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 | \
119 /* Mask of all the possible values for GRegexMatchFlags. */
120 #define G_REGEX_MATCH_MASK (G_REGEX_MATCH_ANCHORED | \
121 G_REGEX_MATCH_NOTBOL | \
122 G_REGEX_MATCH_NOTEOL | \
123 G_REGEX_MATCH_NOTEMPTY | \
124 G_REGEX_MATCH_PARTIAL | \
125 G_REGEX_MATCH_NEWLINE_CR | \
126 G_REGEX_MATCH_NEWLINE_LF | \
127 G_REGEX_MATCH_NEWLINE_CRLF | \
128 G_REGEX_MATCH_NEWLINE_ANY | \
129 G_REGEX_MATCH_NEWLINE_ANYCRLF | \
130 G_REGEX_MATCH_BSR_ANYCRLF | \
131 G_REGEX_MATCH_BSR_ANY)
133 /* we rely on these flags having the same values */
134 G_STATIC_ASSERT (G_REGEX_CASELESS == PCRE_CASELESS);
135 G_STATIC_ASSERT (G_REGEX_MULTILINE == PCRE_MULTILINE);
136 G_STATIC_ASSERT (G_REGEX_DOTALL == PCRE_DOTALL);
137 G_STATIC_ASSERT (G_REGEX_EXTENDED == PCRE_EXTENDED);
138 G_STATIC_ASSERT (G_REGEX_ANCHORED == PCRE_ANCHORED);
139 G_STATIC_ASSERT (G_REGEX_DOLLAR_ENDONLY == PCRE_DOLLAR_ENDONLY);
140 G_STATIC_ASSERT (G_REGEX_UNGREEDY == PCRE_UNGREEDY);
141 G_STATIC_ASSERT (G_REGEX_NO_AUTO_CAPTURE == PCRE_NO_AUTO_CAPTURE);
142 G_STATIC_ASSERT (G_REGEX_DUPNAMES == PCRE_DUPNAMES);
143 G_STATIC_ASSERT (G_REGEX_NEWLINE_CR == PCRE_NEWLINE_CR);
144 G_STATIC_ASSERT (G_REGEX_NEWLINE_LF == PCRE_NEWLINE_LF);
145 G_STATIC_ASSERT (G_REGEX_NEWLINE_CRLF == PCRE_NEWLINE_CRLF);
146 G_STATIC_ASSERT (G_REGEX_NEWLINE_ANYCRLF == PCRE_NEWLINE_ANYCRLF);
147 G_STATIC_ASSERT (G_REGEX_BSR_ANYCRLF == PCRE_BSR_ANYCRLF);
149 G_STATIC_ASSERT (G_REGEX_MATCH_ANCHORED == PCRE_ANCHORED);
150 G_STATIC_ASSERT (G_REGEX_MATCH_NOTBOL == PCRE_NOTBOL);
151 G_STATIC_ASSERT (G_REGEX_MATCH_NOTEOL == PCRE_NOTEOL);
152 G_STATIC_ASSERT (G_REGEX_MATCH_NOTEMPTY == PCRE_NOTEMPTY);
153 G_STATIC_ASSERT (G_REGEX_MATCH_PARTIAL == PCRE_PARTIAL);
154 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_CR == PCRE_NEWLINE_CR);
155 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_LF == PCRE_NEWLINE_LF);
156 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_CRLF == PCRE_NEWLINE_CRLF);
157 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_ANY == PCRE_NEWLINE_ANY);
158 G_STATIC_ASSERT (G_REGEX_MATCH_NEWLINE_ANYCRLF == PCRE_NEWLINE_ANYCRLF);
159 G_STATIC_ASSERT (G_REGEX_MATCH_BSR_ANYCRLF == PCRE_BSR_ANYCRLF);
160 G_STATIC_ASSERT (G_REGEX_MATCH_BSR_ANY == PCRE_BSR_UNICODE);
162 /* if the string is in UTF-8 use g_utf8_ functions, else use
164 #define NEXT_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
165 g_utf8_next_char (s) : \
167 #define PREV_CHAR(re, s) (((re)->compile_opts & PCRE_UTF8) ? \
168 g_utf8_prev_char (s) : \
173 volatile gint ref_count; /* the ref count */
174 GRegex *regex; /* the regex */
175 GRegexMatchFlags match_opts; /* options used at match time on the regex */
176 gint matches; /* number of matching sub patterns */
177 gint pos; /* position in the string where last match left off */
178 gint n_offsets; /* number of offsets */
179 gint *offsets; /* array of offsets paired 0,1 ; 2,3 ; 3,4 etc */
180 gint *workspace; /* workspace for pcre_dfa_exec() */
181 gint n_workspace; /* number of workspace elements */
182 const gchar *string; /* string passed to the match function */
183 gssize string_len; /* length of string */
188 volatile gint ref_count; /* the ref count for the immutable part */
189 gchar *pattern; /* the pattern */
190 pcre *pcre_re; /* compiled form of the pattern */
191 GRegexCompileFlags compile_opts; /* options used at compile time on the pattern */
192 GRegexMatchFlags match_opts; /* options used at match time on the regex */
193 pcre_extra *extra; /* data stored when G_REGEX_OPTIMIZE is used */
196 /* TRUE if ret is an error code, FALSE otherwise. */
197 #define IS_PCRE_ERROR(ret) ((ret) < PCRE_ERROR_NOMATCH && (ret) != PCRE_ERROR_PARTIAL)
199 typedef struct _InterpolationData InterpolationData;
200 static gboolean interpolation_list_needs_match (GList *list);
201 static gboolean interpolate_replacement (const GMatchInfo *match_info,
204 static GList *split_replacement (const gchar *replacement,
206 static void free_interpolation_data (InterpolationData *data);
210 match_error (gint errcode)
214 case PCRE_ERROR_NOMATCH:
217 case PCRE_ERROR_NULL:
218 /* NULL argument, this should not happen in GRegex */
219 g_warning ("A NULL argument was passed to PCRE");
221 case PCRE_ERROR_BADOPTION:
222 return "bad options";
223 case PCRE_ERROR_BADMAGIC:
224 return _("corrupted object");
225 case PCRE_ERROR_UNKNOWN_OPCODE:
226 return N_("internal error or corrupted object");
227 case PCRE_ERROR_NOMEMORY:
228 return _("out of memory");
229 case PCRE_ERROR_NOSUBSTRING:
230 /* not used by pcre_exec() */
232 case PCRE_ERROR_MATCHLIMIT:
233 return _("backtracking limit reached");
234 case PCRE_ERROR_CALLOUT:
235 /* callouts are not implemented */
237 case PCRE_ERROR_BADUTF8:
238 case PCRE_ERROR_BADUTF8_OFFSET:
239 /* we do not check if strings are valid */
241 case PCRE_ERROR_PARTIAL:
244 case PCRE_ERROR_BADPARTIAL:
245 return _("the pattern contains items not supported for partial matching");
246 case PCRE_ERROR_INTERNAL:
247 return _("internal error");
248 case PCRE_ERROR_BADCOUNT:
249 /* negative ovecsize, this should not happen in GRegex */
250 g_warning ("A negative ovecsize was passed to PCRE");
252 case PCRE_ERROR_DFA_UITEM:
253 return _("the pattern contains items not supported for partial matching");
254 case PCRE_ERROR_DFA_UCOND:
255 return _("back references as conditions are not supported for partial matching");
256 case PCRE_ERROR_DFA_UMLIMIT:
257 /* the match_field field is not used in GRegex */
259 case PCRE_ERROR_DFA_WSSIZE:
260 /* handled expanding the workspace */
262 case PCRE_ERROR_DFA_RECURSE:
263 case PCRE_ERROR_RECURSIONLIMIT:
264 return _("recursion limit reached");
265 case PCRE_ERROR_BADNEWLINE:
266 return _("invalid combination of newline flags");
267 case PCRE_ERROR_BADOFFSET:
268 return _("bad offset");
269 case PCRE_ERROR_SHORTUTF8:
270 return _("short utf8");
271 case PCRE_ERROR_RECURSELOOP:
272 return _("recursion loop");
276 return _("unknown error");
280 translate_compile_error (gint *errcode, const gchar **errmsg)
282 /* Compile errors are created adding 100 to the error code returned
284 * If errcode is known we put the translatable error message in
285 * erromsg. If errcode is unknown we put the generic
286 * G_REGEX_ERROR_COMPILE error code in errcode and keep the
287 * untranslated error message returned by PCRE.
