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 | \
116 G_REGEX_NEWLINE_ANYCRLF)
118 /* Mask of all the possible values for GRegexMatchFlags. */
119 #define G_REGEX_MATCH_MASK (G_REGEX_MATCH_ANCHORED | \
120 G_REGEX_MATCH_NOTBOL | \
121 G_REGEX_MATCH_NOTEOL | \
122 G_REGEX_MATCH_NOTEMPTY | \
123 G_REGEX_MATCH_PARTIAL | \
124 G_REGEX_MATCH_NEWLINE_CR | \
125 G_REGEX_MATCH_NEWLINE_LF | \
126 G_REGEX_MATCH_NEWLINE_CRLF | \
127 G_REGEX_MATCH_NEWLINE_ANY)
129 /* we rely on these flags having the same values */
130 G_STATIC_ASSERT (G_REGEX_CASELESS == PCRE_CASELESS);
131 G_STATIC_ASSERT (G_REGEX_MULTILINE == PCRE_MULTILINE);
132 G_STATIC_ASSERT (G_REGEX_DOTALL == PCRE_DOTALL);
133 G_STATIC_ASSERT (G_REGEX_EXTENDED == PCRE_EXTENDED);
134 G_STATIC_ASSERT (G_REGEX_ANCHORED == PCRE_ANCHORED);
135 G_STATIC_ASSERT (G_REGEX_DOLLAR_ENDONLY == PCRE_DOLLAR_ENDONLY);
136 G_STATIC_ASSERT (G_REGEX_UNGREEDY == PCRE_UNGREEDY);
137 G_STATIC_ASSERT (G_REGEX_NO_AUTO_CAPTURE == PCRE_NO_AUTO_CAPTURE);
138 G_STATIC_ASSERT (G_REGEX_DUPNAMES == PCRE_DUPNAMES);
139 G_STATIC_ASSERT (G_REGEX_NEWLINE_CR == PCRE_NEWLINE_CR);
140 G_STATIC_ASSERT (G_REGEX_NEWLINE_LF == PCRE_NEWLINE_LF);
141 G_STATIC_ASSERT (G_REGEX_NEWLINE_CRLF == PCRE_NEWLINE_CRLF);
142 G_STATIC_ASSERT (G_REGEX_NEWLINE_ANYCRLF == PCRE_NEWLINE_ANYCRLF);
144 G_STATIC_ASSERT (G_REGEX_MATCH_ANCHORED == PCRE_ANCHORED);
145 G_STATIC_ASSERT (G_REGEX_MATCH_NOTBOL == PCRE_NOTBOL);
146 G_STATIC_ASSERT (G_REGEX_MATCH_NOTEOL == PCRE_NOTEOL);
147 G_STATIC_ASSERT (G_REGEX_MATCH_NOTEMPTY == PCRE_NOTEMPTY);
148 G_STATIC_ASSERT (G_REGEX_MATCH_PARTIAL == PCRE_PARTIAL);
149 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_CR == PCRE_NEWLINE_CR);
150 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_LF == PCRE_NEWLINE_LF);
151 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_CRLF == PCRE_NEWLINE_CRLF);
152 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_ANY == PCRE_NEWLINE_ANY);
154 /* if the string is in UTF-8 use g_utf8_ functions, else use
156 #define NEXT_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
157 g_utf8_next_char (s) : \
159 #define PREV_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
160 g_utf8_prev_char (s) : \
165 volatile gint ref_count; /* the ref count */
166 GRegex *regex; /* the regex */
167 GRegexMatchFlags match_opts; /* options used at match time on the regex */
168 gint matches; /* number of matching sub patterns */
169 gint pos; /* position in the string where last match left off */
170 gint n_offsets; /* number of offsets */
171 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
172 gint *workspace; /* workspace for pcre_dfa_exec() */
173 gint n_workspace; /* number of workspace elements */
174 const gchar *string; /* string passed to the match function */
175 gssize string_len; /* length of string */
180 volatile gint ref_count; /* the ref count for the immutable part */
181 gchar *pattern; /* the pattern */
182 pcre *pcre_re; /* compiled form of the pattern */
183 GRegexCompileFlags compile_opts; /* options used at compile time on the pattern */
184 GRegexMatchFlags match_opts; /* options used at match time on the regex */
185 pcre_extra *extra; /* data stored when G_REGEX_OPTIMIZE is used */
188 /* TRUE if ret is an error code, FALSE otherwise. */
189 #define IS_PCRE_ERROR(ret) ((ret) < PCRE_ERROR_NOMATCH && (ret) != PCRE_ERROR_PARTIAL)
191 typedef struct _InterpolationData InterpolationData;
192 static gboolean interpolation_list_needs_match (GList *list);
193 static gboolean interpolate_replacement (const GMatchInfo *match_info,
196 static GList *split_replacement (const gchar *replacement,
198 static void free_interpolation_data (InterpolationData *data);
202 match_error (gint errcode)
206 case PCRE_ERROR_NOMATCH:
209 case PCRE_ERROR_NULL:
210 /* NULL argument, this should not happen in GRegex */
211 g_warning ("A NULL argument was passed to PCRE");
213 case PCRE_ERROR_BADOPTION:
214 return "bad options";
215 case PCRE_ERROR_BADMAGIC:
216 return _("corrupted object");
217 case PCRE_ERROR_UNKNOWN_OPCODE:
218 return N_("internal error or corrupted object");
219 case PCRE_ERROR_NOMEMORY:
220 return _("out of memory");
221 case PCRE_ERROR_NOSUBSTRING:
222 /* not used by pcre_exec() */
224 case PCRE_ERROR_MATCHLIMIT:
225 return _("backtracking limit reached");
226 case PCRE_ERROR_CALLOUT:
227 /* callouts are not implemented */
229 case PCRE_ERROR_BADUTF8:
230 case PCRE_ERROR_BADUTF8_OFFSET:
231 /* we do not check if strings are valid */
233 case PCRE_ERROR_PARTIAL:
236 case PCRE_ERROR_BADPARTIAL:
237 return _("the pattern contains items not supported for partial matching");
238 case PCRE_ERROR_INTERNAL:
239 return _("internal error");
240 case PCRE_ERROR_BADCOUNT:
241 /* negative ovecsize, this should not happen in GRegex */
242 g_warning ("A negative ovecsize was passed to PCRE");
244 case PCRE_ERROR_DFA_UITEM:
245 return _("the pattern contains items not supported for partial matching");
246 case PCRE_ERROR_DFA_UCOND:
247 return _("back references as conditions are not supported for partial matching");
248 case PCRE_ERROR_DFA_UMLIMIT:
249 /* the match_field field is not used in GRegex */
251 case PCRE_ERROR_DFA_WSSIZE:
252 /* handled expanding the workspace */
254 case PCRE_ERROR_DFA_RECURSE:
255 case PCRE_ERROR_RECURSIONLIMIT:
256 return _("recursion limit reached");
257 case PCRE_ERROR_NULLWSLIMIT:
258 return _("workspace limit for empty substrings reached");
259 case PCRE_ERROR_BADNEWLINE:
260 return _("invalid combination of newline flags");
261 case PCRE_ERROR_BADOFFSET:
262 return _("bad offset");
263 case PCRE_ERROR_SHORTUTF8:
264 return _("short utf8");
265 case PCRE_ERROR_RECURSELOOP:
266 return _("recursion loop");
270 return _("unknown error");
274 translate_compile_error (gint *errcode, const gchar **errmsg)
276 /* Compile errors are created adding 100 to the error code returned
278 * If errcode is known we put the translatable error message in
279 * erromsg. If errcode is unknown we put the generic
280 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
281 * untranslated error message returned by PCRE.
282 * Note that there can be more PCRE errors with the same GRegexError
283 * and that some PCRE errors are useless for us.
289 case G_REGEX_ERROR_STRAY_BACKSLASH:
290 *errmsg = _("\\ at end of pattern");
292 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
293 *errmsg = _("\\c at end of pattern");
295 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
296 *errmsg = _("unrecognized character follows \\");
298 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
299 *errmsg = _("numbers out of order in {} quantifier");
301 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
302 *errmsg = _("number too big in {} quantifier");
304 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
305 *errmsg = _("missing terminating ] for character class");
307 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
308 *errmsg = _("invalid escape sequence in character class");
310 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
311 *errmsg = _("range out of order in character class");
313 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
314 *errmsg = _("nothing to repeat");
316 case 111: /* internal error: unexpected repeat */
317 *errcode = G_REGEX_ERROR_INTERNAL;
318 *errmsg = _("unexpected repeat");
320 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
321 *errmsg = _("unrecognized character after (? or (?-");
323 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
324 *errmsg = _("POSIX named classes are supported only within a class");
326 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
327 *errmsg = _("missing terminating )");
329 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
330 *errmsg = _("reference to non-existent subpattern");
332 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
333 *errmsg = _("missing ) after comment");
335 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
336 *errmsg = _("regular expression is too large");
338 case G_REGEX_ERROR_MEMORY_ERROR:
339 *errmsg = _("failed to get memory");
341 case 122: /* unmatched parentheses */
342 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
343 *errmsg = _(") without opening (");
345 case 123: /* internal error: code overflow */
346 *errcode = G_REGEX_ERROR_INTERNAL;
347 *errmsg = _("code overflow");
349 case 124: /* "unrecognized character after (?<\0 */
350 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
351 *errmsg = _("unrecognized character after (?<");
353 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
354 *errmsg = _("lookbehind assertion is not fixed length");
356 case G_REGEX_ERROR_MALFORMED_CONDITION:
357 *errmsg = _("malformed number or name after (?(");
359 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
360 *errmsg = _("conditional group contains more than two branches");
362 case G_REGEX_ERROR_ASSERTION_EXPECTED:
363 *errmsg = _("assertion expected after (?(");
366 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
367 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
368 * sequences here, '(?-54' would be an example for the second group.
