets[1])
333 if (match_info->pos > match_info->string_len)
335 /* we have reached the end of the string */
336 match_info->pos = -1;
337 match_info->matches = PCRE_ERROR_NOMATCH;
341 match_info->pos = NEXT_CHAR (match_info->regex,
342 &match_info->string[match_info->pos]) -
347 match_info->pos = match_info->offsets[1];
350 return match_info->matches >= 0;
354 * g_match_info_matches:
355 * @match_info: a #GMatchInfo structure
357 * Returns whether the previous match operation succeeded.
359 * Returns: %TRUE if the previous match operation succeeded,
365 g_match_info_matches (const GMatchInfo *match_info)
367 g_return_val_if_fail (match_info != NULL, FALSE);
369 return match_info->matches >= 0;
373 * g_match_info_get_match_count:
374 * @match_info: a #GMatchInfo structure
376 * Retrieves the number of matched substrings (including substring 0,
377 * that is the whole matched text), so 1 is returned if the pattern
378 * has no substrings in it and 0 is returned if the match failed.
380 * If the last match was obtained using the DFA algorithm, that is
381 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
382 * count is not that of the number of capturing parentheses but that of
383 * the number of matched substrings.
385 * Returns: Number of matched substrings, or -1 if an error occurred
390 g_match_info_get_match_count (const GMatchInfo *match_info)
392 g_return_val_if_fail (match_info, -1);
394 if (match_info->matches == PCRE_ERROR_NOMATCH)
397 else if (match_info->matches < PCRE_ERROR_NOMATCH)
402 return match_info->matches;
406 * g_match_info_is_partial_match:
407 * @match_info: a #GMatchInfo structure
409 * Usually if the string passed to g_regex_match*() matches as far as
410 * it goes, but is too short to match the entire pattern, %FALSE is
411 * returned. There are circumstances where it might be helpful to
412 * distinguish this case from other cases in which there is no match.
414 * Consider, for example, an application where a human is required to
415 * type in data for a field with specific formatting requirements. An
416 * example might be a date in the form ddmmmyy, defined by the pattern
417 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
418 * If the application sees the user’s keystrokes one by one, and can
419 * check that what has been typed so far is potentially valid, it is
420 * able to raise an error as soon as a mistake is made.
422 * GRegex supports the concept of partial matching by means of the
423 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
424 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
425 * for a complete match, %FALSE otherwise. But, when these functions
426 * return %FALSE, you can check if the match was partial calling
427 * g_match_info_is_partial_match().
429 * When using partial matching you cannot use g_match_info_fetch*().
431 * Because of the way certain internal optimizations are implemented
432 * the partial matching algorithm cannot be used with all patterns.
433 * So repeated single characters such as "a{2,4}" and repeated single
434 * meta-sequences such as "\d+" are not permitted if the maximum number
435 * of occurrences is greater than one. Optional items such as "\d?"
436 * (where the maximum is one) are permitted. Quantifiers with any values
437 * are permitted after parentheses, so the invalid examples above can be
438 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
439 * for a pattern that does not conform to the restrictions, matching
440 * functions return an error.
442 * Returns: %TRUE if the match was partial, %FALSE otherwise
447 g_match_info_is_partial_match (const GMatchInfo *match_info)
449 g_return_val_if_fail (match_info != NULL, FALSE);
451 return match_info->matches == PCRE_ERROR_PARTIAL;
455 * g_match_info_expand_references:
456 * @match_info: a #GMatchInfo or %NULL
457 * @string_to_expand: the string to expand
458 * @error: location to store the error occuring, or %NULL to ignore errors
460 * Returns a new string containing the text in @string_to_expand with
461 * references and escape sequences expanded. References refer to the last
462 * match done with @string against @regex and have the same syntax used by
465 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
466 * passed to g_regex_new().
468 * The backreferences are extracted from the string passed to the match
469 * function, so you cannot call this function after freeing the string.
471 * @match_info may be %NULL in which case @string_to_expand must not
472 * contain references. For instance "foo\n" does not refer to an actual
473 * pattern and '\n' merely will be replaced with \n character,
474 * while to expand "\0" (whole match) one needs the result of a match.
475 * Use g_regex_check_replacement() to find out whether @string_to_expand
476 * contains references.
478 * Returns: the expanded string, or %NULL if an error occurred
483 g_match_info_expand_references (const GMatchInfo *match_info,
484 const gchar *string_to_expand,
489 GError *tmp_error = NULL;
491 g_return_val_if_fail (string_to_expand != NULL, NULL);
492 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
494 list = split_replacement (string_to_expand, &tmp_error);
495 if (tmp_error != NULL)
497 g_propagate_error (error, tmp_error);
501 if (!match_info && interpolation_list_needs_match (list))
503 g_critical ("String '%s' contains references to the match, can't "
504 "expand references without GMatchInfo object",
509 result = g_string_sized_new (strlen (string_to_expand));
510 interpolate_replacement (match_info, result, list);
512 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
515 return g_string_free (result, FALSE);
519 * g_match_info_fetch:
520 * @match_info: #GMatchInfo structure
521 * @match_num: number of the sub expression
523 * Retrieves the text matching the @match_num<!-- -->'th capturing
524 * parentheses. 0 is the full text of the match, 1 is the first paren
525 * set, 2 the second, and so on.
527 * If @match_num is a valid sub pattern but it didn't match anything
528 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
529 * string is returned.
531 * If the match was obtained using the DFA algorithm, that is using
532 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
533 * string is not that of a set of parentheses but that of a matched
534 * substring. Substrings are matched in reverse order of length, so
535 * 0 is the longest match.
537 * The string is fetched from the string passed to the match function,
538 * so you cannot call this function after freeing the string.
540 * Returns: The matched substring, or %NULL if an error occurred.