288 * Note that there can be more PCRE errors with the same GRegexError
289 * and that some PCRE errors are useless for us.
295 case G_REGEX_ERROR_STRAY_BACKSLASH:
296 *errmsg = _("\\ at end of pattern");
298 case G_REGEX_ERROR_MISSING_CONTROL_CHAR:
299 *errmsg = _("\\c at end of pattern");
301 case G_REGEX_ERROR_UNRECOGNIZED_ESCAPE:
302 *errmsg = _("unrecognized character follows \\");
304 case G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER:
305 *errmsg = _("numbers out of order in {} quantifier");
307 case G_REGEX_ERROR_QUANTIFIER_TOO_BIG:
308 *errmsg = _("number too big in {} quantifier");
310 case G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS:
311 *errmsg = _("missing terminating ] for character class");
313 case G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS:
314 *errmsg = _("invalid escape sequence in character class");
316 case G_REGEX_ERROR_RANGE_OUT_OF_ORDER:
317 *errmsg = _("range out of order in character class");
319 case G_REGEX_ERROR_NOTHING_TO_REPEAT:
320 *errmsg = _("nothing to repeat");
322 case 111: /* internal error: unexpected repeat */
323 *errcode = G_REGEX_ERROR_INTERNAL;
324 *errmsg = _("unexpected repeat");
326 case G_REGEX_ERROR_UNRECOGNIZED_CHARACTER:
327 *errmsg = _("unrecognized character after (? or (?-");
329 case G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS:
330 *errmsg = _("POSIX named classes are supported only within a class");
332 case G_REGEX_ERROR_UNMATCHED_PARENTHESIS:
333 *errmsg = _("missing terminating )");
335 case G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE:
336 *errmsg = _("reference to non-existent subpattern");
338 case G_REGEX_ERROR_UNTERMINATED_COMMENT:
339 *errmsg = _("missing ) after comment");
341 case G_REGEX_ERROR_EXPRESSION_TOO_LARGE:
342 *errmsg = _("regular expression is too large");
344 case G_REGEX_ERROR_MEMORY_ERROR:
345 *errmsg = _("failed to get memory");
347 case 122: /* unmatched parentheses */
348 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
349 *errmsg = _(") without opening (");
351 case 123: /* internal error: code overflow */
352 *errcode = G_REGEX_ERROR_INTERNAL;
353 *errmsg = _("code overflow");
355 case 124: /* "unrecognized character after (?<\0 */
356 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
357 *errmsg = _("unrecognized character after (?<");
359 case G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND:
360 *errmsg = _("lookbehind assertion is not fixed length");
362 case G_REGEX_ERROR_MALFORMED_CONDITION:
363 *errmsg = _("malformed number or name after (?(");
365 case G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES:
366 *errmsg = _("conditional group contains more than two branches");
368 case G_REGEX_ERROR_ASSERTION_EXPECTED:
369 *errmsg = _("assertion expected after (?(");
372 *errcode = G_REGEX_ERROR_UNMATCHED_PARENTHESIS;
373 /* translators: '(?R' and '(?[+-]digits' are both meant as (groups of)
374 * sequences here, '(?-54' would be an example for the second group.
376 *errmsg = _("(?R or (?[+-]digits must be followed by )");
378 case G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME:
379 *errmsg = _("unknown POSIX class name");
381 case G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED:
382 *errmsg = _("POSIX collating elements are not supported");
384 case G_REGEX_ERROR_HEX_CODE_TOO_LARGE:
385 *errmsg = _("character value in \\x{...} sequence is too large");
387 case G_REGEX_ERROR_INVALID_CONDITION:
388 *errmsg = _("invalid condition (?(0)");
390 case G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND:
391 *errmsg = _("\\C not allowed in lookbehind assertion");
393 case 137: /* PCRE does not support \\L, \\l, \\N{name}, \\U, or \\u\0 */
394 /* A number of Perl escapes are not handled by PCRE.
395 * Therefore it explicitly raises ERR37.
397 *errcode = G_REGEX_ERROR_UNRECOGNIZED_ESCAPE;
398 *errmsg = _("escapes \\L, \\l, \\N{name}, \\U, and \\u are not supported");
400 case G_REGEX_ERROR_INFINITE_LOOP:
401 *errmsg = _("recursive call could loop indefinitely");
403 case 141: /* unrecognized character after (?P\0 */
404 *errcode = G_REGEX_ERROR_UNRECOGNIZED_CHARACTER;
405 *errmsg = _("unrecognized character after (?P");
407 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR:
408 *errmsg = _("missing terminator in subpattern name");
410 case G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME:
411 *errmsg = _("two named subpatterns have the same name");
413 case G_REGEX_ERROR_MALFORMED_PROPERTY:
414 *errmsg = _("malformed \\P or \\p sequence");
416 case G_REGEX_ERROR_UNKNOWN_PROPERTY:
417 *errmsg = _("unknown property name after \\P or \\p");
419 case G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG:
420 *errmsg = _("subpattern name is too long (maximum 32 characters)");
422 case G_REGEX_ERROR_TOO_MANY_SUBPATTERNS:
423 *errmsg = _("too many named subpatterns (maximum 10,000)");
425 case G_REGEX_ERROR_INVALID_OCTAL_VALUE:
426 *errmsg = _("octal value is greater than \\377");
428 case 152: /* internal error: overran compiling workspace */
429 *errcode = G_REGEX_ERROR_INTERNAL;
430 *errmsg = _("overran compiling workspace");
432 case 153: /* internal error: previously-checked referenced subpattern not found */
433 *errcode = G_REGEX_ERROR_INTERNAL;
434 *errmsg = _("previously-checked referenced subpattern not found");
436 case G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE:
437 *errmsg = _("DEFINE group contains more than one branch");
439 case G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS:
440 *errmsg = _("inconsistent NEWLINE options");
442 case G_REGEX_ERROR_MISSING_BACK_REFERENCE:
443 *errmsg = _("\\g is not followed by a braced, angle-bracketed, or quoted name or "
444 "number, or by a plain number");
446 case G_REGEX_ERROR_INVALID_RELATIVE_REFERENCE:
447 *errmsg = _("a numbered reference must not be zero");
449 case G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_FORBIDDEN:
450 *errmsg = _("an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)");
452 case G_REGEX_ERROR_UNKNOWN_BACKTRACKING_CONTROL_VERB:
453 *errmsg = _("(*VERB) not recognized");
455 case G_REGEX_ERROR_NUMBER_TOO_BIG:
456 *errmsg = _("number is too bug");
458 case G_REGEX_ERROR_MISSING_SUBPATTERN_NAME:
459 *errmsg = _("missing subpattern name after (?&");
461 case G_REGEX_ERROR_MISSING_DIGIT:
462 *errmsg = _("digit expected after (?+");
464 case G_REGEX_ERROR_EXTRA_SUBPATTERN_NAME:
465 *errmsg = _("different names for subpatterns of the same number are not allowed");
467 case G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_REQUIRED:
468 *errmsg = _("(*MARK) must have an argument");
470 case G_REGEX_ERROR_INVALID_CONTROL_CHAR:
471 *errmsg = _( "\\c must be followed by an ASCII character");
473 case G_REGEX_ERROR_MISSING_NAME:
474 *errmsg = _("\\k is not followed by a braced, angle-bracketed, or quoted name");
476 case G_REGEX_ERROR_NOT_SUPPORTED_IN_CLASS:
477 *errmsg = _("\\N is not supported in a class");
479 case G_REGEX_ERROR_TOO_MANY_FORWARD_REFERENCES:
480 *errmsg = _("too many forward references");
482 case G_REGEX_ERROR_NAME_TOO_LONG:
483 *errmsg = _("name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)");
486 case 116: /* erroffset passed as NULL */
487 /* This should not happen as we never pass a NULL erroffset */
488 g_warning ("erroffset passed as NULL");
489 *errcode = G_REGEX_ERROR_COMPILE;
491 case 117: /* unknown option bit(s) set */
492 /* This should not happen as we check options before passing them
493 * to pcre_compile2() */
494 g_warning ("unknown option bit(s) set");
495 *errcode = G_REGEX_ERROR_COMPILE;
497 case 132: /* this version of PCRE is compiled without UTF support */
498 case 144: /* invalid UTF-8 string */
499 case 145: /* support for \\P, \\p, and \\X has not been compiled */
500 case 167: /* this version of PCRE is not compiled with Unicode property support */
501 case 173: /* disallowed Unicode code point (>= 0xd800 && <= 0xdfff) */
502 case 174: /* invalid UTF-16 string */
503 /* These errors should not happen as we are using an UTF-8 and UCP-enabled PCRE
504 * and we do not check if strings are valid */
505 case 164: /* ] is an invalid data character in JavaScript compatibility mode */
506 /* This should not happen as we don't use PCRE_JAVASCRIPT_COMPAT */
507 g_warning ("%s", *errmsg);
508 *errcode = G_REGEX_ERROR_COMPILE;
510 case 170: /* internal error: unknown opcode in find_fixedlength() */
511 *errcode = G_REGEX_ERROR_INTERNAL;
515 *errcode = G_REGEX_ERROR_COMPILE;
522 match_info_new (const GRegex *regex,
529 GMatchInfo *match_info;
532 string_len = strlen (string);
534 match_info = g_new0 (GMatchInfo, 1);
535 match_info->ref_count = 1;
536 match_info->regex = g_regex_ref ((GRegex *)regex);
537 match_info->string = string;
538 match_info->string_len = string_len;
539 match_info->matches = PCRE_ERROR_NOMATCH;
540 match_info->pos = start_position;
541 match_info->match_opts = match_options;
545 /* These values should be enough for most cases, if they are not
546 * enough g_regex_match_all_full() will expand them. */
547 match_info->n_offsets = 24;
548 match_info->n_workspace = 100;
549 match_info->workspace = g_new (gint, match_info->n_workspace);
554 pcre_fullinfo (regex->pcre_re, regex->extra,
555 PCRE_INFO_CAPTURECOUNT, &capture_count);
556 match_info->n_offsets = (capture_count + 1) * 3;
559 match_info->offsets = g_new0 (gint, match_info->n_offsets);
560 /* Set an invalid position for the previous match. */
561 match_info->offsets[0] = -1;
562 match_info->offsets[1] = -1;
568 * g_match_info_get_regex:
569 * @match_info: a #GMatchInfo
571 * Returns #GRegex object used in @match_info. It belongs to Glib
572 * and must not be freed. Use g_regex_ref() if you need to keep it
573 * after you free @match_info object.