370 *errmsg = _("(?R or (?[+-]digits must be followed by )");
372 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
373 *errmsg = _("unknown POSIX class name");
375 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
376 *errmsg = _("POSIX collating elements are not supported");
378 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
379 *errmsg = _("character value in \\x{...} sequence is too large");
381 case G_REGEX_ERROR_INVALID_CONDITION:
382 *errmsg = _("invalid condition (?(0)");
384 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
385 *errmsg = _("\\C not allowed in lookbehind assertion");
387 case 137: /* PCRE does not support \\L, \\l, \\N{name}, \\U, or \\u\0 */
388 /* A number of Perl escapes are not handled by PCRE.
389 * Therefore it explicitly raises ERR37.
391 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
392 *errmsg = _("escapes \\L, \\l, \\N{name}, \\U, and \\u are not supported");
394 case G_REGEX_ERROR_INFINITE_LOOP:
395 *errmsg = _("recursive call could loop indefinitely");
397 case 141: /* unrecognized character after (?P\0 */
398 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
399 *errmsg = _("unrecognized character after (?P");
401 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
402 *errmsg = _("missing terminator in subpattern name");
404 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
405 *errmsg = _("two named subpatterns have the same name");
407 case G_REGEX_ERROR_MALFORMED_PROPERTY:
408 *errmsg = _("malformed \\P or \\p sequence");
410 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
411 *errmsg = _("unknown property name after \\P or \\p");
413 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
414 *errmsg = _("subpattern name is too long (maximum 32 characters)");
416 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
417 *errmsg = _("too many named subpatterns (maximum 10,000)");
419 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
420 *errmsg = _("octal value is greater than \\377");
422 case 152: /* internal error: overran compiling workspace */
423 *errcode = G_REGEX_ERROR_INTERNAL;
424 *errmsg = _("overran compiling workspace");
426 case 153: /* internal error: previously-checked referenced subpattern not found */
427 *errcode = G_REGEX_ERROR_INTERNAL;
428 *errmsg = _("previously-checked referenced subpattern not found");
430 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
431 *errmsg = _("DEFINE group contains more than one branch");
433 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
434 *errmsg = _("inconsistent NEWLINE options");
436 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
437 *errmsg = _("\\g is not followed by a braced, angle-bracketed, or quoted name or "
438 "number, or by a plain number");
440 case G_REGEX_ERROR_INVALID_RELATIVE_REFERENCE:
441 *errmsg = _("a numbered reference must not be zero");
443 case G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_FORBIDDEN:
444 *errmsg = _("an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)");
446 case G_REGEX_ERROR_UNKNOWN_BACKTRACKING_CONTROL_VERB:
447 *errmsg = _("(*VERB) not recognized");
449 case G_REGEX_ERROR_NUMBER_TOO_BIG:
450 *errmsg = _("number is too bug");
452 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME:
453 *errmsg = _("missing subpattern name after (?&");
455 case G_REGEX_ERROR_MISSING_DIGIT:
456 *errmsg = _("digit expected after (?+");
458 case G_REGEX_ERROR_EXTRA_SUBPATTERN_NAME:
459 *errmsg = _("different names for subpatterns of the same number are not allowed");
461 case G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_REQUIRED:
462 *errmsg = _("(*MARK) must have an argument");
464 case G_REGEX_ERROR_INVALID_CONTROL_CHAR:
465 *errmsg = _( "\\c must be followed by an ASCII character");
467 case G_REGEX_ERROR_MISSING_NAME:
468 *errmsg = _("\\k is not followed by a braced, angle-bracketed, or quoted name");
470 case G_REGEX_ERROR_NOT_SUPPORTED_IN_CLASS:
471 *errmsg = _("\\N is not supported in a class");
473 case G_REGEX_ERROR_TOO_MANY_FORWARD_REFERENCES:
474 *errmsg = _("too many forward references");
476 case G_REGEX_ERROR_NAME_TOO_LONG:
477 *errmsg = _("name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)");
480 case 116: /* erroffset passed as NULL */
481 /* This should not happen as we never pass a NULL erroffset */
482 g_warning ("erroffset passed as NULL");
483 *errcode = G_REGEX_ERROR_COMPILE;
485 case 117: /* unknown option bit(s) set */
486 /* This should not happen as we check options before passing them
487 * to pcre_compile2() */
488 g_warning ("unknown option bit(s) set");
489 *errcode = G_REGEX_ERROR_COMPILE;
491 case 132: /* this version of PCRE is compiled without UTF support */
492 case 144: /* invalid UTF-8 string */
493 case 145: /* support for \\P, \\p, and \\X has not been compiled */
494 case 167: /* this version of PCRE is not compiled with Unicode property support */
495 case 173: /* disallowed Unicode code point (>= 0xd800 && <= 0xdfff) */
496 case 174: /* invalid UTF-16 string */
497 /* These errors should not happen as we are using an UTF-8 and UCP-enabled PCRE
498 * and we do not check if strings are valid */
499 case 164: /* ] is an invalid data character in JavaScript compatibility mode */
500 /* This should not happen as we don't use PCRE_JAVASCRIPT_COMPAT */
501 g_warning ("%s", *errmsg);
502 *errcode = G_REGEX_ERROR_COMPILE;
504 case 170: /* internal error: unknown opcode in find_fixedlength() */
505 *errcode = G_REGEX_ERROR_INTERNAL;
509 *errcode = G_REGEX_ERROR_COMPILE;
516 match_info_new (const GRegex *regex,
523 GMatchInfo *match_info;
526 string_len = strlen (string);
528 match_info = g_new0 (GMatchInfo, 1);
529 match_info->ref_count = 1;
530 match_info->regex = g_regex_ref ((GRegex *)regex);
531 match_info->string = string;
532 match_info->string_len = string_len;
533 match_info->matches = PCRE_ERROR_NOMATCH;
534 match_info->pos = start_position;
535 match_info->match_opts = match_options;
539 /* These values should be enough for most cases, if they are not
540 * enough g_regex_match_all_full() will expand them. */
541 match_info->n_offsets = 24;
542 match_info->n_workspace = 100;
543 match_info->workspace = g_new (gint, match_info->n_workspace);
548 pcre_fullinfo (regex->pcre_re, regex->extra,
549 PCRE_INFO_CAPTURECOUNT, &capture_count);
550 match_info->n_offsets = (capture_count + 1) * 3;
553 match_info->offsets = g_new0 (gint, match_info->n_offsets);
554 /* Set an invalid position for the previous match. */
555 match_info->offsets[0] = -1;
556 match_info->offsets[1] = -1;
562 * g_match_info_get_regex:
563 * @match_info: a #GMatchInfo
565 * Returns #GRegex object used in @match_info. It belongs to Glib
566 * and must not be freed. Use g_regex_ref() if you need to keep it
567 * after you free @match_info object.
569 * Returns: #GRegex object used in @match_info
574 g_match_info_get_regex (const GMatchInfo *match_info)
576 g_return_val_if_fail (match_info != NULL, NULL);
577 return match_info->regex;
581 * g_match_info_get_string:
582 * @match_info: a #GMatchInfo
584 * Returns the string searched with @match_info. This is the
585 * string passed to g_regex_match() or g_regex_replace() so
586 * you may not free it before calling this function.
588 * Returns: the string searched with @match_info
593 g_match_info_get_string (const GMatchInfo *match_info)
595 g_return_val_if_fail (match_info != NULL, NULL);
596 return match_info->string;
601 * @match_info: a #GMatchInfo
603 * Increases reference count of @match_info by 1.
605 * Returns: @match_info
610 g_match_info_ref (GMatchInfo *match_info)
612 g_return_val_if_fail (match_info != NULL, NULL);
613 g_atomic_int_inc (&match_info->ref_count);
618 * g_match_info_unref:
619 * @match_info: a #GMatchInfo
621 * Decreases reference count of @match_info by 1. When reference count drops
622 * to zero, it frees all the memory associated with the match_info structure.