541 * You have to free the string yourself
546 g_match_info_fetch (const GMatchInfo *match_info,
549 /* we cannot use pcre_get_substring() because it allocates the
550 * string using pcre_malloc(). */
554 g_return_val_if_fail (match_info != NULL, NULL);
555 g_return_val_if_fail (match_num >= 0, NULL);
557 /* match_num does not exist or it didn't matched, i.e. matching "b"
558 * against "(a)?b" then group 0 is empty. */
559 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
561 else if (start == -1)
562 match = g_strdup ("");
564 match = g_strndup (&match_info->string[start], end - start);
570 * g_match_info_fetch_pos:
571 * @match_info: #GMatchInfo structure
572 * @match_num: number of the sub expression
573 * @start_pos: pointer to location where to store the start position
574 * @end_pos: pointer to location where to store the end position
576 * Retrieves the position of the @match_num<!-- -->'th capturing
577 * parentheses. 0 is the full text of the match, 1 is the first
578 * paren set, 2 the second, and so on.
580 * If @match_num is a valid sub pattern but it didn't match anything
581 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
582 * and @end_pos are set to -1 and %TRUE is returned.
584 * If the match was obtained using the DFA algorithm, that is using
585 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
586 * position is not that of a set of parentheses but that of a matched
587 * substring. Substrings are matched in reverse order of length, so
588 * 0 is the longest match.
590 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
591 * the position cannot be fetched, @start_pos and @end_pos are left
597 g_match_info_fetch_pos (const GMatchInfo *match_info,
602 g_return_val_if_fail (match_info != NULL, FALSE);
603 g_return_val_if_fail (match_num >= 0, FALSE);
605 /* make sure the sub expression number they're requesting is less than
606 * the total number of sub expressions that were matched. */
607 if (match_num >= match_info->matches)
610 if (start_pos != NULL)
611 *start_pos = match_info->offsets[2 * match_num];
614 *end_pos = match_info->offsets[2 * match_num + 1];
620 * Returns number of first matched subpattern with name @name.
621 * There may be more than one in case when DUPNAMES is used,
622 * and not all subpatterns with that name match;
623 * pcre_get_stringnumber() does not work in that case.
626 get_matched_substring_number (const GMatchInfo *match_info,
633 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
634 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
636 /* This code is copied from pcre_get.c: get_first_set() */
637 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
645 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
647 gint n = (entry[0] << 8) + entry[1];
648 if (match_info->offsets[n*2] >= 0)
652 return (first[0] << 8) + first[1];
656 * g_match_info_fetch_named:
657 * @match_info: #GMatchInfo structure
658 * @name: name of the subexpression
660 * Retrieves the text matching the capturing parentheses named @name.
662 * If @name is a valid sub pattern name but it didn't match anything
663 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
664 * then an empty string is returned.
666 * The string is fetched from the string passed to the match function,
667 * so you cannot call this function after freeing the string.
669 * Returns: The matched substring, or %NULL if an error occurred.
670 * You have to free the string yourself
675 g_match_info_fetch_named (const GMatchInfo *match_info,
678 /* we cannot use pcre_get_named_substring() because it allocates the
679 * string using pcre_malloc(). */
682 g_return_val_if_fail (match_info != NULL, NULL);
683 g_return_val_if_fail (name != NULL, NULL);
685 num = get_matched_substring_number (match_info, name);
689 return g_match_info_fetch (match_info, num);
693 * g_match_info_fetch_named_pos:
694 * @match_info: #GMatchInfo structure
695 * @name: name of the subexpression
696 * @start_pos: pointer to location where to store the start position
697 * @end_pos: pointer to location where to store the end position
699 * Retrieves the position of the capturing parentheses named @name.
701 * If @name is a valid sub pattern name but it didn't match anything
702 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
703 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
705 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
706 * the position cannot be fetched, @start_pos and @end_pos are left
712 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
719 g_return_val_if_fail (match_info != NULL, FALSE);
720 g_return_val_if_fail (name != NULL, FALSE);
722 num = get_matched_substring_number (match_info, name);
726 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
730 * g_match_info_fetch_all:
731 * @match_info: a #GMatchInfo structure
733 * Bundles up pointers to each of the matching substrings from a match
734 * and stores them in an array of gchar pointers. The first element in
735 * the returned array is the match number 0, i.e. the entire matched
738 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
739 * "b" against "(a)?b") then an empty string is inserted.
741 * If the last match was obtained using the DFA algorithm, that is using
742 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
743 * strings are not that matched by sets of parentheses but that of the
744 * matched substring. Substrings are matched in reverse order of length,
745 * so the first one is the longest match.
747 * The strings are fetched from the string passed to the match function,
748 * so you cannot call this function after freeing the string.
750 * Returns: a %NULL-terminated array of gchar * pointers. It must be
751 * freed using g_strfreev(). If the previous match failed %NULL is
757 g_match_info_fetch_all (const GMatchInfo *match_info)
759 /* we cannot use pcre_get_substring_list() because the returned value
760 * isn't suitable for g_strfreev(). */
764 g_return_val_if_fail (match_info != NULL, NULL);
766 if (match_info->matches < 0)
769 result = g_new (gchar *, match_info->matches + 1);
770 for (i = 0; i < match_info->matches; i++)
771 result[i] = g_match_info_fetch (match_info, i);
781 g_regex_error_quark (void)
783 static GQuark error_quark = 0;
785 if (error_quark == 0)
786 error_quark = g_quark_from_static_string ("g-regex-error-quark");
795 * Increases reference count of @regex by 1.
802 g_regex_ref (GRegex *regex)
804 g_return_val_if_fail (regex != NULL, NULL);
805 g_atomic_int_inc (®ex->ref_count);
813 * Decreases reference count of @regex by 1. When reference count drops
814 * to zero, it frees all the memory associated with the regex structure.
819 g_regex_unref (GRegex *regex)
821 g_return_if_fail (regex != NULL);
823 if (g_atomic_int_exchange_and_add (®ex->ref_count, -1) - 1 == 0)
825 g_free (regex->pattern);
826 if (regex->pcre_re != NULL)
827 pcre_free (regex->pcre_re);
828 if (regex->extra != NULL)
829 pcre_free (regex->extra);
836 * @pattern: the regular expression
837 * @compile_options: compile options for the regular expression
838 * @match_options: match options for the regular expression
839 * @error: return location for a #GError
841 * Compiles the regular expression to an internal form, and does
842 * the initial setup of the #GRegex structure.