575 * Returns: #GRegex object used in @match_info
580 g_match_info_get_regex (const GMatchInfo *match_info)
582 g_return_val_if_fail (match_info != NULL, NULL);
583 return match_info->regex;
587 * g_match_info_get_string:
588 * @match_info: a #GMatchInfo
590 * Returns the string searched with @match_info. This is the
591 * string passed to g_regex_match() or g_regex_replace() so
592 * you may not free it before calling this function.
594 * Returns: the string searched with @match_info
599 g_match_info_get_string (const GMatchInfo *match_info)
601 g_return_val_if_fail (match_info != NULL, NULL);
602 return match_info->string;
607 * @match_info: a #GMatchInfo
609 * Increases reference count of @match_info by 1.
611 * Returns: @match_info
616 g_match_info_ref (GMatchInfo *match_info)
618 g_return_val_if_fail (match_info != NULL, NULL);
619 g_atomic_int_inc (&match_info->ref_count);
624 * g_match_info_unref:
625 * @match_info: a #GMatchInfo
627 * Decreases reference count of @match_info by 1. When reference count drops
628 * to zero, it frees all the memory associated with the match_info structure.
633 g_match_info_unref (GMatchInfo *match_info)
635 if (g_atomic_int_dec_and_test (&match_info->ref_count))
637 g_regex_unref (match_info->regex);
638 g_free (match_info->offsets);
639 g_free (match_info->workspace);
646 * @match_info: (allow-none): a #GMatchInfo, or %NULL
648 * If @match_info is not %NULL, calls g_match_info_unref(); otherwise does
654 g_match_info_free (GMatchInfo *match_info)
656 if (match_info == NULL)
659 g_match_info_unref (match_info);
664 * @match_info: a #GMatchInfo structure
665 * @error: location to store the error occurring, or %NULL to ignore errors
667 * Scans for the next match using the same parameters of the previous
668 * call to g_regex_match_full() or g_regex_match() that returned
671 * The match is done on the string passed to the match function, so you
672 * cannot free it before calling this function.
674 * Returns: %TRUE is the string matched, %FALSE otherwise
679 g_match_info_next (GMatchInfo *match_info,
682 gint prev_match_start;
685 g_return_val_if_fail (match_info != NULL, FALSE);
686 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
687 g_return_val_if_fail (match_info->pos >= 0, FALSE);
689 prev_match_start = match_info->offsets[0];
690 prev_match_end = match_info->offsets[1];
692 if (match_info->pos > match_info->string_len)
694 /* we have reached the end of the string */
695 match_info->pos = -1;
696 match_info->matches = PCRE_ERROR_NOMATCH;
700 match_info->matches = pcre_exec (match_info->regex->pcre_re,
701 match_info->regex->extra,
703 match_info->string_len,
705 match_info->regex->match_opts | match_info->match_opts,
707 match_info->n_offsets);
708 if (IS_PCRE_ERROR (match_info->matches))
710 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
711 _("Error while matching regular expression %s: %s"),
712 match_info->regex->pattern, match_error (match_info->matches));
716 /* avoid infinite loops if the pattern is an empty string or something
718 if (match_info->pos == match_info->offsets[1])
720 if (match_info->pos > match_info->string_len)
722 /* we have reached the end of the string */
723 match_info->pos = -1;
724 match_info->matches = PCRE_ERROR_NOMATCH;
728 match_info->pos = NEXT_CHAR (match_info->regex,
729 &match_info->string[match_info->pos]) -
734 match_info->pos = match_info->offsets[1];
737 /* it's possible to get two identical matches when we are matching
738 * empty strings, for instance if the pattern is "(?=[A-Z0-9])" and
739 * the string is "RegExTest" we have:
740 * - search at position 0: match from 0 to 0
741 * - search at position 1: match from 3 to 3
742 * - search at position 3: match from 3 to 3 (duplicate)
743 * - search at position 4: match from 5 to 5
744 * - search at position 5: match from 5 to 5 (duplicate)
745 * - search at position 6: no match -> stop
746 * so we have to ignore the duplicates.
747 * see bug #515944: http://bugzilla.gnome.org/show_bug.cgi?id=515944 */
748 if (match_info->matches >= 0 &&
749 prev_match_start == match_info->offsets[0] &&
750 prev_match_end == match_info->offsets[1])
752 /* ignore this match and search the next one */
753 return g_match_info_next (match_info, error);
756 return match_info->matches >= 0;
760 * g_match_info_matches:
761 * @match_info: a #GMatchInfo structure
763 * Returns whether the previous match operation succeeded.
765 * Returns: %TRUE if the previous match operation succeeded,
771 g_match_info_matches (const GMatchInfo *match_info)
773 g_return_val_if_fail (match_info != NULL, FALSE);
775 return match_info->matches >= 0;
779 * g_match_info_get_match_count:
780 * @match_info: a #GMatchInfo structure
782 * Retrieves the number of matched substrings (including substring 0,
783 * that is the whole matched text), so 1 is returned if the pattern
784 * has no substrings in it and 0 is returned if the match failed.
786 * If the last match was obtained using the DFA algorithm, that is
787 * using g_regex_match_all() or g_regex_match_all_full(), the retrieved
788 * count is not that of the number of capturing parentheses but that of
789 * the number of matched substrings.
791 * Returns: Number of matched substrings, or -1 if an error occurred
796 g_match_info_get_match_count (const GMatchInfo *match_info)
798 g_return_val_if_fail (match_info, -1);
800 if (match_info->matches == PCRE_ERROR_NOMATCH)
803 else if (match_info->matches < PCRE_ERROR_NOMATCH)
808 return match_info->matches;
812 * g_match_info_is_partial_match:
813 * @match_info: a #GMatchInfo structure
815 * Usually if the string passed to g_regex_match*() matches as far as
816 * it goes, but is too short to match the entire pattern, %FALSE is
817 * returned. There are circumstances where it might be helpful to
818 * distinguish this case from other cases in which there is no match.