627 g_match_info_unref (GMatchInfo *match_info)
629 if (g_atomic_int_dec_and_test (&match_info->ref_count))
631 g_regex_unref (match_info->regex);
632 g_free (match_info->offsets);
633 g_free (match_info->workspace);
640 * @match_info: (allow-none): a #GMatchInfo, or %NULL
642 * If @match_info is not %NULL, calls g_match_info_unref(); otherwise does
648 g_match_info_free (GMatchInfo *match_info)
650 if (match_info == NULL)
653 g_match_info_unref (match_info);
658 * @match_info: a #GMatchInfo structure
659 * @error: location to store the error occurring, or %NULL to ignore errors
661 * Scans for the next match using the same parameters of the previous
662 * call to g_regex_match_full() or g_regex_match() that returned
665 * The match is done on the string passed to the match function, so you
666 * cannot free it before calling this function.
668 * Returns: %TRUE is the string matched, %FALSE otherwise
673 g_match_info_next (GMatchInfo *match_info,
676 gint prev_match_start;
679 g_return_val_if_fail (match_info != NULL, FALSE);
680 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
681 g_return_val_if_fail (match_info->pos >= 0, FALSE);
683 prev_match_start = match_info->offsets[0];
684 prev_match_end = 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->matches = pcre_exec (match_info->regex->pcre_re,
695 match_info->regex->extra,
697 match_info->string_len,
699 match_info->regex->match_opts | match_info->match_opts,
701 match_info->n_offsets);
702 if (IS_PCRE_ERROR (match_info->matches))
704 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
705 _("Error while matching regular expression %s: %s"),
706 match_info->regex->pattern, match_error (match_info->matches));
710 /* avoid infinite loops if the pattern is an empty string or something
712 if (match_info->pos == match_info->offsets[1])
714 if (match_info->pos > match_info->string_len)
716 /* we have reached the end of the string */
717 match_info->pos = -1;
718 match_info->matches = PCRE_ERROR_NOMATCH;
722 match_info->pos = NEXT_CHAR (match_info->regex,
723 &match_info->string[match_info->pos]) -
728 match_info->pos = match_info->offsets[1];
731 /* it's possible to get two identical matches when we are matching
732 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
733 * the string is "RegExTest" we have:
734 * - search at position 0: match from 0 to 0
735 * - search at position 1: match from 3 to 3
736 * - search at position 3: match from 3 to 3 (duplicate)
737 * - search at position 4: match from 5 to 5
738 * - search at position 5: match from 5 to 5 (duplicate)
739 * - search at position 6: no match -> stop
740 * so we have to ignore the duplicates.
741 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
742 if (match_info->matches >= 0 &&
743 prev_match_start == match_info->offsets[0] &&
744 prev_match_end == match_info->offsets[1])
746 /* ignore this match and search the next one */
747 return g_match_info_next (match_info, error);
750 return match_info->matches >= 0;
754 * g_match_info_matches:
755 * @match_info: a #GMatchInfo structure
757 * Returns whether the previous match operation succeeded.
759 * Returns: %TRUE if the previous match operation succeeded,
765 g_match_info_matches (const GMatchInfo *match_info)
767 g_return_val_if_fail (match_info != NULL, FALSE);
769 return match_info->matches >= 0;
773 * g_match_info_get_match_count:
774 * @match_info: a #GMatchInfo structure
776 * Retrieves the number of matched substrings (including substring 0,
777 * that is the whole matched text), so 1 is returned if the pattern
778 * has no substrings in it and 0 is returned if the match failed.
780 * If the last match was obtained using the DFA algorithm, that is
781 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
782 * count is not that of the number of capturing parentheses but that of
783 * the number of matched substrings.
785 * Returns: Number of matched substrings, or -1 if an error occurred
790 g_match_info_get_match_count (const GMatchInfo *match_info)
792 g_return_val_if_fail (match_info, -1);
794 if (match_info->matches == PCRE_ERROR_NOMATCH)
797 else if (match_info->matches < PCRE_ERROR_NOMATCH)
802 return match_info->matches;
806 * g_match_info_is_partial_match:
807 * @match_info: a #GMatchInfo structure
809 * Usually if the string passed to g_regex_match*() matches as far as
810 * it goes, but is too short to match the entire pattern, %FALSE is
811 * returned. There are circumstances where it might be helpful to
812 * distinguish this case from other cases in which there is no match.
814 * Consider, for example, an application where a human is required to
815 * type in data for a field with specific formatting requirements. An
816 * example might be a date in the form ddmmmyy, defined by the pattern
817 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
818 * If the application sees the user’s keystrokes one by one, and can
819 * check that what has been typed so far is potentially valid, it is
820 * able to raise an error as soon as a mistake is made.
822 * GRegex supports the concept of partial matching by means of the
823 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
824 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
825 * for a complete match, %FALSE otherwise. But, when these functions
826 * return %FALSE, you can check if the match was partial calling
827 * g_match_info_is_partial_match().
829 * When using partial matching you cannot use g_match_info_fetch*().
831 * Because of the way certain internal optimizations are implemented
832 * the partial matching algorithm cannot be used with all patterns.
833 * So repeated single characters such as "a{2,4}" and repeated single
834 * meta-sequences such as "\d+" are not permitted if the maximum number
835 * of occurrences is greater than one. Optional items such as "\d?"
836 * (where the maximum is one) are permitted. Quantifiers with any values
837 * are permitted after parentheses, so the invalid examples above can be
838 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
839 * for a pattern that does not conform to the restrictions, matching
840 * functions return an error.
842 * Returns: %TRUE if the match was partial, %FALSE otherwise
847 g_match_info_is_partial_match (const GMatchInfo *match_info)
849 g_return_val_if_fail (match_info != NULL, FALSE);
851 return match_info->matches == PCRE_ERROR_PARTIAL;
855 * g_match_info_expand_references:
856 * @match_info: (allow-none): a #GMatchInfo or %NULL
857 * @string_to_expand: the string to expand
858 * @error: location to store the error occurring, or %NULL to ignore errors
860 * Returns a new string containing the text in @string_to_expand with
861 * references and escape sequences expanded. References refer to the last
862 * match done with @string against @regex and have the same syntax used by
865 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
866 * passed to g_regex_new().
868 * The backreferences are extracted from the string passed to the match
869 * function, so you cannot call this function after freeing the string.
871 * @match_info may be %NULL in which case @string_to_expand must not
872 * contain references. For instance "foo\n" does not refer to an actual
873 * pattern and '\n' merely will be replaced with \n character,
874 * while to expand "\0" (whole match) one needs the result of a match.
875 * Use g_regex_check_replacement() to find out whether @string_to_expand
876 * contains references.
878 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
883 g_match_info_expand_references (const GMatchInfo *match_info,
884 const gchar *string_to_expand,
889 GError *tmp_error = NULL;
891 g_return_val_if_fail (string_to_expand != NULL, NULL);
892 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
894 list = split_replacement (string_to_expand, &tmp_error);
895 if (tmp_error != NULL)
897 g_propagate_error (error, tmp_error);
901 if (!match_info && interpolation_list_needs_match (list))
903 g_critical ("String '%s' contains references to the match, can't "
904 "expand references without GMatchInfo object",
909 result = g_string_sized_new (strlen (string_to_expand));
910 interpolate_replacement (match_info, result, list);
912 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
914 return g_string_free (result, FALSE);
918 * g_match_info_fetch:
919 * @match_info: #GMatchInfo structure
920 * @match_num: number of the sub expression
922 * Retrieves the text matching the @match_num<!-- -->'th capturing
923 * parentheses. 0 is the full text of the match, 1 is the first paren
924 * set, 2 the second, and so on.
926 * If @match_num is a valid sub pattern but it didn't match anything
927 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
928 * string is returned.
930 * If the match was obtained using the DFA algorithm, that is using
931 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
932 * string is not that of a set of parentheses but that of a matched
933 * substring. Substrings are matched in reverse order of length, so
934 * 0 is the longest match.
936 * The string is fetched from the string passed to the match function,
937 * so you cannot call this function after freeing the string.
939 * Returns: (allow-none): The matched substring, or %NULL if an error
940 * occurred. You have to free the string yourself
945 g_match_info_fetch (const GMatchInfo *match_info,
948 /* we cannot use pcre_get_substring() because it allocates the
949 * string using pcre_malloc(). */
953 g_return_val_if_fail (match_info != NULL, NULL);
954 g_return_val_if_fail (match_num >= 0, NULL);
956 /* match_num does not exist or it didn't matched, i.e. matching "b"
957 * against "(a)?b" then group 0 is empty. */
958 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
960 else if (start == -1)
961 match = g_strdup ("");
963 match = g_strndup (&match_info->string[start], end - start);
969 * g_match_info_fetch_pos:
970 * @match_info: #GMatchInfo structure
971 * @match_num: number of the sub expression
972 * @start_pos: (out) (allow-none): pointer to location where to store
973 * the start position, or %NULL
974 * @end_pos: (out) (allow-none): pointer to location where to store
975 * the end position, or %NULL
977 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
978 * parentheses. 0 is the full text of the match, 1 is the first
979 * paren set, 2 the second, and so on.
981 * If @match_num is a valid sub pattern but it didn't match anything
982 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
983 * and @end_pos are set to -1 and %TRUE is returned.
985 * If the match was obtained using the DFA algorithm, that is using
986 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
987 * position is not that of a set of parentheses but that of a matched
988 * substring. Substrings are matched in reverse order of length, so
989 * 0 is the longest match.