844 * Returns: a #GRegex structure. Call g_regex_unref() when you
850 g_regex_new (const gchar *pattern,
851 GRegexCompileFlags compile_options,
852 GRegexMatchFlags match_options,
859 gboolean optimize = FALSE;
860 static gboolean initialized = FALSE;
861 unsigned long int pcre_compile_options;
863 g_return_val_if_fail (pattern != NULL, NULL);
864 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
865 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
866 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
873 pcre_config (PCRE_CONFIG_UTF8, &support);
876 msg = N_("PCRE library is compiled without UTF8 support");
878 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
882 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
885 msg = N_("PCRE library is compiled without UTF8 properties support");
887 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
894 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
895 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
896 if (compile_options & G_REGEX_OPTIMIZE)
899 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
900 * instead uses UTF-8 only if required with PCRE_UTF8. */
901 if (compile_options & G_REGEX_RAW)
904 compile_options &= ~G_REGEX_RAW;
909 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
910 match_options |= PCRE_NO_UTF8_CHECK;
913 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
914 * not for the system one. */
915 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
916 !(compile_options & G_REGEX_NEWLINE_LF))
918 compile_options |= PCRE_NEWLINE_ANY;
921 /* compile the pattern */
922 re = pcre_compile (pattern, compile_options, &errmsg, &erroffset, NULL);
924 /* if the compilation failed, set the error member and return
930 /* PCRE uses byte offsets but we want to show character offsets */
931 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
933 tmp_error = g_error_new (G_REGEX_ERROR,
934 G_REGEX_ERROR_COMPILE,
935 _("Error while compiling regular "
936 "expression %s at char %d: %s"),
937 pattern, erroffset, errmsg);
938 g_propagate_error (error, tmp_error);
943 /* For options set at the beginning of the pattern, pcre puts them into
944 * compile options, e.g. "(?i)foo" will make the pcre structure store
945 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
946 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
947 compile_options = pcre_compile_options;
949 if (!(compile_options & G_REGEX_DUPNAMES))
951 gboolean jchanged = FALSE;
952 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
954 compile_options |= G_REGEX_DUPNAMES;
957 regex = g_new0 (GRegex, 1);
958 regex->ref_count = 1;
959 regex->pattern = g_strdup (pattern);
961 regex->compile_opts = compile_options;
962 regex->match_opts = match_options;
966 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
969 GError *tmp_error = g_error_new (G_REGEX_ERROR,
970 G_REGEX_ERROR_OPTIMIZE,
971 _("Error while optimizing "
972 "regular expression %s: %s"),
975 g_propagate_error (error, tmp_error);
984 * g_regex_get_pattern:
985 * @regex: a #GRegex structure
987 * Gets the pattern string associated with @regex, i.e. a copy of
988 * the string passed to g_regex_new().
990 * Returns: the pattern of @regex
995 g_regex_get_pattern (const GRegex *regex)
997 g_return_val_if_fail (regex != NULL, NULL);
999 return regex->pattern;
1003 * g_regex_get_max_backref:
1006 * Returns the number of the highest back reference
1007 * in the pattern, or 0 if the pattern does not contain
1010 * Returns: the number of the highest back reference
1015 g_regex_get_max_backref (const GRegex *regex)
1019 pcre_fullinfo (regex->pcre_re, regex->extra,
1020 PCRE_INFO_BACKREFMAX, &value);
1026 * g_regex_get_capture_count:
1029 * Returns the number of capturing subpatterns in the pattern.
1031 * Returns: the number of capturing subpatterns
1036 g_regex_get_capture_count (const GRegex *regex)
1040 pcre_fullinfo (regex->pcre_re, regex->extra,
1041 PCRE_INFO_CAPTURECOUNT, &value);
1047 * g_regex_match_simple:
1048 * @pattern: the regular expression
1049 * @string: the string to scan for matches
1050 * @compile_options: compile options for the regular expression
1051 * @match_options: match options
1053 * Scans for a match in @string for @pattern.
1055 * This function is equivalent to g_regex_match() but it does not
1056 * require to compile the pattern with g_regex_new(), avoiding some
1057 * lines of code when you need just to do a match without extracting
1058 * substrings, capture counts, and so on.
1060 * If this function is to be called on the same @pattern more than
1061 * once, it's more efficient to compile the pattern once with
1062 * g_regex_new() and then use g_regex_match().
1064 * Returns: %TRUE is the string matched, %FALSE otherwise
1069 g_regex_match_simple (const gchar *pattern,
1070 const gchar *string,
1071 GRegexCompileFlags compile_options,
1072 GRegexMatchFlags match_options)
1077 regex = g_regex_new (pattern, compile_options, 0, NULL);
1080 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1081 g_regex_unref (regex);
1087 * @regex: a #GRegex structure from g_regex_new()
1088 * @string: the string to scan for matches
1089 * @match_options: match options
1090 * @match_info: pointer to location where to store the #GMatchInfo,
1091 * or %NULL if you do not need it
1093 * Scans for a match in string for the pattern in @regex.
1094 * The @match_options are combined with the match options specified
1095 * when the @regex structure was created, letting you have more
1096 * flexibility in reusing #GRegex structures.
1098 * A #GMatchInfo structure, used to get information on the match,
1099 * is stored in @match_info if not %NULL. Note that if @match_info
1100 * is not %NULL then it is created even if the function returns %FALSE,
1101 * i.e. you must free it regardless if regular expression actually matched.
1103 * To retrieve all the non-overlapping matches of the pattern in
1104 * string you can use g_match_info_next().