820 * Consider, for example, an application where a human is required to
821 * type in data for a field with specific formatting requirements. An
822 * example might be a date in the form ddmmmyy, defined by the pattern
823 * "^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$".
824 * If the application sees the user’s keystrokes one by one, and can
825 * check that what has been typed so far is potentially valid, it is
826 * able to raise an error as soon as a mistake is made.
828 * GRegex supports the concept of partial matching by means of the
829 * #G_REGEX_MATCH_PARTIAL flag. When this is set the return code for
830 * g_regex_match() or g_regex_match_full() is, as usual, %TRUE
831 * for a complete match, %FALSE otherwise. But, when these functions
832 * return %FALSE, you can check if the match was partial calling
833 * g_match_info_is_partial_match().
835 * When using partial matching you cannot use g_match_info_fetch*().
837 * Because of the way certain internal optimizations are implemented
838 * the partial matching algorithm cannot be used with all patterns.
839 * So repeated single characters such as "a{2,4}" and repeated single
840 * meta-sequences such as "\d+" are not permitted if the maximum number
841 * of occurrences is greater than one. Optional items such as "\d?"
842 * (where the maximum is one) are permitted. Quantifiers with any values
843 * are permitted after parentheses, so the invalid examples above can be
844 * coded thus "(a){2,4}" and "(\d)+". If #G_REGEX_MATCH_PARTIAL is set
845 * for a pattern that does not conform to the restrictions, matching
846 * functions return an error.
848 * Returns: %TRUE if the match was partial, %FALSE otherwise
853 g_match_info_is_partial_match (const GMatchInfo *match_info)
855 g_return_val_if_fail (match_info != NULL, FALSE);
857 return match_info->matches == PCRE_ERROR_PARTIAL;
861 * g_match_info_expand_references:
862 * @match_info: (allow-none): a #GMatchInfo or %NULL
863 * @string_to_expand: the string to expand
864 * @error: location to store the error occurring, or %NULL to ignore errors
866 * Returns a new string containing the text in @string_to_expand with
867 * references and escape sequences expanded. References refer to the last
868 * match done with @string against @regex and have the same syntax used by
871 * The @string_to_expand must be UTF-8 encoded even if #G_REGEX_RAW was
872 * passed to g_regex_new().
874 * The backreferences are extracted from the string passed to the match
875 * function, so you cannot call this function after freeing the string.
877 * @match_info may be %NULL in which case @string_to_expand must not
878 * contain references. For instance "foo\n" does not refer to an actual
879 * pattern and '\n' merely will be replaced with \n character,
880 * while to expand "\0" (whole match) one needs the result of a match.
881 * Use g_regex_check_replacement() to find out whether @string_to_expand
882 * contains references.
884 * Returns: (allow-none): the expanded string, or %NULL if an error occurred
889 g_match_info_expand_references (const GMatchInfo *match_info,
890 const gchar *string_to_expand,
895 GError *tmp_error = NULL;
897 g_return_val_if_fail (string_to_expand != NULL, NULL);
898 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
900 list = split_replacement (string_to_expand, &tmp_error);
901 if (tmp_error != NULL)
903 g_propagate_error (error, tmp_error);
907 if (!match_info && interpolation_list_needs_match (list))
909 g_critical ("String '%s' contains references to the match, can't "
910 "expand references without GMatchInfo object",
915 result = g_string_sized_new (strlen (string_to_expand));
916 interpolate_replacement (match_info, result, list);
918 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
920 return g_string_free (result, FALSE);
924 * g_match_info_fetch:
925 * @match_info: #GMatchInfo structure
926 * @match_num: number of the sub expression
928 * Retrieves the text matching the @match_num<!-- -->'th capturing
929 * parentheses. 0 is the full text of the match, 1 is the first paren
930 * set, 2 the second, and so on.
932 * If @match_num is a valid sub pattern but it didn't match anything
933 * (e.g. sub pattern 1, matching "b" against "(a)?b") then an empty
934 * string is returned.
936 * If the match was obtained using the DFA algorithm, that is using
937 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
938 * string is not that of a set of parentheses but that of a matched
939 * substring. Substrings are matched in reverse order of length, so
940 * 0 is the longest match.
942 * The string is fetched from the string passed to the match function,
943 * so you cannot call this function after freeing the string.
945 * Returns: (allow-none): The matched substring, or %NULL if an error
946 * occurred. You have to free the string yourself
951 g_match_info_fetch (const GMatchInfo *match_info,
954 /* we cannot use pcre_get_substring() because it allocates the
955 * string using pcre_malloc(). */
959 g_return_val_if_fail (match_info != NULL, NULL);
960 g_return_val_if_fail (match_num >= 0, NULL);
962 /* match_num does not exist or it didn't matched, i.e. matching "b"
963 * against "(a)?b" then group 0 is empty. */
964 if (!g_match_info_fetch_pos (match_info, match_num, &start, &end))
966 else if (start == -1)
967 match = g_strdup ("");
969 match = g_strndup (&match_info->string[start], end - start);
975 * g_match_info_fetch_pos:
976 * @match_info: #GMatchInfo structure
977 * @match_num: number of the sub expression
978 * @start_pos: (out) (allow-none): pointer to location where to store
979 * the start position, or %NULL
980 * @end_pos: (out) (allow-none): pointer to location where to store
981 * the end position, or %NULL
983 * Retrieves the position in bytes of the @match_num<!-- -->'th capturing
984 * parentheses. 0 is the full text of the match, 1 is the first
985 * paren set, 2 the second, and so on.
987 * If @match_num is a valid sub pattern but it didn't match anything
988 * (e.g. sub pattern 1, matching "b" against "(a)?b") then @start_pos
989 * and @end_pos are set to -1 and %TRUE is returned.
991 * If the match was obtained using the DFA algorithm, that is using
992 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
993 * position is not that of a set of parentheses but that of a matched
994 * substring. Substrings are matched in reverse order of length, so
995 * 0 is the longest match.
997 * Returns: %TRUE if the position was fetched, %FALSE otherwise. If
998 * the position cannot be fetched, @start_pos and @end_pos are left
1004 g_match_info_fetch_pos (const GMatchInfo *match_info,
1009 g_return_val_if_fail (match_info != NULL, FALSE);
1010 g_return_val_if_fail (match_num >= 0, FALSE);
1012 /* make sure the sub expression number they're requesting is less than
1013 * the total number of sub expressions that were matched. */
1014 if (match_num >= match_info->matches)
1017 if (start_pos != NULL)
1018 *start_pos = match_info->offsets[2 * match_num];
1020 if (end_pos != NULL)
1021 *end_pos = match_info->offsets[2 * match_num + 1];
1027 * Returns number of first matched subpattern with name @name.
1028 * There may be more than one in case when DUPNAMES is used,
1029 * and not all subpatterns with that name match;
1030 * pcre_get_stringnumber() does not work in that case.
1033 get_matched_substring_number (const GMatchInfo *match_info,
1037 gchar *first, *last;
1040 if (!(match_info->regex->compile_opts & G_REGEX_DUPNAMES))
1041 return pcre_get_stringnumber (match_info->regex->pcre_re, name);
1043 /* This code is copied from pcre_get.c: get_first_set() */
1044 entrysize = pcre_get_stringtable_entries (match_info->regex->pcre_re,
1052 for (entry = (guchar*) first; entry <= (guchar*) last; entry += entrysize)
1054 gint n = (entry[0] << 8) + entry[1];
1055 if (match_info->offsets[n*2] >= 0)
1059 return (first[0] << 8) + first[1];
1063 * g_match_info_fetch_named:
1064 * @match_info: #GMatchInfo structure
1065 * @name: name of the subexpression
1067 * Retrieves the text matching the capturing parentheses named @name.
1069 * If @name is a valid sub pattern name but it didn't match anything
1070 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1071 * then an empty string is returned.
1073 * The string is fetched from the string passed to the match function,
1074 * so you cannot call this function after freeing the string.