991 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
992 * the position cannot be fetched, @start_pos and @end_pos are left
998 g_match_info_fetch_pos (const GMatchInfo *match_info,
1003 g_return_val_if_fail (match_info != NULL, FALSE);
1004 g_return_val_if_fail (match_num >= 0, FALSE);
1006 /* make sure the sub expression number they're requesting is less than
1007 * the total number of sub expressions that were matched. */
1008 if (match_num >= match_info->matches)
1011 if (start_pos != NULL)
1012 *start_pos = match_info->offsets[2 * match_num];
1014 if (end_pos != NULL)
1015 *end_pos = match_info->offsets[2 * match_num + 1];
1021 * Returns number of first matched subpattern with name @name.
1022 * There may be more than one in case when DUPNAMES is used,
1023 * and not all subpatterns with that name match;
1024 * pcre_get_stringnumber() does not work in that case.
1027 get_matched_substring_number (const GMatchInfo *match_info,
1031 gchar *first, *last;
1034 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
1035 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
1037 /* This code is copied from pcre_get.c: get_first_set() */
1038 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
1046 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
1048 gint n = (entry[0] << 8) + entry[1];
1049 if (match_info->offsets[n*2] >= 0)
1053 return (first[0] << 8) + first[1];
1057 * g_match_info_fetch_named:
1058 * @match_info: #GMatchInfo structure
1059 * @name: name of the subexpression
1061 * Retrieves the text matching the capturing parentheses named @name.
1063 * If @name is a valid sub pattern name but it didn't match anything
1064 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1065 * then an empty string is returned.
1067 * The string is fetched from the string passed to the match function,
1068 * so you cannot call this function after freeing the string.
1070 * Returns: (allow-none): The matched substring, or %NULL if an error
1071 * occurred. You have to free the string yourself
1076 g_match_info_fetch_named (const GMatchInfo *match_info,
1079 /* we cannot use pcre_get_named_substring() because it allocates the
1080 * string using pcre_malloc(). */
1083 g_return_val_if_fail (match_info != NULL, NULL);
1084 g_return_val_if_fail (name != NULL, NULL);
1086 num = get_matched_substring_number (match_info, name);
1090 return g_match_info_fetch (match_info, num);
1094 * g_match_info_fetch_named_pos:
1095 * @match_info: #GMatchInfo structure
1096 * @name: name of the subexpression
1097 * @start_pos: (out) (allow-none): pointer to location where to store
1098 * the start position, or %NULL
1099 * @end_pos: (out) (allow-none): pointer to location where to store
1100 * the end position, or %NULL
1102 * Retrieves the position in bytes of the capturing parentheses named @name.
1104 * If @name is a valid sub pattern name but it didn't match anything
1105 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1106 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
1108 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
1109 * If the position cannot be fetched, @start_pos and @end_pos
1110 * are left unchanged.
1115 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
1122 g_return_val_if_fail (match_info != NULL, FALSE);
1123 g_return_val_if_fail (name != NULL, FALSE);
1125 num = get_matched_substring_number (match_info, name);
1129 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1133 * g_match_info_fetch_all:
1134 * @match_info: a #GMatchInfo structure
1136 * Bundles up pointers to each of the matching substrings from a match
1137 * and stores them in an array of gchar pointers. The first element in
1138 * the returned array is the match number 0, i.e. the entire matched
1141 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1142 * "b" against "(a)?b") then an empty string is inserted.
1144 * If the last match was obtained using the DFA algorithm, that is using
1145 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1146 * strings are not that matched by sets of parentheses but that of the
1147 * matched substring. Substrings are matched in reverse order of length,
1148 * so the first one is the longest match.
1150 * The strings are fetched from the string passed to the match function,
1151 * so you cannot call this function after freeing the string.
1153 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1154 * It must be freed using g_strfreev(). If the previous match failed
1160 g_match_info_fetch_all (const GMatchInfo *match_info)
1162 /* we cannot use pcre_get_substring_list() because the returned value
1163 * isn't suitable for g_strfreev(). */
1167 g_return_val_if_fail (match_info != NULL, NULL);
1169 if (match_info->matches < 0)
1172 result = g_new (gchar *, match_info->matches + 1);
1173 for (i = 0; i < match_info->matches; i++)
1174 result[i] = g_match_info_fetch (match_info, i);
1184 g_regex_error_quark (void)
1186 static GQuark error_quark = 0;
1188 if (error_quark == 0)
1189 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1198 * Increases reference count of @regex by 1.
1205 g_regex_ref (GRegex *regex)
1207 g_return_val_if_fail (regex != NULL, NULL);
1208 g_atomic_int_inc (®ex->ref_count);
1216 * Decreases reference count of @regex by 1. When reference count drops
1217 * to zero, it frees all the memory associated with the regex structure.
1222 g_regex_unref (GRegex *regex)
1224 g_return_if_fail (regex != NULL);
1226 if (g_atomic_int_dec_and_test (®ex->ref_count))
1228 g_free (regex->pattern);
1229 if (regex->pcre_re != NULL)
1230 pcre_free (regex->pcre_re);
1231 if (regex->extra != NULL)
1232 pcre_free (regex->extra);
1239 * @pattern: the regular expression
1240 * @compile_options: compile options for the regular expression, or 0
1241 * @match_options: match options for the regular expression, or 0
1242 * @error: return location for a #GError
1244 * Compiles the regular expression to an internal form, and does
1245 * the initial setup of the #GRegex structure.
1247 * Returns: a #GRegex structure. Call g_regex_unref() when you
1253 g_regex_new (const gchar *pattern,
1254 GRegexCompileFlags compile_options,
1255 GRegexMatchFlags match_options,
1260 const gchar *errmsg;
1263 gboolean optimize = FALSE;
1264 static gsize initialised;
1265 unsigned long int pcre_compile_options;
1267 g_return_val_if_fail (pattern != NULL, NULL);
1268 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1269 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1270 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1272 if (g_once_init_enter (&initialised))
1277 pcre_config (PCRE_CONFIG_UTF8, &support);
1280 msg = N_("PCRE library is compiled without UTF8 support");
1281 g_critical ("%s", msg);
1282 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1286 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &support);
1289 msg = N_("PCRE library is compiled without UTF8 properties support");
1290 g_critical ("%s", msg);
1291 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE, gettext (msg));
1295 g_once_init_leave (&initialised, TRUE);
1298 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1299 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1300 if (compile_options & G_REGEX_OPTIMIZE)
1303 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1304 * instead uses UTF-8 only if required with PCRE_UTF8. */
1305 if (compile_options & G_REGEX_RAW)
1308 compile_options &= ~G_REGEX_RAW;
1313 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1314 match_options |= PCRE_NO_UTF8_CHECK;
1317 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1318 * not for the system one. */
1319 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1320 !(compile_options & G_REGEX_NEWLINE_LF))
1322 compile_options |= PCRE_NEWLINE_ANY;
1325 compile_options |= PCRE_UCP;
1327 /* compile the pattern */
1328 re = pcre_compile2 (pattern, compile_options, &errcode,
1329 &errmsg, &erroffset, NULL);
1331 /* if the compilation failed, set the error member and return
1337 /* Translate the PCRE error code to GRegexError and use a translated
1338 * error message if possible */
1339 translate_compile_error (&errcode, &errmsg);
1341 /* PCRE uses byte offsets but we want to show character offsets */
1342 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1344 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1345 _("Error while compiling regular "
1346 "expression %s at char %d: %s"),
1347 pattern, erroffset, errmsg);
1348 g_propagate_error (error, tmp_error);
1353 /* For options set at the beginning of the pattern, pcre puts them into
1354 * compile options, e.g. "(?i)foo" will make the pcre structure store
1355 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1356 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1357 compile_options = pcre_compile_options;
1359 if (!(compile_options & G_REGEX_DUPNAMES))
1361 gboolean jchanged = FALSE;
1362 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1364 compile_options |= G_REGEX_DUPNAMES;
1367 regex = g_new0 (GRegex, 1);
1368 regex->ref_count = 1;
1369 regex->pattern = g_strdup (pattern);
1370 regex->pcre_re = re;
1371 regex->compile_opts = compile_options;
1372 regex->match_opts = match_options;
1376 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1379 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1380 G_REGEX_ERROR_OPTIMIZE,
1381 _("Error while optimizing "
1382 "regular expression %s: %s"),
1385 g_propagate_error (error, tmp_error);
1387 g_regex_unref (regex);
1396 * g_regex_get_pattern:
1397 * @regex: a #GRegex structure
1399 * Gets the pattern string associated with @regex, i.e. a copy of
1400 * the string passed to g_regex_new().
1402 * Returns: the pattern of @regex
1407 g_regex_get_pattern (const GRegex *regex)
1409 g_return_val_if_fail (regex != NULL, NULL);
1411 return regex->pattern;
1415 * g_regex_get_max_backref:
1418 * Returns the number of the highest back reference
1419 * in the pattern, or 0 if the pattern does not contain
1422 * Returns: the number of the highest back reference
1427 g_regex_get_max_backref (const GRegex *regex)
1431 pcre_fullinfo (regex->pcre_re, regex->extra,
1432 PCRE_INFO_BACKREFMAX, &value);
1438 * g_regex_get_capture_count:
1441 * Returns the number of capturing subpatterns in the pattern.