1106 * <informalexample><programlisting>
1108 * print_uppercase_words (const gchar *string)
1110 * /* Print all uppercase-only words. */
1112 * GMatchInfo *match_info;
1114 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1115 * g_regex_match (regex, string, 0, &match_info);
1116 * while (g_match_info_matches (match_info))
1118 * gchar *word = g_match_info_fetch (match_info, 0);
1119 * g_print ("Found: %s\n", word);
1121 * g_match_info_next (match_info, NULL);
1123 * g_match_info_free (match_info);
1124 * g_regex_unref (regex);
1126 * </programlisting></informalexample>
1128 * Returns: %TRUE is the string matched, %FALSE otherwise
1133 g_regex_match (const GRegex *regex,
1134 const gchar *string,
1135 GRegexMatchFlags match_options,
1136 GMatchInfo **match_info)
1138 return g_regex_match_full (regex, string, -1, 0, match_options,
1143 * g_regex_match_full:
1144 * @regex: a #GRegex structure from g_regex_new()
1145 * @string: the string to scan for matches
1146 * @string_len: the length of @string, or -1 if @string is nul-terminated
1147 * @start_position: starting index of the string to match
1148 * @match_options: match options
1149 * @match_info: pointer to location where to store the #GMatchInfo,
1150 * or %NULL if you do not need it
1151 * @error: location to store the error occuring, or %NULL to ignore errors
1153 * Scans for a match in string for the pattern in @regex.
1154 * The @match_options are combined with the match options specified
1155 * when the @regex structure was created, letting you have more
1156 * flexibility in reusing #GRegex structures.
1158 * Setting @start_position differs from just passing over a shortened
1159 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1160 * that begins with any kind of lookbehind assertion, such as "\b".
1162 * A #GMatchInfo structure, used to get information on the match, is
1163 * stored in @match_info if not %NULL. Note that if @match_info is
1164 * not %NULL then it is created even if the function returns %FALSE,
1165 * i.e. you must free it regardless if regular expression actually
1168 * @string is not copied and is used in #GMatchInfo internally. If
1169 * you use any #GMatchInfo method (except g_match_info_free()) after
1170 * freeing or modifying @string then the behaviour is undefined.
1172 * To retrieve all the non-overlapping matches of the pattern in
1173 * string you can use g_match_info_next().
1175 * <informalexample><programlisting>
1177 * print_uppercase_words (const gchar *string)
1179 * /* Print all uppercase-only words. */
1181 * GMatchInfo *match_info;
1182 * GError *error = NULL;
1184 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1185 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1186 * while (g_match_info_matches (match_info))
1188 * gchar *word = g_match_info_fetch (match_info, 0);
1189 * g_print ("Found: %s\n", word);
1191 * g_match_info_next (match_info, &error);
1193 * g_match_info_free (match_info);
1194 * g_regex_unref (regex);
1195 * if (error != NULL)
1197 * g_printerr ("Error while matching: %s\n", error->message);
1198 * g_error_free (error);
1201 * </programlisting></informalexample>
1203 * Returns: %TRUE is the string matched, %FALSE otherwise
1208 g_regex_match_full (const GRegex *regex,
1209 const gchar *string,
1211 gint start_position,
1212 GRegexMatchFlags match_options,
1213 GMatchInfo **match_info,
1219 g_return_val_if_fail (regex != NULL, FALSE);
1220 g_return_val_if_fail (string != NULL, FALSE);
1221 g_return_val_if_fail (start_position >= 0, FALSE);
1222 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1223 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1225 info = match_info_new (regex, string, string_len, start_position,
1226 match_options, FALSE);
1227 match_ok = g_match_info_next (info, error);
1228 if (match_info != NULL)
1231 g_match_info_free (info);
1237 * g_regex_match_all:
1238 * @regex: a #GRegex structure from g_regex_new()
1239 * @string: the string to scan for matches
1240 * @match_options: match options
1241 * @match_info: pointer to location where to store the #GMatchInfo,
1242 * or %NULL if you do not need it
1244 * Using the standard algorithm for regular expression matching only
1245 * the longest match in the string is retrieved. This function uses
1246 * a different algorithm so it can retrieve all the possible matches.
1247 * For more documentation see g_regex_match_all_full().
1249 * A #GMatchInfo structure, used to get information on the match, is
1250 * stored in @match_info if not %NULL. Note that if @match_info is
1251 * not %NULL then it is created even if the function returns %FALSE,
1252 * i.e. you must free it regardless if regular expression actually
1255 * Returns: %TRUE is the string matched, %FALSE otherwise
1260 g_regex_match_all (const GRegex *regex,
1261 const gchar *string,
1262 GRegexMatchFlags match_options,
1263 GMatchInfo **match_info)
1265 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1270 * g_regex_match_all_full:
1271 * @regex: a #GRegex structure from g_regex_new()
1272 * @string: the string to scan for matches
1273 * @string_len: the length of @string, or -1 if @string is nul-terminated
1274 * @start_position: starting index of the string to match
1275 * @match_options: match options
1276 * @match_info: pointer to location where to store the #GMatchInfo,
1277 * or %NULL if you do not need it
1278 * @error: location to store the error occuring, or %NULL to ignore errors
1280 * Using the standard algorithm for regular expression matching only
1281 * the longest match in the string is retrieved, it is not possibile
1282 * to obtain all the available matches. For instance matching
1283 * "<a> <b> <c>" against the pattern "<.*>"
1284 * you get "<a> <b> <c>".
1286 * This function uses a different algorithm (called DFA, i.e. deterministic
1287 * finite automaton), so it can retrieve all the possible matches, all
1288 * starting at the same point in the string. For instance matching
1289 * "<a> <b> <c>" against the pattern "<.*>"
1290 * you would obtain three matches: "<a> <b> <c>",
1291 * "<a> <b>" and "<a>".
1293 * The number of matched strings is retrieved using
1294 * g_match_info_get_match_count(). To obtain the matched strings and
1295 * their position you can use, respectively, g_match_info_fetch() and
1296 * g_match_info_fetch_pos(). Note that the strings are returned in
1297 * reverse order of length; that is, the longest matching string is
1300 * Note that the DFA algorithm is slower than the standard one and it
1301 * is not able to capture substrings, so backreferences do not work.
1303 * Setting @start_position differs from just passing over a shortened
1304 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1305 * that begins with any kind of lookbehind assertion, such as "\b".