1076 * Returns: (allow-none): The matched substring, or %NULL if an error
1077 * occurred. You have to free the string yourself
1082 g_match_info_fetch_named (const GMatchInfo *match_info,
1085 /* we cannot use pcre_get_named_substring() because it allocates the
1086 * string using pcre_malloc(). */
1089 g_return_val_if_fail (match_info != NULL, NULL);
1090 g_return_val_if_fail (name != NULL, NULL);
1092 num = get_matched_substring_number (match_info, name);
1096 return g_match_info_fetch (match_info, num);
1100 * g_match_info_fetch_named_pos:
1101 * @match_info: #GMatchInfo structure
1102 * @name: name of the subexpression
1103 * @start_pos: (out) (allow-none): pointer to location where to store
1104 * the start position, or %NULL
1105 * @end_pos: (out) (allow-none): pointer to location where to store
1106 * the end position, or %NULL
1108 * Retrieves the position in bytes of the capturing parentheses named @name.
1110 * If @name is a valid sub pattern name but it didn't match anything
1111 * (e.g. sub pattern "X", matching "b" against "(?P<X>a)?b")
1112 * then @start_pos and @end_pos are set to -1 and %TRUE is returned.
1114 * Returns: %TRUE if the position was fetched, %FALSE otherwise.
1115 * If the position cannot be fetched, @start_pos and @end_pos
1116 * are left unchanged.
1121 g_match_info_fetch_named_pos (const GMatchInfo *match_info,
1128 g_return_val_if_fail (match_info != NULL, FALSE);
1129 g_return_val_if_fail (name != NULL, FALSE);
1131 num = get_matched_substring_number (match_info, name);
1135 return g_match_info_fetch_pos (match_info, num, start_pos, end_pos);
1139 * g_match_info_fetch_all:
1140 * @match_info: a #GMatchInfo structure
1142 * Bundles up pointers to each of the matching substrings from a match
1143 * and stores them in an array of gchar pointers. The first element in
1144 * the returned array is the match number 0, i.e. the entire matched
1147 * If a sub pattern didn't match anything (e.g. sub pattern 1, matching
1148 * "b" against "(a)?b") then an empty string is inserted.
1150 * If the last match was obtained using the DFA algorithm, that is using
1151 * g_regex_match_all() or g_regex_match_all_full(), the retrieved
1152 * strings are not that matched by sets of parentheses but that of the
1153 * matched substring. Substrings are matched in reverse order of length,
1154 * so the first one is the longest match.
1156 * The strings are fetched from the string passed to the match function,
1157 * so you cannot call this function after freeing the string.
1159 * Returns: (allow-none): a %NULL-terminated array of gchar * pointers.
1160 * It must be freed using g_strfreev(). If the previous match failed
1166 g_match_info_fetch_all (const GMatchInfo *match_info)
1168 /* we cannot use pcre_get_substring_list() because the returned value
1169 * isn't suitable for g_strfreev(). */
1173 g_return_val_if_fail (match_info != NULL, NULL);
1175 if (match_info->matches < 0)
1178 result = g_new (gchar *, match_info->matches + 1);
1179 for (i = 0; i < match_info->matches; i++)
1180 result[i] = g_match_info_fetch (match_info, i);
1190 g_regex_error_quark (void)
1192 static GQuark error_quark = 0;
1194 if (error_quark == 0)
1195 error_quark = g_quark_from_static_string ("g-regex-error-quark");
1204 * Increases reference count of @regex by 1.
1211 g_regex_ref (GRegex *regex)
1213 g_return_val_if_fail (regex != NULL, NULL);
1214 g_atomic_int_inc (®ex->ref_count);
1222 * Decreases reference count of @regex by 1. When reference count drops
1223 * to zero, it frees all the memory associated with the regex structure.
1228 g_regex_unref (GRegex *regex)
1230 g_return_if_fail (regex != NULL);
1232 if (g_atomic_int_dec_and_test (®ex->ref_count))
1234 g_free (regex->pattern);
1235 if (regex->pcre_re != NULL)
1236 pcre_free (regex->pcre_re);
1237 if (regex->extra != NULL)
1238 pcre_free (regex->extra);
1245 * @pattern: the regular expression
1246 * @compile_options: compile options for the regular expression, or 0
1247 * @match_options: match options for the regular expression, or 0
1248 * @error: return location for a #GError
1250 * Compiles the regular expression to an internal form, and does
1251 * the initial setup of the #GRegex structure.
1253 * Returns: a #GRegex structure. Call g_regex_unref() when you
1259 g_regex_new (const gchar *pattern,
1260 GRegexCompileFlags compile_options,
1261 GRegexMatchFlags match_options,
1266 const gchar *errmsg;
1269 gboolean optimize = FALSE;
1270 static volatile gsize initialised = 0;
1271 unsigned long int pcre_compile_options;
1273 g_return_val_if_fail (pattern != NULL, NULL);
1274 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
1275 g_return_val_if_fail ((compile_options & ~G_REGEX_COMPILE_MASK) == 0, NULL);
1276 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
1278 if (g_once_init_enter (&initialised))
1280 int supports_utf8, supports_ucp;
1282 pcre_config (PCRE_CONFIG_UTF8, &supports_utf8);
1284 g_critical (_("PCRE library is compiled without UTF8 support"));
1286 pcre_config (PCRE_CONFIG_UNICODE_PROPERTIES, &supports_ucp);
1288 g_critical (_("PCRE library is compiled without UTF8 properties support"));
1290 g_once_init_leave (&initialised, supports_utf8 && supports_ucp ? 1 : 2);
1293 if (G_UNLIKELY (initialised != 1))
1295 g_set_error_literal (error, G_REGEX_ERROR, G_REGEX_ERROR_COMPILE,
1296 _("PCRE library is compiled with incompatible options"));
1300 /* G_REGEX_OPTIMIZE has the same numeric value of PCRE_NO_UTF8_CHECK,
1301 * as we do not need to wrap PCRE_NO_UTF8_CHECK. */
1302 if (compile_options & G_REGEX_OPTIMIZE)
1305 /* In GRegex the string are, by default, UTF-8 encoded. PCRE
1306 * instead uses UTF-8 only if required with PCRE_UTF8. */
1307 if (compile_options & G_REGEX_RAW)
1310 compile_options &= ~G_REGEX_RAW;
1315 compile_options |= PCRE_UTF8 | PCRE_NO_UTF8_CHECK;
1316 match_options |= PCRE_NO_UTF8_CHECK;
1319 /* PCRE_NEWLINE_ANY is the default for the internal PCRE but
1320 * not for the system one. */
1321 if (!(compile_options & G_REGEX_NEWLINE_CR) &&
1322 !(compile_options & G_REGEX_NEWLINE_LF))
1324 compile_options |= PCRE_NEWLINE_ANY;
1327 compile_options |= PCRE_UCP;
1329 /* PCRE_BSR_UNICODE is the default for the internal PCRE but
1330 * possibly not for the system one.
1332 if (~compile_options & G_REGEX_BSR_ANYCRLF)
1333 compile_options |= PCRE_BSR_UNICODE;
1335 /* compile the pattern */
1336 re = pcre_compile2 (pattern, compile_options, &errcode,
1337 &errmsg, &erroffset, NULL);
1339 /* if the compilation failed, set the error member and return
1345 /* Translate the PCRE error code to GRegexError and use a translated
1346 * error message if possible */
1347 translate_compile_error (&errcode, &errmsg);
1349 /* PCRE uses byte offsets but we want to show character offsets */
1350 erroffset = g_utf8_pointer_to_offset (pattern, &pattern[erroffset]);
1352 tmp_error = g_error_new (G_REGEX_ERROR, errcode,
1353 _("Error while compiling regular "
1354 "expression %s at char %d: %s"),
1355 pattern, erroffset, errmsg);
1356 g_propagate_error (error, tmp_error);
1361 /* For options set at the beginning of the pattern, pcre puts them into
1362 * compile options, e.g. "(?i)foo" will make the pcre structure store
1363 * PCRE_CASELESS even though it wasn't explicitly given for compilation. */
1364 pcre_fullinfo (re, NULL, PCRE_INFO_OPTIONS, &pcre_compile_options);
1365 compile_options = pcre_compile_options;
1367 if (!(compile_options & G_REGEX_DUPNAMES))
1369 gboolean jchanged = FALSE;
1370 pcre_fullinfo (re, NULL, PCRE_INFO_JCHANGED, &jchanged);
1372 compile_options |= G_REGEX_DUPNAMES;
1375 regex = g_new0 (GRegex, 1);
1376 regex->ref_count = 1;
1377 regex->pattern = g_strdup (pattern);
1378 regex->pcre_re = re;
1379 regex->compile_opts = compile_options;
1380 regex->match_opts = match_options;
1384 regex->extra = pcre_study (regex->pcre_re, 0, &errmsg);
1387 GError *tmp_error = g_error_new (G_REGEX_ERROR,
1388 G_REGEX_ERROR_OPTIMIZE,
1389 _("Error while optimizing "
1390 "regular expression %s: %s"),
1393 g_propagate_error (error, tmp_error);
1395 g_regex_unref (regex);
1404 * g_regex_get_pattern:
1405 * @regex: a #GRegex structure
1407 * Gets the pattern string associated with @regex, i.e. a copy of
1408 * the string passed to g_regex_new().