1443 * Returns: the number of capturing subpatterns
1448 g_regex_get_capture_count (const GRegex *regex)
1452 pcre_fullinfo (regex->pcre_re, regex->extra,
1453 PCRE_INFO_CAPTURECOUNT, &value);
1459 * g_regex_get_compile_flags:
1462 * Returns the compile options that @regex was created with.
1464 * Returns: flags from #GRegexCompileFlags
1469 g_regex_get_compile_flags (const GRegex *regex)
1471 g_return_val_if_fail (regex != NULL, 0);
1473 return regex->compile_opts;
1477 * g_regex_get_match_flags:
1480 * Returns the match options that @regex was created with.
1482 * Returns: flags from #GRegexMatchFlags
1487 g_regex_get_match_flags (const GRegex *regex)
1489 g_return_val_if_fail (regex != NULL, 0);
1491 return regex->match_opts;
1495 * g_regex_match_simple:
1496 * @pattern: the regular expression
1497 * @string: the string to scan for matches
1498 * @compile_options: compile options for the regular expression, or 0
1499 * @match_options: match options, or 0
1501 * Scans for a match in @string for @pattern.
1503 * This function is equivalent to g_regex_match() but it does not
1504 * require to compile the pattern with g_regex_new(), avoiding some
1505 * lines of code when you need just to do a match without extracting
1506 * substrings, capture counts, and so on.
1508 * If this function is to be called on the same @pattern more than
1509 * once, it's more efficient to compile the pattern once with
1510 * g_regex_new() and then use g_regex_match().
1512 * Returns: %TRUE if the string matched, %FALSE otherwise
1517 g_regex_match_simple (const gchar *pattern,
1518 const gchar *string,
1519 GRegexCompileFlags compile_options,
1520 GRegexMatchFlags match_options)
1525 regex = g_regex_new (pattern, compile_options, 0, NULL);
1528 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1529 g_regex_unref (regex);
1535 * @regex: a #GRegex structure from g_regex_new()
1536 * @string: the string to scan for matches
1537 * @match_options: match options
1538 * @match_info: (out) (allow-none): pointer to location where to store
1539 * the #GMatchInfo, or %NULL if you do not need it
1541 * Scans for a match in string for the pattern in @regex.
1542 * The @match_options are combined with the match options specified
1543 * when the @regex structure was created, letting you have more
1544 * flexibility in reusing #GRegex structures.
1546 * A #GMatchInfo structure, used to get information on the match,
1547 * is stored in @match_info if not %NULL. Note that if @match_info
1548 * is not %NULL then it is created even if the function returns %FALSE,
1549 * i.e. you must free it regardless if regular expression actually matched.
1551 * To retrieve all the non-overlapping matches of the pattern in
1552 * string you can use g_match_info_next().
1556 * print_uppercase_words (const gchar *string)
1558 * /* Print all uppercase-only words. */
1560 * GMatchInfo *match_info;
1562 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1563 * g_regex_match (regex, string, 0, &match_info);
1564 * while (g_match_info_matches (match_info))
1566 * gchar *word = g_match_info_fetch (match_info, 0);
1567 * g_print ("Found: %s\n", word);
1569 * g_match_info_next (match_info, NULL);
1571 * g_match_info_free (match_info);
1572 * g_regex_unref (regex);
1576 * @string is not copied and is used in #GMatchInfo internally. If
1577 * you use any #GMatchInfo method (except g_match_info_free()) after
1578 * freeing or modifying @string then the behaviour is undefined.
1580 * Returns: %TRUE is the string matched, %FALSE otherwise
1585 g_regex_match (const GRegex *regex,
1586 const gchar *string,
1587 GRegexMatchFlags match_options,
1588 GMatchInfo **match_info)
1590 return g_regex_match_full (regex, string, -1, 0, match_options,
1595 * g_regex_match_full:
1596 * @regex: a #GRegex structure from g_regex_new()
1597 * @string: (array length=string_len): the string to scan for matches
1598 * @string_len: the length of @string, or -1 if @string is nul-terminated
1599 * @start_position: starting index of the string to match
1600 * @match_options: match options
1601 * @match_info: (out) (allow-none): pointer to location where to store
1602 * the #GMatchInfo, or %NULL if you do not need it
1603 * @error: location to store the error occurring, or %NULL to ignore errors
1605 * Scans for a match in string for the pattern in @regex.
1606 * The @match_options are combined with the match options specified
1607 * when the @regex structure was created, letting you have more
1608 * flexibility in reusing #GRegex structures.
1610 * Setting @start_position differs from just passing over a shortened
1611 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1612 * that begins with any kind of lookbehind assertion, such as "\b".
1614 * A #GMatchInfo structure, used to get information on the match, is
1615 * stored in @match_info if not %NULL. Note that if @match_info is
1616 * not %NULL then it is created even if the function returns %FALSE,
1617 * i.e. you must free it regardless if regular expression actually
1620 * @string is not copied and is used in #GMatchInfo internally. If
1621 * you use any #GMatchInfo method (except g_match_info_free()) after
1622 * freeing or modifying @string then the behaviour is undefined.
1624 * To retrieve all the non-overlapping matches of the pattern in
1625 * string you can use g_match_info_next().
1629 * print_uppercase_words (const gchar *string)
1631 * /* Print all uppercase-only words. */
1633 * GMatchInfo *match_info;
1634 * GError *error = NULL;
1636 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1637 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1638 * while (g_match_info_matches (match_info))
1640 * gchar *word = g_match_info_fetch (match_info, 0);
1641 * g_print ("Found: %s\n", word);
1643 * g_match_info_next (match_info, &error);
1645 * g_match_info_free (match_info);
1646 * g_regex_unref (regex);
1647 * if (error != NULL)
1649 * g_printerr ("Error while matching: %s\n", error->message);
1650 * g_error_free (error);
1655 * Returns: %TRUE is the string matched, %FALSE otherwise
1660 g_regex_match_full (const GRegex *regex,
1661 const gchar *string,
1663 gint start_position,
1664 GRegexMatchFlags match_options,
1665 GMatchInfo **match_info,
1671 g_return_val_if_fail (regex != NULL, FALSE);
1672 g_return_val_if_fail (string != NULL, FALSE);
1673 g_return_val_if_fail (start_position >= 0, FALSE);
1674 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1675 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1677 info = match_info_new (regex, string, string_len, start_position,
1678 match_options, FALSE);
1679 match_ok = g_match_info_next (info, error);
1680 if (match_info != NULL)
1683 g_match_info_free (info);
1689 * g_regex_match_all:
1690 * @regex: a #GRegex structure from g_regex_new()
1691 * @string: the string to scan for matches
1692 * @match_options: match options
1693 * @match_info: (out) (allow-none): pointer to location where to store
1694 * the #GMatchInfo, or %NULL if you do not need it
1696 * Using the standard algorithm for regular expression matching only
1697 * the longest match in the string is retrieved. This function uses
1698 * a different algorithm so it can retrieve all the possible matches.
1699 * For more documentation see g_regex_match_all_full().
1701 * A #GMatchInfo structure, used to get information on the match, is
1702 * stored in @match_info if not %NULL. Note that if @match_info is
1703 * not %NULL then it is created even if the function returns %FALSE,
1704 * i.e. you must free it regardless if regular expression actually
1707 * @string is not copied and is used in #GMatchInfo internally. If
1708 * you use any #GMatchInfo method (except g_match_info_free()) after
1709 * freeing or modifying @string then the behaviour is undefined.
1711 * Returns: %TRUE is the string matched, %FALSE otherwise
1716 g_regex_match_all (const GRegex *regex,
1717 const gchar *string,
1718 GRegexMatchFlags match_options,
1719 GMatchInfo **match_info)
1721 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1726 * g_regex_match_all_full:
1727 * @regex: a #GRegex structure from g_regex_new()
1728 * @string: (array length=string_len): the string to scan for matches
1729 * @string_len: the length of @string, or -1 if @string is nul-terminated
1730 * @start_position: starting index of the string to match
1731 * @match_options: match options
1732 * @match_info: (out) (allow-none): pointer to location where to store
1733 * the #GMatchInfo, or %NULL if you do not need it
1734 * @error: location to store the error occurring, or %NULL to ignore errors
1736 * Using the standard algorithm for regular expression matching only
1737 * the longest match in the string is retrieved, it is not possible
1738 * to obtain all the available matches. For instance matching
1739 * "<a> <b> <c>" against the pattern "<.*>"
1740 * you get "<a> <b> <c>".
1742 * This function uses a different algorithm (called DFA, i.e. deterministic
1743 * finite automaton), so it can retrieve all the possible matches, all
1744 * starting at the same point in the string. For instance matching
1745 * "<a> <b> <c>" against the pattern "<.*>"
1746 * you would obtain three matches: "<a> <b> <c>",
1747 * "<a> <b>" and "<a>".