1307 * A #GMatchInfo structure, used to get information on the match, is
1308 * stored in @match_info if not %NULL. Note that if @match_info is
1309 * not %NULL then it is created even if the function returns %FALSE,
1310 * i.e. you must free it regardless if regular expression actually
1313 * Returns: %TRUE is the string matched, %FALSE otherwise
1318 g_regex_match_all_full (const GRegex *regex,
1319 const gchar *string,
1321 gint start_position,
1322 GRegexMatchFlags match_options,
1323 GMatchInfo **match_info,
1329 g_return_val_if_fail (regex != NULL, FALSE);
1330 g_return_val_if_fail (string != NULL, FALSE);
1331 g_return_val_if_fail (start_position >= 0, FALSE);
1332 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1333 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1335 info = match_info_new (regex, string, string_len, start_position,
1336 match_options, TRUE);
1342 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1343 info->string, info->string_len,
1345 regex->match_opts | match_options,
1346 info->offsets, info->n_offsets,
1347 info->workspace, info->n_workspace);
1348 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1350 /* info->workspace is too small. */
1351 info->n_workspace *= 2;
1352 info->workspace = g_realloc (info->workspace,
1353 info->n_workspace * sizeof (gint));
1356 else if (info->matches == 0)
1358 /* info->offsets is too small. */
1359 info->n_offsets *= 2;
1360 info->offsets = g_realloc (info->offsets,
1361 info->n_offsets * sizeof (gint));
1364 else if (IS_PCRE_ERROR (info->matches))
1366 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1367 _("Error while matching regular expression %s: %s"),
1368 regex->pattern, match_error (info->matches));
1372 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1375 if (match_info != NULL)
1378 g_match_info_free (info);
1380 return info->matches >= 0;
1384 * g_regex_get_string_number:
1385 * @regex: #GRegex structure
1386 * @name: name of the subexpression
1388 * Retrieves the number of the subexpression named @name.
1390 * Returns: The number of the subexpression or -1 if @name
1396 g_regex_get_string_number (const GRegex *regex,
1401 g_return_val_if_fail (regex != NULL, -1);
1402 g_return_val_if_fail (name != NULL, -1);
1404 num = pcre_get_stringnumber (regex->pcre_re, name);
1405 if (num == PCRE_ERROR_NOSUBSTRING)
1412 * g_regex_split_simple:
1413 * @pattern: the regular expression
1414 * @string: the string to scan for matches
1415 * @compile_options: compile options for the regular expression
1416 * @match_options: match options
1418 * Breaks the string on the pattern, and returns an array of
1419 * the tokens. If the pattern contains capturing parentheses,
1420 * then the text for each of the substrings will also be returned.
1421 * If the pattern does not match anywhere in the string, then the
1422 * whole string is returned as the first token.
1424 * This function is equivalent to g_regex_split() but it does
1425 * not require to compile the pattern with g_regex_new(), avoiding
1426 * some lines of code when you need just to do a split without
1427 * extracting substrings, capture counts, and so on.
1429 * If this function is to be called on the same @pattern more than
1430 * once, it's more efficient to compile the pattern once with
1431 * g_regex_new() and then use g_regex_split().
1433 * As a special case, the result of splitting the empty string ""
1434 * is an empty vector, not a vector containing a single string.
1435 * The reason for this special case is that being able to represent
1436 * a empty vector is typically more useful than consistent handling
1437 * of empty elements. If you do need to represent empty elements,
1438 * you'll need to check for the empty string before calling this
1441 * A pattern that can match empty strings splits @string into
1442 * separate characters wherever it matches the empty string between
1443 * characters. For example splitting "ab c" using as a separator
1444 * "\s*", you will get "a", "b" and "c".
1446 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1451 g_regex_split_simple (const gchar *pattern,
1452 const gchar *string,
1453 GRegexCompileFlags compile_options,
1454 GRegexMatchFlags match_options)
1459 regex = g_regex_new (pattern, compile_options, 0, NULL);
1462 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1463 g_regex_unref (regex);
1469 * @regex: a #GRegex structure
1470 * @string: the string to split with the pattern
1471 * @match_options: match time option flags
1473 * Breaks the string on the pattern, and returns an array of the tokens.
1474 * If the pattern contains capturing parentheses, then the text for each
1475 * of the substrings will also be returned. If the pattern does not match
1476 * anywhere in the string, then the whole string is returned as the first
1479 * As a special case, the result of splitting the empty string "" is an
1480 * empty vector, not a vector containing a single string. The reason for
1481 * this special case is that being able to represent a empty vector is
1482 * typically more useful than consistent handling of empty elements. If
1483 * you do need to represent empty elements, you'll need to check for the
1484 * empty string before calling this function.
1486 * A pattern that can match empty strings splits @string into separate
1487 * characters wherever it matches the empty string between characters.
1488 * For example splitting "ab c" using as a separator "\s*", you will get
1491 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1496 g_regex_split (const GRegex *regex,
1497 const gchar *string,
1498 GRegexMatchFlags match_options)
1500 return g_regex_split_full (regex, string, -1, 0,
1501 match_options, 0, NULL);
1505 * g_regex_split_full:
1506 * @regex: a #GRegex structure
1507 * @string: the string to split with the pattern
1508 * @string_len: the length of @string, or -1 if @string is nul-terminated
1509 * @start_position: starting index of the string to match
1510 * @match_options: match time option flags
1511 * @max_tokens: the maximum number of tokens to split @string into.
1512 * If this is less than 1, the string is split completely
1513 * @error: return location for a #GError
1515 * Breaks the string on the pattern, and returns an array of the tokens.
1516 * If the pattern contains capturing parentheses, then the text for each
1517 * of the substrings will also be returned. If the pattern does not match
1518 * anywhere in the string, then the whole string is returned as the first
1521 * As a special case, the result of splitting the empty string "" is an
1522 * empty vector, not a vector containing a single string. The reason for
1523 * this special case is that being able to represent a empty vector is
1524 * typically more useful than consistent handling of empty elements. If
1525 * you do need to represent empty elements, you'll need to check for the
1526 * empty string before calling this function.
1528 * A pattern that can match empty strings splits @string into separate
1529 * characters wherever it matches the empty string between characters.
1530 * For example splitting "ab c" using as a separator "\s*", you will get
1533 * Setting @start_position differs from just passing over a shortened
1534 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1535 * that begins with any kind of lookbehind assertion, such as "\b".