1410 * Returns: the pattern of @regex
1415 g_regex_get_pattern (const GRegex *regex)
1417 g_return_val_if_fail (regex != NULL, NULL);
1419 return regex->pattern;
1423 * g_regex_get_max_backref:
1426 * Returns the number of the highest back reference
1427 * in the pattern, or 0 if the pattern does not contain
1430 * Returns: the number of the highest back reference
1435 g_regex_get_max_backref (const GRegex *regex)
1439 pcre_fullinfo (regex->pcre_re, regex->extra,
1440 PCRE_INFO_BACKREFMAX, &value);
1446 * g_regex_get_capture_count:
1449 * Returns the number of capturing subpatterns in the pattern.
1451 * Returns: the number of capturing subpatterns
1456 g_regex_get_capture_count (const GRegex *regex)
1460 pcre_fullinfo (regex->pcre_re, regex->extra,
1461 PCRE_INFO_CAPTURECOUNT, &value);
1467 * g_regex_get_has_cr_or_lf:
1468 * @regex: a #GRegex structure
1470 * Checks whether the pattern contains explicit CR or LF references.
1472 * Returns: %TRUE if the pattern contains explicit CR or LF references
1477 g_regex_get_has_cr_or_lf (const GRegex *regex)
1481 pcre_fullinfo (regex->pcre_re, regex->extra,
1482 PCRE_INFO_HASCRORLF, &value);
1488 * g_regex_get_compile_flags:
1491 * Returns the compile options that @regex was created with.
1493 * Returns: flags from #GRegexCompileFlags
1498 g_regex_get_compile_flags (const GRegex *regex)
1500 g_return_val_if_fail (regex != NULL, 0);
1502 return regex->compile_opts;
1506 * g_regex_get_match_flags:
1509 * Returns the match options that @regex was created with.
1511 * Returns: flags from #GRegexMatchFlags
1516 g_regex_get_match_flags (const GRegex *regex)
1518 g_return_val_if_fail (regex != NULL, 0);
1520 return regex->match_opts;
1524 * g_regex_match_simple:
1525 * @pattern: the regular expression
1526 * @string: the string to scan for matches
1527 * @compile_options: compile options for the regular expression, or 0
1528 * @match_options: match options, or 0
1530 * Scans for a match in @string for @pattern.
1532 * This function is equivalent to g_regex_match() but it does not
1533 * require to compile the pattern with g_regex_new(), avoiding some
1534 * lines of code when you need just to do a match without extracting
1535 * substrings, capture counts, and so on.
1537 * If this function is to be called on the same @pattern more than
1538 * once, it's more efficient to compile the pattern once with
1539 * g_regex_new() and then use g_regex_match().
1541 * Returns: %TRUE if the string matched, %FALSE otherwise
1546 g_regex_match_simple (const gchar *pattern,
1547 const gchar *string,
1548 GRegexCompileFlags compile_options,
1549 GRegexMatchFlags match_options)
1554 regex = g_regex_new (pattern, compile_options, 0, NULL);
1557 result = g_regex_match_full (regex, string, -1, 0, match_options, NULL, NULL);
1558 g_regex_unref (regex);
1564 * @regex: a #GRegex structure from g_regex_new()
1565 * @string: the string to scan for matches
1566 * @match_options: match options
1567 * @match_info: (out) (allow-none): pointer to location where to store
1568 * the #GMatchInfo, or %NULL if you do not need it
1570 * Scans for a match in string for the pattern in @regex.
1571 * The @match_options are combined with the match options specified
1572 * when the @regex structure was created, letting you have more
1573 * flexibility in reusing #GRegex structures.
1575 * A #GMatchInfo structure, used to get information on the match,
1576 * is stored in @match_info if not %NULL. Note that if @match_info
1577 * is not %NULL then it is created even if the function returns %FALSE,
1578 * i.e. you must free it regardless if regular expression actually matched.
1580 * To retrieve all the non-overlapping matches of the pattern in
1581 * string you can use g_match_info_next().
1585 * print_uppercase_words (const gchar *string)
1587 * /* Print all uppercase-only words. */
1589 * GMatchInfo *match_info;
1591 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1592 * g_regex_match (regex, string, 0, &match_info);
1593 * while (g_match_info_matches (match_info))
1595 * gchar *word = g_match_info_fetch (match_info, 0);
1596 * g_print ("Found: %s\n", word);
1598 * g_match_info_next (match_info, NULL);
1600 * g_match_info_free (match_info);
1601 * g_regex_unref (regex);
1605 * @string is not copied and is used in #GMatchInfo internally. If
1606 * you use any #GMatchInfo method (except g_match_info_free()) after
1607 * freeing or modifying @string then the behaviour is undefined.
1609 * Returns: %TRUE is the string matched, %FALSE otherwise
1614 g_regex_match (const GRegex *regex,
1615 const gchar *string,
1616 GRegexMatchFlags match_options,
1617 GMatchInfo **match_info)
1619 return g_regex_match_full (regex, string, -1, 0, match_options,
1624 * g_regex_match_full:
1625 * @regex: a #GRegex structure from g_regex_new()
1626 * @string: (array length=string_len): the string to scan for matches
1627 * @string_len: the length of @string, or -1 if @string is nul-terminated
1628 * @start_position: starting index of the string to match
1629 * @match_options: match options
1630 * @match_info: (out) (allow-none): pointer to location where to store
1631 * the #GMatchInfo, or %NULL if you do not need it
1632 * @error: location to store the error occurring, or %NULL to ignore errors
1634 * Scans for a match in string for the pattern in @regex.
1635 * The @match_options are combined with the match options specified
1636 * when the @regex structure was created, letting you have more
1637 * flexibility in reusing #GRegex structures.
1639 * Setting @start_position differs from just passing over a shortened
1640 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1641 * that begins with any kind of lookbehind assertion, such as "\b".
1643 * A #GMatchInfo structure, used to get information on the match, is
1644 * stored in @match_info if not %NULL. Note that if @match_info is
1645 * not %NULL then it is created even if the function returns %FALSE,
1646 * i.e. you must free it regardless if regular expression actually
1649 * @string is not copied and is used in #GMatchInfo internally. If
1650 * you use any #GMatchInfo method (except g_match_info_free()) after
1651 * freeing or modifying @string then the behaviour is undefined.
1653 * To retrieve all the non-overlapping matches of the pattern in
1654 * string you can use g_match_info_next().
1658 * print_uppercase_words (const gchar *string)
1660 * /* Print all uppercase-only words. */
1662 * GMatchInfo *match_info;
1663 * GError *error = NULL;
1665 * regex = g_regex_new ("[A-Z]+", 0, 0, NULL);
1666 * g_regex_match_full (regex, string, -1, 0, 0, &match_info, &error);
1667 * while (g_match_info_matches (match_info))
1669 * gchar *word = g_match_info_fetch (match_info, 0);
1670 * g_print ("Found: %s\n", word);
1672 * g_match_info_next (match_info, &error);
1674 * g_match_info_free (match_info);
1675 * g_regex_unref (regex);
1676 * if (error != NULL)
1678 * g_printerr ("Error while matching: %s\n", error->message);
1679 * g_error_free (error);
1684 * Returns: %TRUE is the string matched, %FALSE otherwise
1689 g_regex_match_full (const GRegex *regex,
1690 const gchar *string,
1692 gint start_position,
1693 GRegexMatchFlags match_options,
1694 GMatchInfo **match_info,
1700 g_return_val_if_fail (regex != NULL, FALSE);
1701 g_return_val_if_fail (string != NULL, FALSE);
1702 g_return_val_if_fail (start_position >= 0, FALSE);
1703 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1704 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1706 info = match_info_new (regex, string, string_len, start_position,
1707 match_options, FALSE);
1708 match_ok = g_match_info_next (info, error);
1709 if (match_info != NULL)
1712 g_match_info_free (info);
1718 * g_regex_match_all:
1719 * @regex: a #GRegex structure from g_regex_new()
1720 * @string: the string to scan for matches
1721 * @match_options: match options
1722 * @match_info: (out) (allow-none): pointer to location where to store
1723 * the #GMatchInfo, or %NULL if you do not need it
1725 * Using the standard algorithm for regular expression matching only
1726 * the longest match in the string is retrieved. This function uses
1727 * a different algorithm so it can retrieve all the possible matches.