1749 * The number of matched strings is retrieved using
1750 * g_match_info_get_match_count(). To obtain the matched strings and
1751 * their position you can use, respectively, g_match_info_fetch() and
1752 * g_match_info_fetch_pos(). Note that the strings are returned in
1753 * reverse order of length; that is, the longest matching string is
1756 * Note that the DFA algorithm is slower than the standard one and it
1757 * is not able to capture substrings, so backreferences do not work.
1759 * Setting @start_position differs from just passing over a shortened
1760 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1761 * that begins with any kind of lookbehind assertion, such as "\b".
1763 * A #GMatchInfo structure, used to get information on the match, is
1764 * stored in @match_info if not %NULL. Note that if @match_info is
1765 * not %NULL then it is created even if the function returns %FALSE,
1766 * i.e. you must free it regardless if regular expression actually
1769 * @string is not copied and is used in #GMatchInfo internally. If
1770 * you use any #GMatchInfo method (except g_match_info_free()) after
1771 * freeing or modifying @string then the behaviour is undefined.
1773 * Returns: %TRUE is the string matched, %FALSE otherwise
1778 g_regex_match_all_full (const GRegex *regex,
1779 const gchar *string,
1781 gint start_position,
1782 GRegexMatchFlags match_options,
1783 GMatchInfo **match_info,
1789 g_return_val_if_fail (regex != NULL, FALSE);
1790 g_return_val_if_fail (string != NULL, FALSE);
1791 g_return_val_if_fail (start_position >= 0, FALSE);
1792 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1793 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1795 info = match_info_new (regex, string, string_len, start_position,
1796 match_options, TRUE);
1802 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1803 info->string, info->string_len,
1805 regex->match_opts | match_options,
1806 info->offsets, info->n_offsets,
1807 info->workspace, info->n_workspace);
1808 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1810 /* info->workspace is too small. */
1811 info->n_workspace *= 2;
1812 info->workspace = g_realloc (info->workspace,
1813 info->n_workspace * sizeof (gint));
1816 else if (info->matches == 0)
1818 /* info->offsets is too small. */
1819 info->n_offsets *= 2;
1820 info->offsets = g_realloc (info->offsets,
1821 info->n_offsets * sizeof (gint));
1824 else if (IS_PCRE_ERROR (info->matches))
1826 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1827 _("Error while matching regular expression %s: %s"),
1828 regex->pattern, match_error (info->matches));
1832 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1835 if (match_info != NULL)
1838 g_match_info_free (info);
1840 return info->matches >= 0;
1844 * g_regex_get_string_number:
1845 * @regex: #GRegex structure
1846 * @name: name of the subexpression
1848 * Retrieves the number of the subexpression named @name.
1850 * Returns: The number of the subexpression or -1 if @name
1856 g_regex_get_string_number (const GRegex *regex,
1861 g_return_val_if_fail (regex != NULL, -1);
1862 g_return_val_if_fail (name != NULL, -1);
1864 num = pcre_get_stringnumber (regex->pcre_re, name);
1865 if (num == PCRE_ERROR_NOSUBSTRING)
1872 * g_regex_split_simple:
1873 * @pattern: the regular expression
1874 * @string: the string to scan for matches
1875 * @compile_options: compile options for the regular expression, or 0
1876 * @match_options: match options, or 0
1878 * Breaks the string on the pattern, and returns an array of
1879 * the tokens. If the pattern contains capturing parentheses,
1880 * then the text for each of the substrings will also be returned.
1881 * If the pattern does not match anywhere in the string, then the
1882 * whole string is returned as the first token.
1884 * This function is equivalent to g_regex_split() but it does
1885 * not require to compile the pattern with g_regex_new(), avoiding
1886 * some lines of code when you need just to do a split without
1887 * extracting substrings, capture counts, and so on.
1889 * If this function is to be called on the same @pattern more than
1890 * once, it's more efficient to compile the pattern once with
1891 * g_regex_new() and then use g_regex_split().
1893 * As a special case, the result of splitting the empty string ""
1894 * is an empty vector, not a vector containing a single string.
1895 * The reason for this special case is that being able to represent
1896 * a empty vector is typically more useful than consistent handling
1897 * of empty elements. If you do need to represent empty elements,
1898 * you'll need to check for the empty string before calling this
1901 * A pattern that can match empty strings splits @string into
1902 * separate characters wherever it matches the empty string between
1903 * characters. For example splitting "ab c" using as a separator
1904 * "\s*", you will get "a", "b" and "c".
1906 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1911 g_regex_split_simple (const gchar *pattern,
1912 const gchar *string,
1913 GRegexCompileFlags compile_options,
1914 GRegexMatchFlags match_options)
1919 regex = g_regex_new (pattern, compile_options, 0, NULL);
1923 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1924 g_regex_unref (regex);
1930 * @regex: a #GRegex structure
1931 * @string: the string to split with the pattern
1932 * @match_options: match time option flags
1934 * Breaks the string on the pattern, and returns an array of the tokens.
1935 * If the pattern contains capturing parentheses, then the text for each
1936 * of the substrings will also be returned. If the pattern does not match
1937 * anywhere in the string, then the whole string is returned as the first
1940 * As a special case, the result of splitting the empty string "" is an
1941 * empty vector, not a vector containing a single string. The reason for
1942 * this special case is that being able to represent a empty vector is
1943 * typically more useful than consistent handling of empty elements. If
1944 * you do need to represent empty elements, you'll need to check for the
1945 * empty string before calling this function.
1947 * A pattern that can match empty strings splits @string into separate
1948 * characters wherever it matches the empty string between characters.
1949 * For example splitting "ab c" using as a separator "\s*", you will get
1952 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1957 g_regex_split (const GRegex *regex,
1958 const gchar *string,
1959 GRegexMatchFlags match_options)
1961 return g_regex_split_full (regex, string, -1, 0,
1962 match_options, 0, NULL);
1966 * g_regex_split_full:
1967 * @regex: a #GRegex structure
1968 * @string: (array length=string_len): the string to split with the pattern
1969 * @string_len: the length of @string, or -1 if @string is nul-terminated
1970 * @start_position: starting index of the string to match
1971 * @match_options: match time option flags
1972 * @max_tokens: the maximum number of tokens to split @string into.
1973 * If this is less than 1, the string is split completely
1974 * @error: return location for a #GError
1976 * Breaks the string on the pattern, and returns an array of the tokens.
1977 * If the pattern contains capturing parentheses, then the text for each
1978 * of the substrings will also be returned. If the pattern does not match
1979 * anywhere in the string, then the whole string is returned as the first
1982 * As a special case, the result of splitting the empty string "" is an
1983 * empty vector, not a vector containing a single string. The reason for
1984 * this special case is that being able to represent a empty vector is
1985 * typically more useful than consistent handling of empty elements. If
1986 * you do need to represent empty elements, you'll need to check for the
1987 * empty string before calling this function.
1989 * A pattern that can match empty strings splits @string into separate
1990 * characters wherever it matches the empty string between characters.
1991 * For example splitting "ab c" using as a separator "\s*", you will get
1994 * Setting @start_position differs from just passing over a shortened
1995 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1996 * that begins with any kind of lookbehind assertion, such as "\b".