1537 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1542 g_regex_split_full (const GRegex *regex,
1543 const gchar *string,
1545 gint start_position,
1546 GRegexMatchFlags match_options,
1550 GError *tmp_error = NULL;
1551 GMatchInfo *match_info;
1556 /* position of the last separator. */
1557 gint last_separator_end;
1558 /* was the last match 0 bytes long? */
1559 gboolean last_match_is_empty;
1560 /* the returned array of char **s */
1561 gchar **string_list;
1563 g_return_val_if_fail (regex != NULL, NULL);
1564 g_return_val_if_fail (string != NULL, NULL);
1565 g_return_val_if_fail (start_position >= 0, NULL);
1566 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1567 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1569 if (max_tokens <= 0)
1570 max_tokens = G_MAXINT;
1573 string_len = strlen (string);
1575 /* zero-length string */
1576 if (string_len - start_position == 0)
1577 return g_new0 (gchar *, 1);
1579 if (max_tokens == 1)
1581 string_list = g_new0 (gchar *, 2);
1582 string_list[0] = g_strndup (&string[start_position],
1583 string_len - start_position);
1589 last_separator_end = start_position;
1590 last_match_is_empty = FALSE;
1592 match_ok = g_regex_match_full (regex, string, string_len, start_position,
1593 match_options, &match_info, &tmp_error);
1594 while (tmp_error == NULL)
1598 last_match_is_empty =
1599 (match_info->offsets[0] == match_info->offsets[1]);
1601 /* we need to skip empty separators at the same position of the end
1602 * of another separator. e.g. the string is "a b" and the separator
1603 * is " *", so from 1 to 2 we have a match and at position 2 we have
1604 * an empty match. */
1605 if (last_separator_end != match_info->offsets[1])
1610 token = g_strndup (string + last_separator_end,
1611 match_info->offsets[0] - last_separator_end);
1612 list = g_list_prepend (list, token);
1615 /* if there were substrings, these need to be added to
1617 match_count = g_match_info_get_match_count (match_info);
1618 if (match_count > 1)
1620 for (i = 1; i < match_count; i++)
1621 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
1627 /* if there was no match, copy to end of string. */
1628 if (!last_match_is_empty)
1630 gchar *token = g_strndup (string + last_separator_end,
1631 match_info->string_len - last_separator_end);
1632 list = g_list_prepend (list, token);
1634 /* no more tokens, end the loop. */
1638 /* -1 to leave room for the last part. */
1639 if (token_count >= max_tokens - 1)
1641 /* we have reached the maximum number of tokens, so we copy
1642 * the remaining part of the string. */
1643 if (last_match_is_empty)
1645 /* the last match was empty, so we have moved one char
1646 * after the real position to avoid empty matches at the
1648 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
1650 /* the if is needed in the case we have terminated the available
1651 * tokens, but we are at the end of the string, so there are no
1652 * characters left to copy. */
1653 if (string_len > match_info->pos)
1655 gchar *token = g_strndup (string + match_info->pos,
1656 string_len - match_info->pos);
1657 list = g_list_prepend (list, token);
1663 last_separator_end = match_info->pos;
1664 if (last_match_is_empty)
1665 /* if the last match was empty, g_match_info_next() has moved
1666 * forward to avoid infinite loops, but we still need to copy that
1668 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
1670 match_ok = g_match_info_next (match_info, &tmp_error);
1672 g_match_info_free (match_info);
1673 if (tmp_error != NULL)
1675 g_propagate_error (error, tmp_error);
1676 g_list_foreach (list, (GFunc)g_free, NULL);
1678 match_info->pos = -1;
1682 string_list = g_new (gchar *, g_list_length (list) + 1);
1684 for (last = g_list_last (list); last; last = g_list_previous (last))
1685 string_list[i++] = last->data;
1686 string_list[i] = NULL;
1695 REPL_TYPE_CHARACTER,
1696 REPL_TYPE_SYMBOLIC_REFERENCE,
1697 REPL_TYPE_NUMERIC_REFERENCE,
1698 REPL_TYPE_CHANGE_CASE
1703 CHANGE_CASE_NONE = 1 << 0,
1704 CHANGE_CASE_UPPER = 1 << 1,
1705 CHANGE_CASE_LOWER = 1 << 2,
1706 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
1707 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
1708 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
1709 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
1710 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
1713 struct _InterpolationData
1719 ChangeCase change_case;
1723 free_interpolation_data (InterpolationData *data)
1725 g_free (data->text);
1729 static const gchar *
1730 expand_escape (const gchar *replacement,
1732 InterpolationData *data,
1737 const gchar *error_detail;
1739 GError *tmp_error = NULL;
1747 data->type = REPL_TYPE_CHARACTER;
1752 data->type = REPL_TYPE_CHARACTER;
1757 data->type = REPL_TYPE_CHARACTER;
1762 data->type = REPL_TYPE_CHARACTER;
1767 data->type = REPL_TYPE_CHARACTER;
1772 data->type = REPL_TYPE_CHARACTER;
1777 data->type = REPL_TYPE_CHARACTER;
1782 data->type = REPL_TYPE_CHARACTER;
1792 h = g_ascii_xdigit_value (*p);
1795 error_detail = _("hexadecimal digit or '}' expected");
1806 for (i = 0; i < 2; i++)
1808 h = g_ascii_xdigit_value (*p);
1811 error_detail = _("hexadecimal digit expected");
1818 data->type = REPL_TYPE_STRING;
1819 data->text = g_new0 (gchar, 8);
1820 g_unichar_to_utf8 (x, data->text);
1824 data->type = REPL_TYPE_CHANGE_CASE;
1825 data->change_case = CHANGE_CASE_LOWER_SINGLE;
1829 data->type = REPL_TYPE_CHANGE_CASE;
1830 data->change_case = CHANGE_CASE_UPPER_SINGLE;
1834 data->type = REPL_TYPE_CHANGE_CASE;
1835 data->change_case = CHANGE_CASE_LOWER;
1839 data->type = REPL_TYPE_CHANGE_CASE;
1840 data->change_case = CHANGE_CASE_UPPER;
1844 data->type = REPL_TYPE_CHANGE_CASE;
1845 data->change_case = CHANGE_CASE_NONE;
1851 error_detail = _("missing '<' in symbolic reference");