1728 * For more documentation see g_regex_match_all_full().
1730 * A #GMatchInfo structure, used to get information on the match, is
1731 * stored in @match_info if not %NULL. Note that if @match_info is
1732 * not %NULL then it is created even if the function returns %FALSE,
1733 * i.e. you must free it regardless if regular expression actually
1736 * @string is not copied and is used in #GMatchInfo internally. If
1737 * you use any #GMatchInfo method (except g_match_info_free()) after
1738 * freeing or modifying @string then the behaviour is undefined.
1740 * Returns: %TRUE is the string matched, %FALSE otherwise
1745 g_regex_match_all (const GRegex *regex,
1746 const gchar *string,
1747 GRegexMatchFlags match_options,
1748 GMatchInfo **match_info)
1750 return g_regex_match_all_full (regex, string, -1, 0, match_options,
1755 * g_regex_match_all_full:
1756 * @regex: a #GRegex structure from g_regex_new()
1757 * @string: (array length=string_len): the string to scan for matches
1758 * @string_len: the length of @string, or -1 if @string is nul-terminated
1759 * @start_position: starting index of the string to match
1760 * @match_options: match options
1761 * @match_info: (out) (allow-none): pointer to location where to store
1762 * the #GMatchInfo, or %NULL if you do not need it
1763 * @error: location to store the error occurring, or %NULL to ignore errors
1765 * Using the standard algorithm for regular expression matching only
1766 * the longest match in the string is retrieved, it is not possible
1767 * to obtain all the available matches. For instance matching
1768 * "<a> <b> <c>" against the pattern "<.*>"
1769 * you get "<a> <b> <c>".
1771 * This function uses a different algorithm (called DFA, i.e. deterministic
1772 * finite automaton), so it can retrieve all the possible matches, all
1773 * starting at the same point in the string. For instance matching
1774 * "<a> <b> <c>" against the pattern "<.*>"
1775 * you would obtain three matches: "<a> <b> <c>",
1776 * "<a> <b>" and "<a>".
1778 * The number of matched strings is retrieved using
1779 * g_match_info_get_match_count(). To obtain the matched strings and
1780 * their position you can use, respectively, g_match_info_fetch() and
1781 * g_match_info_fetch_pos(). Note that the strings are returned in
1782 * reverse order of length; that is, the longest matching string is
1785 * Note that the DFA algorithm is slower than the standard one and it
1786 * is not able to capture substrings, so backreferences do not work.
1788 * Setting @start_position differs from just passing over a shortened
1789 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
1790 * that begins with any kind of lookbehind assertion, such as "\b".
1792 * A #GMatchInfo structure, used to get information on the match, is
1793 * stored in @match_info if not %NULL. Note that if @match_info is
1794 * not %NULL then it is created even if the function returns %FALSE,
1795 * i.e. you must free it regardless if regular expression actually
1798 * @string is not copied and is used in #GMatchInfo internally. If
1799 * you use any #GMatchInfo method (except g_match_info_free()) after
1800 * freeing or modifying @string then the behaviour is undefined.
1802 * Returns: %TRUE is the string matched, %FALSE otherwise
1807 g_regex_match_all_full (const GRegex *regex,
1808 const gchar *string,
1810 gint start_position,
1811 GRegexMatchFlags match_options,
1812 GMatchInfo **match_info,
1818 g_return_val_if_fail (regex != NULL, FALSE);
1819 g_return_val_if_fail (string != NULL, FALSE);
1820 g_return_val_if_fail (start_position >= 0, FALSE);
1821 g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
1822 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, FALSE);
1824 info = match_info_new (regex, string, string_len, start_position,
1825 match_options, TRUE);
1831 info->matches = pcre_dfa_exec (regex->pcre_re, regex->extra,
1832 info->string, info->string_len,
1834 regex->match_opts | match_options,
1835 info->offsets, info->n_offsets,
1836 info->workspace, info->n_workspace);
1837 if (info->matches == PCRE_ERROR_DFA_WSSIZE)
1839 /* info->workspace is too small. */
1840 info->n_workspace *= 2;
1841 info->workspace = g_realloc (info->workspace,
1842 info->n_workspace * sizeof (gint));
1845 else if (info->matches == 0)
1847 /* info->offsets is too small. */
1848 info->n_offsets *= 2;
1849 info->offsets = g_realloc (info->offsets,
1850 info->n_offsets * sizeof (gint));
1853 else if (IS_PCRE_ERROR (info->matches))
1855 g_set_error (error, G_REGEX_ERROR, G_REGEX_ERROR_MATCH,
1856 _("Error while matching regular expression %s: %s"),
1857 regex->pattern, match_error (info->matches));
1861 /* set info->pos to -1 so that a call to g_match_info_next() fails. */
1864 if (match_info != NULL)
1867 g_match_info_free (info);
1869 return info->matches >= 0;
1873 * g_regex_get_string_number:
1874 * @regex: #GRegex structure
1875 * @name: name of the subexpression
1877 * Retrieves the number of the subexpression named @name.
1879 * Returns: The number of the subexpression or -1 if @name
1885 g_regex_get_string_number (const GRegex *regex,
1890 g_return_val_if_fail (regex != NULL, -1);
1891 g_return_val_if_fail (name != NULL, -1);
1893 num = pcre_get_stringnumber (regex->pcre_re, name);
1894 if (num == PCRE_ERROR_NOSUBSTRING)
1901 * g_regex_split_simple:
1902 * @pattern: the regular expression
1903 * @string: the string to scan for matches
1904 * @compile_options: compile options for the regular expression, or 0
1905 * @match_options: match options, or 0
1907 * Breaks the string on the pattern, and returns an array of
1908 * the tokens. If the pattern contains capturing parentheses,
1909 * then the text for each of the substrings will also be returned.
1910 * If the pattern does not match anywhere in the string, then the
1911 * whole string is returned as the first token.
1913 * This function is equivalent to g_regex_split() but it does
1914 * not require to compile the pattern with g_regex_new(), avoiding
1915 * some lines of code when you need just to do a split without
1916 * extracting substrings, capture counts, and so on.
1918 * If this function is to be called on the same @pattern more than
1919 * once, it's more efficient to compile the pattern once with
1920 * g_regex_new() and then use g_regex_split().
1922 * As a special case, the result of splitting the empty string ""
1923 * is an empty vector, not a vector containing a single string.
1924 * The reason for this special case is that being able to represent
1925 * a empty vector is typically more useful than consistent handling
1926 * of empty elements. If you do need to represent empty elements,
1927 * you'll need to check for the empty string before calling this
1930 * A pattern that can match empty strings splits @string into
1931 * separate characters wherever it matches the empty string between
1932 * characters. For example splitting "ab c" using as a separator
1933 * "\s*", you will get "a", "b" and "c".
1935 * Returns: a %NULL-terminated array of strings. Free it using g_strfreev()
1940 g_regex_split_simple (const gchar *pattern,
1941 const gchar *string,
1942 GRegexCompileFlags compile_options,
1943 GRegexMatchFlags match_options)
1948 regex = g_regex_new (pattern, compile_options, 0, NULL);
1952 result = g_regex_split_full (regex, string, -1, 0, match_options, 0, NULL);
1953 g_regex_unref (regex);
1959 * @regex: a #GRegex structure
1960 * @string: the string to split with the pattern
1961 * @match_options: match time option flags
1963 * Breaks the string on the pattern, and returns an array of the tokens.