1998 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
2003 g_regex_split_full (const GRegex *regex,
2004 const gchar *string,
2006 gint start_position,
2007 GRegexMatchFlags match_options,
2011 GError *tmp_error = NULL;
2012 GMatchInfo *match_info;
2017 /* position of the last separator. */
2018 gint last_separator_end;
2019 /* was the last match 0 bytes long? */
2020 gboolean last_match_is_empty;
2021 /* the returned array of char **s */
2022 gchar **string_list;
2024 g_return_val_if_fail (regex != NULL, NULL);
2025 g_return_val_if_fail (string != NULL, NULL);
2026 g_return_val_if_fail (start_position >= 0, NULL);
2027 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2028 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2030 if (max_tokens <= 0)
2031 max_tokens = G_MAXINT;
2034 string_len = strlen (string);
2036 /* zero-length string */
2037 if (string_len - start_position == 0)
2038 return g_new0 (gchar *, 1);
2040 if (max_tokens == 1)
2042 string_list = g_new0 (gchar *, 2);
2043 string_list[0] = g_strndup (&string[start_position],
2044 string_len - start_position);
2050 last_separator_end = start_position;
2051 last_match_is_empty = FALSE;
2053 match_ok = g_regex_match_full (regex, string, string_len, start_position,
2054 match_options, &match_info, &tmp_error);
2056 while (tmp_error == NULL)
2060 last_match_is_empty =
2061 (match_info->offsets[0] == match_info->offsets[1]);
2063 /* we need to skip empty separators at the same position of the end
2064 * of another separator. e.g. the string is "a b" and the separator
2065 * is " *", so from 1 to 2 we have a match and at position 2 we have
2066 * an empty match. */
2067 if (last_separator_end != match_info->offsets[1])
2072 token = g_strndup (string + last_separator_end,
2073 match_info->offsets[0] - last_separator_end);
2074 list = g_list_prepend (list, token);
2077 /* if there were substrings, these need to be added to
2079 match_count = g_match_info_get_match_count (match_info);
2080 if (match_count > 1)
2082 for (i = 1; i < match_count; i++)
2083 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
2089 /* if there was no match, copy to end of string. */
2090 if (!last_match_is_empty)
2092 gchar *token = g_strndup (string + last_separator_end,
2093 match_info->string_len - last_separator_end);
2094 list = g_list_prepend (list, token);
2096 /* no more tokens, end the loop. */
2100 /* -1 to leave room for the last part. */
2101 if (token_count >= max_tokens - 1)
2103 /* we have reached the maximum number of tokens, so we copy
2104 * the remaining part of the string. */
2105 if (last_match_is_empty)
2107 /* the last match was empty, so we have moved one char
2108 * after the real position to avoid empty matches at the
2110 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
2112 /* the if is needed in the case we have terminated the available
2113 * tokens, but we are at the end of the string, so there are no
2114 * characters left to copy. */
2115 if (string_len > match_info->pos)
2117 gchar *token = g_strndup (string + match_info->pos,
2118 string_len - match_info->pos);
2119 list = g_list_prepend (list, token);
2125 last_separator_end = match_info->pos;
2126 if (last_match_is_empty)
2127 /* if the last match was empty, g_match_info_next() has moved
2128 * forward to avoid infinite loops, but we still need to copy that
2130 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
2132 match_ok = g_match_info_next (match_info, &tmp_error);
2134 g_match_info_free (match_info);
2135 if (tmp_error != NULL)
2137 g_propagate_error (error, tmp_error);
2138 g_list_free_full (list, g_free);
2139 match_info->pos = -1;
2143 string_list = g_new (gchar *, g_list_length (list) + 1);
2145 for (last = g_list_last (list); last; last = g_list_previous (last))
2146 string_list[i++] = last->data;
2147 string_list[i] = NULL;
2156 REPL_TYPE_CHARACTER,
2157 REPL_TYPE_SYMBOLIC_REFERENCE,
2158 REPL_TYPE_NUMERIC_REFERENCE,
2159 REPL_TYPE_CHANGE_CASE
2164 CHANGE_CASE_NONE = 1 << 0,
2165 CHANGE_CASE_UPPER = 1 << 1,
2166 CHANGE_CASE_LOWER = 1 << 2,
2167 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2168 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2169 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2170 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2171 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2174 struct _InterpolationData
2180 ChangeCase change_case;
2184 free_interpolation_data (InterpolationData *data)
2186 g_free (data->text);
2190 static const gchar *
2191 expand_escape (const gchar *replacement,
2193 InterpolationData *data,
2198 const gchar *error_detail;
2200 GError *tmp_error = NULL;
2208 data->type = REPL_TYPE_CHARACTER;
2213 data->type = REPL_TYPE_CHARACTER;
2218 data->type = REPL_TYPE_CHARACTER;
2223 data->type = REPL_TYPE_CHARACTER;
2228 data->type = REPL_TYPE_CHARACTER;
2233 data->type = REPL_TYPE_CHARACTER;
2238 data->type = REPL_TYPE_CHARACTER;
2243 data->type = REPL_TYPE_CHARACTER;
2253 h = g_ascii_xdigit_value (*p);
2256 error_detail = _("hexadecimal digit or '}' expected");
2267 for (i = 0; i < 2; i++)
2269 h = g_ascii_xdigit_value (*p);
2272 error_detail = _("hexadecimal digit expected");
2279 data->type = REPL_TYPE_STRING;
2280 data->text = g_new0 (gchar, 8);
2281 g_unichar_to_utf8 (x, data->text);
2285 data->type = REPL_TYPE_CHANGE_CASE;
2286 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2290 data->type = REPL_TYPE_CHANGE_CASE;
2291 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2295 data->type = REPL_TYPE_CHANGE_CASE;
2296 data->change_case = CHANGE_CASE_LOWER;
2300 data->type = REPL_TYPE_CHANGE_CASE;
2301 data->change_case = CHANGE_CASE_UPPER;
2305 data->type = REPL_TYPE_CHANGE_CASE;
2306 data->change_case = CHANGE_CASE_NONE;
2312 error_detail = _("missing '<' in symbolic reference");
2321 error_detail = _("unfinished symbolic reference");
2328 error_detail = _("zero-length symbolic reference");
2331 if (g_ascii_isdigit (*q))
2336 h = g_ascii_digit_value (*q);
2339 error_detail = _("digit expected");
2348 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2355 if (!g_ascii_isalnum (*r))
2357 error_detail = _("illegal symbolic reference");
2364 data->text = g_strndup (q, p - q);
2365 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2370 /* if \0 is followed by a number is an octal number representing a
2371 * character, else it is a numeric reference. */
2372 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2375 p = g_utf8_next_char (p);
2388 for (i = 0; i < 3; i++)
2390 h = g_ascii_digit_value (*p);
2400 if (i == 2 && base == 10)
2406 if (base == 8 || i == 3)
2408 data->type = REPL_TYPE_STRING;
2409 data->text = g_new0 (gchar, 8);
2410 g_unichar_to_utf8 (x, data->text);
2414 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2419 error_detail = _("stray final '\\'");
2423 error_detail = _("unknown escape sequence");
2430 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2431 tmp_error = g_error_new (G_REGEX_ERROR,
2432 G_REGEX_ERROR_REPLACE,
2433 _("Error while parsing replacement "
2434 "text \"%s\" at char %lu: %s"),
2436 (gulong)(p - replacement),
2438 g_propagate_error (error, tmp_error);
2444 split_replacement (const gchar *replacement,
2448 InterpolationData *data;
2449 const gchar *p, *start;
2451 start = p = replacement;
2456 data = g_new0 (InterpolationData, 1);
2457 start = p = expand_escape (replacement, p, data, error);
2460 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2461 free_interpolation_data (data);
2465 list = g_list_prepend (list, data);
2470 if (*p == '\\' || *p == '\0')
2474 data = g_new0 (InterpolationData, 1);
2475 data->text = g_strndup (start, p - start);
2476 data->type = REPL_TYPE_STRING;
2477 list = g_list_prepend (list, data);
2483 return g_list_reverse (list);
2486 /* Change the case of c based on change_case. */
2487 #define CHANGE_CASE(c, change_case) \
2488 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2489 g_unichar_tolower (c) : \
2490 g_unichar_toupper (c))
2493 string_append (GString *string,
2495 ChangeCase *change_case)
2499 if (text[0] == '\0')
2502 if (*change_case == CHANGE_CASE_NONE)
2504 g_string_append (string, text);
2506 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2508 c = g_utf8_get_char (text);
2509 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2510 g_string_append (string, g_utf8_next_char (text));
2511 *change_case = CHANGE_CASE_NONE;
2515 while (*text != '\0')
2517 c = g_utf8_get_char (text);
2518 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2519 text = g_utf8_next_char (text);
2525 interpolate_replacement (const GMatchInfo *match_info,
2530 InterpolationData *idata;
2532 ChangeCase change_case = CHANGE_CASE_NONE;
2534 for (list = data; list; list = list->next)
2537 switch (idata->type)
2539 case REPL_TYPE_STRING:
2540 string_append (result, idata->text, &change_case);
2542 case REPL_TYPE_CHARACTER:
2543 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2544 if (change_case & CHANGE_CASE_SINGLE_MASK)
2545 change_case = CHANGE_CASE_NONE;
2547 case REPL_TYPE_NUMERIC_REFERENCE:
2548 match = g_match_info_fetch (match_info, idata->num);
2551 string_append (result, match, &change_case);
2555 case REPL_TYPE_SYMBOLIC_REFERENCE:
2556 match = g_match_info_fetch_named (match_info, idata->text);
2559 string_append (result, match, &change_case);
2563 case REPL_TYPE_CHANGE_CASE:
2564 change_case = idata->change_case;
2572 /* whether actual match_info is needed for replacement, i.e.
2573 * whether there are references
2576 interpolation_list_needs_match (GList *list)
2578 while (list != NULL)
2580 InterpolationData *data = list->data;
2582 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2583 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2596 * @regex: a #GRegex structure
2597 * @string: (array length=string_len): the string to perform matches against
2598 * @string_len: the length of @string, or -1 if @string is nul-terminated
2599 * @start_position: starting index of the string to match
2600 * @replacement: text to replace each match with
2601 * @match_options: options for the match
2602 * @error: location to store the error occurring, or %NULL to ignore errors
2604 * Replaces all occurrences of the pattern in @regex with the
2605 * replacement text. Backreferences of the form '\number' or
2606 * '\g<number>' in the replacement text are interpolated by the
2607 * number-th captured subexpression of the match, '\g<name>' refers
2608 * to the captured subexpression with the given name. '\0' refers to the
2609 * complete match, but '\0' followed by a number is the octal representation
2610 * of a character. To include a literal '\' in the replacement, write '\\'.