1860 error_detail = _("unfinished symbolic reference");
1867 error_detail = _("zero-length symbolic reference");
1870 if (g_ascii_isdigit (*q))
1875 h = g_ascii_digit_value (*q);
1878 error_detail = _("digit expected");
1887 data->type = REPL_TYPE_NUMERIC_REFERENCE;
1894 if (!g_ascii_isalnum (*r))
1896 error_detail = _("illegal symbolic reference");
1903 data->text = g_strndup (q, p - q);
1904 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
1909 /* if \0 is followed by a number is an octal number representing a
1910 * character, else it is a numeric reference. */
1911 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
1914 p = g_utf8_next_char (p);
1927 for (i = 0; i < 3; i++)
1929 h = g_ascii_digit_value (*p);
1939 if (i == 2 && base == 10)
1945 if (base == 8 || i == 3)
1947 data->type = REPL_TYPE_STRING;
1948 data->text = g_new0 (gchar, 8);
1949 g_unichar_to_utf8 (x, data->text);
1953 data->type = REPL_TYPE_NUMERIC_REFERENCE;
1958 error_detail = _("stray final '\\'");
1962 error_detail = _("unknown escape sequence");
1969 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
1970 tmp_error = g_error_new (G_REGEX_ERROR,
1971 G_REGEX_ERROR_REPLACE,
1972 _("Error while parsing replacement "
1973 "text \"%s\" at char %lu: %s"),
1975 (gulong)(p - replacement),
1977 g_propagate_error (error, tmp_error);
1983 split_replacement (const gchar *replacement,
1987 InterpolationData *data;
1988 const gchar *p, *start;
1990 start = p = replacement;
1995 data = g_new0 (InterpolationData, 1);
1996 start = p = expand_escape (replacement, p, data, error);
1999 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2001 free_interpolation_data (data);
2005 list = g_list_prepend (list, data);
2010 if (*p == '\\' || *p == '\0')
2014 data = g_new0 (InterpolationData, 1);
2015 data->text = g_strndup (start, p - start);
2016 data->type = REPL_TYPE_STRING;
2017 list = g_list_prepend (list, data);
2023 return g_list_reverse (list);
2026 /* Change the case of c based on change_case. */
2027 #define CHANGE_CASE(c, change_case) \
2028 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2029 g_unichar_tolower (c) : \
2030 g_unichar_toupper (c))
2033 string_append (GString *string,
2035 ChangeCase *change_case)
2039 if (text[0] == '\0')
2042 if (*change_case == CHANGE_CASE_NONE)
2044 g_string_append (string, text);
2046 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2048 c = g_utf8_get_char (text);
2049 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2050 g_string_append (string, g_utf8_next_char (text));
2051 *change_case = CHANGE_CASE_NONE;
2055 while (*text != '\0')
2057 c = g_utf8_get_char (text);
2058 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2059 text = g_utf8_next_char (text);
2065 interpolate_replacement (const GMatchInfo *match_info,
2070 InterpolationData *idata;
2072 ChangeCase change_case = CHANGE_CASE_NONE;
2074 for (list = data; list; list = list->next)
2077 switch (idata->type)
2079 case REPL_TYPE_STRING:
2080 string_append (result, idata->text, &change_case);
2082 case REPL_TYPE_CHARACTER:
2083 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2084 if (change_case & CHANGE_CASE_SINGLE_MASK)
2085 change_case = CHANGE_CASE_NONE;
2087 case REPL_TYPE_NUMERIC_REFERENCE:
2088 match = g_match_info_fetch (match_info, idata->num);
2091 string_append (result, match, &change_case);
2095 case REPL_TYPE_SYMBOLIC_REFERENCE:
2096 match = g_match_info_fetch_named (match_info, idata->text);
2099 string_append (result, match, &change_case);
2103 case REPL_TYPE_CHANGE_CASE:
2104 change_case = idata->change_case;
2112 /* whether actual match_info is needed for replacement, i.e.
2113 * whether there are references
2116 interpolation_list_needs_match (GList *list)
2118 while (list != NULL)
2120 InterpolationData *data = list->data;
2122 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2123 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2136 * @regex: a #GRegex structure
2137 * @string: the string to perform matches against
2138 * @string_len: the length of @string, or -1 if @string is nul-terminated
2139 * @start_position: starting index of the string to match
2140 * @replacement: text to replace each match with
2141 * @match_options: options for the match
2142 * @error: location to store the error occuring, or %NULL to ignore errors
2144 * Replaces all occurances of the pattern in @regex with the
2145 * replacement text. Backreferences of the form '\number' or
2146 * '\g<number>' in the replacement text are interpolated by the
2147 * number-th captured subexpression of the match, '\g<name>' refers
2148 * to the captured subexpression with the given name. '\0' refers to the
2149 * complete match, but '\0' followed by a number is the octal representation
2150 * of a character. To include a literal '\' in the replacement, write '\\'.
2151 * There are also escapes that changes the case of the following text:
2154 * <varlistentry><term>\l</term>
2156 * <para>Convert to lower case the next character</para>
2159 * <varlistentry><term>\u</term>
2161 * <para>Convert to upper case the next character</para>
2164 * <varlistentry><term>\L</term>
2166 * <para>Convert to lower case till \E</para>
2169 * <varlistentry><term>\U</term>
2171 * <para>Convert to upper case till \E</para>
2174 * <varlistentry><term>\E</term>
2176 * <para>End case modification</para>
2181 * If you do not need to use backreferences use g_regex_replace_literal().
2183 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2184 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2185 * you can use g_regex_replace_literal().
2187 * Setting @start_position differs from just passing over a shortened
2188 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2189 * begins with any kind of lookbehind assertion, such as "\b".