2611 * There are also escapes that changes the case of the following text:
2614 * <varlistentry><term>\l</term>
2616 * <para>Convert to lower case the next character</para>
2619 * <varlistentry><term>\u</term>
2621 * <para>Convert to upper case the next character</para>
2624 * <varlistentry><term>\L</term>
2626 * <para>Convert to lower case till \E</para>
2629 * <varlistentry><term>\U</term>
2631 * <para>Convert to upper case till \E</para>
2634 * <varlistentry><term>\E</term>
2636 * <para>End case modification</para>
2641 * If you do not need to use backreferences use g_regex_replace_literal().
2643 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2644 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2645 * you can use g_regex_replace_literal().
2647 * Setting @start_position differs from just passing over a shortened
2648 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2649 * begins with any kind of lookbehind assertion, such as "\b".
2651 * Returns: a newly allocated string containing the replacements
2656 g_regex_replace (const GRegex *regex,
2657 const gchar *string,
2659 gint start_position,
2660 const gchar *replacement,
2661 GRegexMatchFlags match_options,
2666 GError *tmp_error = NULL;
2668 g_return_val_if_fail (regex != NULL, NULL);
2669 g_return_val_if_fail (string != NULL, NULL);
2670 g_return_val_if_fail (start_position >= 0, NULL);
2671 g_return_val_if_fail (replacement != NULL, NULL);
2672 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2673 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2675 list = split_replacement (replacement, &tmp_error);
2676 if (tmp_error != NULL)
2678 g_propagate_error (error, tmp_error);
2682 result = g_regex_replace_eval (regex,
2683 string, string_len, start_position,
2685 interpolate_replacement,
2688 if (tmp_error != NULL)
2689 g_propagate_error (error, tmp_error);
2691 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2697 literal_replacement (const GMatchInfo *match_info,
2701 g_string_append (result, data);
2706 * g_regex_replace_literal:
2707 * @regex: a #GRegex structure
2708 * @string: (array length=string_len): the string to perform matches against
2709 * @string_len: the length of @string, or -1 if @string is nul-terminated
2710 * @start_position: starting index of the string to match
2711 * @replacement: text to replace each match with
2712 * @match_options: options for the match
2713 * @error: location to store the error occurring, or %NULL to ignore errors
2715 * Replaces all occurrences of the pattern in @regex with the
2716 * replacement text. @replacement is replaced literally, to
2717 * include backreferences use g_regex_replace().
2719 * Setting @start_position differs from just passing over a
2720 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2721 * case of a pattern that begins with any kind of lookbehind
2722 * assertion, such as "\b".
2724 * Returns: a newly allocated string containing the replacements
2729 g_regex_replace_literal (const GRegex *regex,
2730 const gchar *string,
2732 gint start_position,
2733 const gchar *replacement,
2734 GRegexMatchFlags match_options,
2737 g_return_val_if_fail (replacement != NULL, NULL);
2738 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2740 return g_regex_replace_eval (regex,
2741 string, string_len, start_position,
2743 literal_replacement,
2744 (gpointer)replacement,
2749 * g_regex_replace_eval:
2750 * @regex: a #GRegex structure from g_regex_new()
2751 * @string: (array length=string_len): string to perform matches against
2752 * @string_len: the length of @string, or -1 if @string is nul-terminated
2753 * @start_position: starting index of the string to match
2754 * @match_options: options for the match
2755 * @eval: a function to call for each match
2756 * @user_data: user data to pass to the function
2757 * @error: location to store the error occurring, or %NULL to ignore errors
2759 * Replaces occurrences of the pattern in regex with the output of
2760 * @eval for that occurrence.
2762 * Setting @start_position differs from just passing over a shortened
2763 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2764 * that begins with any kind of lookbehind assertion, such as "\b".
2766 * The following example uses g_regex_replace_eval() to replace multiple
2770 * eval_cb (const GMatchInfo *info,
2777 * match = g_match_info_fetch (info, 0);
2778 * r = g_hash_table_lookup ((GHashTable *)data, match);
2779 * g_string_append (res, r);
2791 * h = g_hash_table_new (g_str_hash, g_str_equal);
2793 * g_hash_table_insert (h, "1", "ONE");
2794 * g_hash_table_insert (h, "2", "TWO");
2795 * g_hash_table_insert (h, "3", "THREE");
2796 * g_hash_table_insert (h, "4", "FOUR");
2798 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2799 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2800 * g_hash_table_destroy (h);
2805 * Returns: a newly allocated string containing the replacements
2810 g_regex_replace_eval (const GRegex *regex,
2811 const gchar *string,
2813 gint start_position,
2814 GRegexMatchFlags match_options,
2815 GRegexEvalCallback eval,
2819 GMatchInfo *match_info;
2822 gboolean done = FALSE;
2823 GError *tmp_error = NULL;
2825 g_return_val_if_fail (regex != NULL, NULL);
2826 g_return_val_if_fail (string != NULL, NULL);
2827 g_return_val_if_fail (start_position >= 0, NULL);
2828 g_return_val_if_fail (eval != NULL, NULL);
2829 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2832 string_len = strlen (string);
2834 result = g_string_sized_new (string_len);
2836 /* run down the string making matches. */
2837 g_regex_match_full (regex, string, string_len, start_position,
2838 match_options, &match_info, &tmp_error);
2839 while (!done && g_match_info_matches (match_info))
2841 g_string_append_len (result,
2843 match_info->offsets[0] - str_pos);
2844 done = (*eval) (match_info, result, user_data);
2845 str_pos = match_info->offsets[1];
2846 g_match_info_next (match_info, &tmp_error);
2848 g_match_info_free (match_info);
2849 if (tmp_error != NULL)
2851 g_propagate_error (error, tmp_error);
2852 g_string_free (result, TRUE);
2856 g_string_append_len (result, string + str_pos, string_len - str_pos);
2857 return g_string_free (result, FALSE);
2861 * g_regex_check_replacement:
2862 * @replacement: the replacement string
2863 * @has_references: (out) (allow-none): location to store information about
2864 * references in @replacement or %NULL
2865 * @error: location to store error
2867 * Checks whether @replacement is a valid replacement string
2868 * (see g_regex_replace()), i.e. that all escape sequences in
2871 * If @has_references is not %NULL then @replacement is checked
2872 * for pattern references. For instance, replacement text 'foo\n'
2873 * does not contain references and may be evaluated without information
2874 * about actual match, but '\0\1' (whole match followed by first
2875 * subpattern) requires valid #GMatchInfo object.
2877 * Returns: whether @replacement is a valid replacement string
2882 g_regex_check_replacement (const gchar *replacement,
2883 gboolean *has_references,
2889 list = split_replacement (replacement, &tmp);
2893 g_propagate_error (error, tmp);
2898 *has_references = interpolation_list_needs_match (list);
2900 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2906 * g_regex_escape_nul:
2907 * @string: the string to escape
2908 * @length: the length of @string
2910 * Escapes the nul characters in @string to "\x00". It can be used
2911 * to compile a regex with embedded nul characters.
2913 * For completeness, @length can be -1 for a nul-terminated string.
2914 * In this case the output string will be of course equal to @string.
2916 * Returns: a newly-allocated escaped string
2921 g_regex_escape_nul (const gchar *string,
2925 const gchar *p, *piece_start, *end;
2928 g_return_val_if_fail (string != NULL, NULL);
2931 return g_strdup (string);
2933 end = string + length;
2934 p = piece_start = string;
2935 escaped = g_string_sized_new (length + 1);
2943 if (p != piece_start)
2945 /* copy the previous piece. */
2946 g_string_append_len (escaped, piece_start, p - piece_start);
2948 if ((backslashes & 1) == 0)
2949 g_string_append_c (escaped, '\\');
2950 g_string_append_c (escaped, 'x');
2951 g_string_append_c (escaped, '0');
2952 g_string_append_c (escaped, '0');
2962 p = g_utf8_next_char (p);
2967 if (piece_start < end)
2968 g_string_append_len (escaped, piece_start, end - piece_start);
2970 return g_string_free (escaped, FALSE);
2974 * g_regex_escape_string:
2975 * @string: (array length=length): the string to escape
2976 * @length: the length of @string, or -1 if @string is nul-terminated
2978 * Escapes the special characters used for regular expressions
2979 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
2980 * function is useful to dynamically generate regular expressions.
2982 * @string can contain nul characters that are replaced with "\0",
2983 * in this case remember to specify the correct length of @string
2986 * Returns: a newly-allocated escaped string
2991 g_regex_escape_string (const gchar *string,
2995 const char *p, *piece_start, *end;
2997 g_return_val_if_fail (string != NULL, NULL);
3000 length = strlen (string);
3002 end = string + length;
3003 p = piece_start = string;
3004 escaped = g_string_sized_new (length + 1);
3025 if (p != piece_start)
3026 /* copy the previous piece. */
3027 g_string_append_len (escaped, piece_start, p - piece_start);
3028 g_string_append_c (escaped, '\\');
3030 g_string_append_c (escaped, '0');
3032 g_string_append_c (escaped, *p);
3036 p = g_utf8_next_char (p);
3041 if (piece_start < end)
3042 g_string_append_len (escaped, piece_start, end - piece_start);
3044 return g_string_free (escaped, FALSE);