2191 * Returns: a newly allocated string containing the replacements
2196 g_regex_replace (const GRegex *regex,
2197 const gchar *string,
2199 gint start_position,
2200 const gchar *replacement,
2201 GRegexMatchFlags match_options,
2206 GError *tmp_error = NULL;
2208 g_return_val_if_fail (regex != NULL, NULL);
2209 g_return_val_if_fail (string != NULL, NULL);
2210 g_return_val_if_fail (start_position >= 0, NULL);
2211 g_return_val_if_fail (replacement != NULL, NULL);
2212 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2213 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2215 list = split_replacement (replacement, &tmp_error);
2216 if (tmp_error != NULL)
2218 g_propagate_error (error, tmp_error);
2222 result = g_regex_replace_eval (regex,
2223 string, string_len, start_position,
2225 interpolate_replacement,
2228 if (tmp_error != NULL)
2229 g_propagate_error (error, tmp_error);
2231 g_list_foreach (list, (GFunc)free_interpolation_data, NULL);
2238 literal_replacement (const GMatchInfo *match_info,
2242 g_string_append (result, data);
2247 * g_regex_replace_literal:
2248 * @regex: a #GRegex structure
2249 * @string: the string to perform matches against
2250 * @string_len: the length of @string, or -1 if @string is nul-terminated
2251 * @start_position: starting index of the string to match
2252 * @replacement: text to replace each match with
2253 * @match_options: options for the match
2254 * @error: location to store the error occuring, or %NULL to ignore errors
2256 * Replaces all occurances of the pattern in @regex with the
2257 * replacement text. @replacement is replaced literally, to
2258 * include backreferences use g_regex_replace().
2260 * Setting @start_position differs from just passing over a
2261 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2262 * case of a pattern that begins with any kind of lookbehind
2263 * assertion, such as "\b".
2265 * Returns: a newly allocated string containing the replacements
2270 g_regex_replace_literal (const GRegex *regex,
2271 const gchar *string,
2273 gint start_position,
2274 const gchar *replacement,
2275 GRegexMatchFlags match_options,
2278 g_return_val_if_fail (replacement != NULL, NULL);
2279 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2281 return g_regex_replace_eval (regex,
2282 string, string_len, start_position,
2284 literal_replacement,
2285 (gpointer)replacement,
2290 * g_regex_replace_eval:
2291 * @regex: a #GRegex structure from g_regex_new()
2292 * @string: string to perform matches against
2293 * @string_len: the length of @string, or -1 if @string is nul-terminated
2294 * @start_position: starting index of the string to match
2295 * @match_options: options for the match
2296 * @eval: a function to call for each match
2297 * @user_data: user data to pass to the function
2298 * @error: location to store the error occuring, or %NULL to ignore errors
2300 * Replaces occurances of the pattern in regex with the output of
2301 * @eval for that occurance.
2303 * Setting @start_position differs from just passing over a shortened
2304 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2305 * that begins with any kind of lookbehind assertion, such as "\b".
2307 * Returns: a newly allocated string containing the replacements
2312 g_regex_replace_eval (const GRegex *regex,
2313 const gchar *string,
2315 gint start_position,
2316 GRegexMatchFlags match_options,
2317 GRegexEvalCallback eval,
2321 GMatchInfo *match_info;
2324 gboolean done = FALSE;
2325 GError *tmp_error = NULL;
2327 g_return_val_if_fail (regex != NULL, NULL);
2328 g_return_val_if_fail (string != NULL, NULL);
2329 g_return_val_if_fail (start_position >= 0, NULL);
2330 g_return_val_if_fail (eval != NULL, NULL);
2331 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2334 string_len = strlen (string);
2336 result = g_string_sized_new (string_len);
2338 /* run down the string making matches. */
2339 g_regex_match_full (regex, string, string_len, start_position,
2340 match_options, &match_info, &tmp_error);
2341 while (!done && g_match_info_matches (match_info))
2343 g_string_append_len (result,
2345 match_info->offsets[0] - str_pos);
2346 done = (*eval) (match_info, result, user_data);
2347 str_pos = match_info->offsets[1];
2348 g_match_info_next (match_info, &tmp_error);
2350 g_match_info_free (match_info);
2351 if (tmp_error != NULL)
2353 g_propagate_error (error, tmp_error);
2354 g_string_free (result, TRUE);
2358 g_string_append_len (result, string + str_pos, string_len - str_pos);
2359 return g_string_free (result, FALSE);
2363 * g_regex_check_replacement:
2364 * @replacement: the replacement string
2365 * @has_references: location to store information about
2366 * references in @replacement or %NULL
2367 * @error: location to store error
2369 * Checks whether @replacement is a valid replacement string
2370 * (see g_regex_replace()), i.e. that all escape sequences in
2373 * If @has_references is not %NULL then @replacement is checked
2374 * for pattern references. For instance, replacement text 'foo\n'
2375 * does not contain references and may be evaluated without information
2376 * about actual match, but '\0\1' (whole match followed by first
2377 * subpattern) requires valid #GMatchInfo object.
2379 * Returns: whether @replacement is a valid replacement string
2384 g_regex_check_replacement (const gchar *replacement,
2385 gboolean *has_references,
2391 list = split_replacement (replacement, &tmp);
2395 g_propagate_error (error, tmp);
2400 *has_references = interpolation_list_needs_match (list);
2402 g_list_foreach (list, (GFunc) free_interpolation_data, NULL);
2409 * g_regex_escape_string:
2410 * @string: the string to escape
2411 * @length: the length of @string, or -1 if @string is nul-terminated
2413 * Escapes the special characters used for regular expressions
2414 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2415 * function is useful to dynamically generate regular expressions.
2417 * @string can contain nul characters that are replaced with "\0",
2418 * in this case remember to specify the correct length of @string
2421 * Returns: a newly-allocated escaped string
2426 g_regex_escape_string (const gchar *string,
2430 const char *p, *piece_start, *end;
2432 g_return_val_if_fail (string != NULL, NULL);
2435 length = strlen (string);
2437 end = string + length;
2438 p = piece_start = string;
2439 escaped = g_string_sized_new (length + 1);
2460 if (p != piece_start)
2461 /* copy the previous piece. */
2462 g_string_append_len (escaped, piece_start, p - piece_start);
2463 g_string_append_c (escaped, '\\');
2465 g_string_append_c (escaped, '0');
2467 g_string_append_c (escaped, *p);
2471 p = g_utf8_next_char (p);
2476 if (piece_start < end)
2477 g_string_append_len (escaped, piece_start, end - piece_start);
2479 return g_string_free (escaped, FALSE);
2482 #define __G_REGEX_C__
2483 #include "galiasdef.c"