1964 * If the pattern contains capturing parentheses, then the text for each
1965 * of the substrings will also be returned. If the pattern does not match
1966 * anywhere in the string, then the whole string is returned as the first
1969 * As a special case, the result of splitting the empty string "" is an
1970 * empty vector, not a vector containing a single string. The reason for
1971 * this special case is that being able to represent a empty vector is
1972 * typically more useful than consistent handling of empty elements. If
1973 * you do need to represent empty elements, you'll need to check for the
1974 * empty string before calling this function.
1976 * A pattern that can match empty strings splits @string into separate
1977 * characters wherever it matches the empty string between characters.
1978 * For example splitting "ab c" using as a separator "\s*", you will get
1981 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
1986 g_regex_split (const GRegex *regex,
1987 const gchar *string,
1988 GRegexMatchFlags match_options)
1990 return g_regex_split_full (regex, string, -1, 0,
1991 match_options, 0, NULL);
1995 * g_regex_split_full:
1996 * @regex: a #GRegex structure
1997 * @string: (array length=string_len): the string to split with the pattern
1998 * @string_len: the length of @string, or -1 if @string is nul-terminated
1999 * @start_position: starting index of the string to match
2000 * @match_options: match time option flags
2001 * @max_tokens: the maximum number of tokens to split @string into.
2002 * If this is less than 1, the string is split completely
2003 * @error: return location for a #GError
2005 * Breaks the string on the pattern, and returns an array of the tokens.
2006 * If the pattern contains capturing parentheses, then the text for each
2007 * of the substrings will also be returned. If the pattern does not match
2008 * anywhere in the string, then the whole string is returned as the first
2011 * As a special case, the result of splitting the empty string "" is an
2012 * empty vector, not a vector containing a single string. The reason for
2013 * this special case is that being able to represent a empty vector is
2014 * typically more useful than consistent handling of empty elements. If
2015 * you do need to represent empty elements, you'll need to check for the
2016 * empty string before calling this function.
2018 * A pattern that can match empty strings splits @string into separate
2019 * characters wherever it matches the empty string between characters.
2020 * For example splitting "ab c" using as a separator "\s*", you will get
2023 * Setting @start_position differs from just passing over a shortened
2024 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2025 * that begins with any kind of lookbehind assertion, such as "\b".
2027 * Returns: a %NULL-terminated gchar ** array. Free it using g_strfreev()
2032 g_regex_split_full (const GRegex *regex,
2033 const gchar *string,
2035 gint start_position,
2036 GRegexMatchFlags match_options,
2040 GError *tmp_error = NULL;
2041 GMatchInfo *match_info;
2046 /* position of the last separator. */
2047 gint last_separator_end;
2048 /* was the last match 0 bytes long? */
2049 gboolean last_match_is_empty;
2050 /* the returned array of char **s */
2051 gchar **string_list;
2053 g_return_val_if_fail (regex != NULL, NULL);
2054 g_return_val_if_fail (string != NULL, NULL);
2055 g_return_val_if_fail (start_position >= 0, NULL);
2056 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2057 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2059 if (max_tokens <= 0)
2060 max_tokens = G_MAXINT;
2063 string_len = strlen (string);
2065 /* zero-length string */
2066 if (string_len - start_position == 0)
2067 return g_new0 (gchar *, 1);
2069 if (max_tokens == 1)
2071 string_list = g_new0 (gchar *, 2);
2072 string_list[0] = g_strndup (&string[start_position],
2073 string_len - start_position);
2079 last_separator_end = start_position;
2080 last_match_is_empty = FALSE;
2082 match_ok = g_regex_match_full (regex, string, string_len, start_position,
2083 match_options, &match_info, &tmp_error);
2085 while (tmp_error == NULL)
2089 last_match_is_empty =
2090 (match_info->offsets[0] == match_info->offsets[1]);
2092 /* we need to skip empty separators at the same position of the end
2093 * of another separator. e.g. the string is "a b" and the separator
2094 * is " *", so from 1 to 2 we have a match and at position 2 we have
2095 * an empty match. */
2096 if (last_separator_end != match_info->offsets[1])
2101 token = g_strndup (string + last_separator_end,
2102 match_info->offsets[0] - last_separator_end);
2103 list = g_list_prepend (list, token);
2106 /* if there were substrings, these need to be added to
2108 match_count = g_match_info_get_match_count (match_info);
2109 if (match_count > 1)
2111 for (i = 1; i < match_count; i++)
2112 list = g_list_prepend (list, g_match_info_fetch (match_info, i));
2118 /* if there was no match, copy to end of string. */
2119 if (!last_match_is_empty)
2121 gchar *token = g_strndup (string + last_separator_end,
2122 match_info->string_len - last_separator_end);
2123 list = g_list_prepend (list, token);
2125 /* no more tokens, end the loop. */
2129 /* -1 to leave room for the last part. */
2130 if (token_count >= max_tokens - 1)
2132 /* we have reached the maximum number of tokens, so we copy
2133 * the remaining part of the string. */
2134 if (last_match_is_empty)
2136 /* the last match was empty, so we have moved one char
2137 * after the real position to avoid empty matches at the
2139 match_info->pos = PREV_CHAR (regex, &string[match_info->pos]) - string;
2141 /* the if is needed in the case we have terminated the available
2142 * tokens, but we are at the end of the string, so there are no
2143 * characters left to copy. */
2144 if (string_len > match_info->pos)
2146 gchar *token = g_strndup (string + match_info->pos,
2147 string_len - match_info->pos);
2148 list = g_list_prepend (list, token);
2154 last_separator_end = match_info->pos;
2155 if (last_match_is_empty)
2156 /* if the last match was empty, g_match_info_next() has moved
2157 * forward to avoid infinite loops, but we still need to copy that
2159 last_separator_end = PREV_CHAR (regex, &string[last_separator_end]) - string;
2161 match_ok = g_match_info_next (match_info, &tmp_error);
2163 g_match_info_free (match_info);
2164 if (tmp_error != NULL)
2166 g_propagate_error (error, tmp_error);
2167 g_list_free_full (list, g_free);
2168 match_info->pos = -1;
2172 string_list = g_new (gchar *, g_list_length (list) + 1);
2174 for (last = g_list_last (list); last; last = g_list_previous (last))
2175 string_list[i++] = last->data;
2176 string_list[i] = NULL;
2185 REPL_TYPE_CHARACTER,
2186 REPL_TYPE_SYMBOLIC_REFERENCE,
2187 REPL_TYPE_NUMERIC_REFERENCE,
2188 REPL_TYPE_CHANGE_CASE
2193 CHANGE_CASE_NONE = 1 << 0,
2194 CHANGE_CASE_UPPER = 1 << 1,
2195 CHANGE_CASE_LOWER = 1 << 2,
2196 CHANGE_CASE_UPPER_SINGLE = 1 << 3,
2197 CHANGE_CASE_LOWER_SINGLE = 1 << 4,
2198 CHANGE_CASE_SINGLE_MASK = CHANGE_CASE_UPPER_SINGLE | CHANGE_CASE_LOWER_SINGLE,
2199 CHANGE_CASE_LOWER_MASK = CHANGE_CASE_LOWER | CHANGE_CASE_LOWER_SINGLE,
2200 CHANGE_CASE_UPPER_MASK = CHANGE_CASE_UPPER | CHANGE_CASE_UPPER_SINGLE
2203 struct _InterpolationData
2209 ChangeCase change_case;
2213 free_interpolation_data (InterpolationData *data)
2215 g_free (data->text);
2219 static const gchar *
2220 expand_escape (const gchar *replacement,
2222 InterpolationData *data,
2227 const gchar *error_detail;
2229 GError *tmp_error = NULL;
2237 data->type = REPL_TYPE_CHARACTER;
2242 data->type = REPL_TYPE_CHARACTER;
2247 data->type = REPL_TYPE_CHARACTER;
2252 data->type = REPL_TYPE_CHARACTER;
2257 data->type = REPL_TYPE_CHARACTER;
2262 data->type = REPL_TYPE_CHARACTER;
2267 data->type = REPL_TYPE_CHARACTER;
2272 data->type = REPL_TYPE_CHARACTER;
2282 h = g_ascii_xdigit_value (*p);
2285 error_detail = _("hexadecimal digit or '}' expected");
2296 for (i = 0; i < 2; i++)
2298 h = g_ascii_xdigit_value (*p);
2301 error_detail = _("hexadecimal digit expected");
2308 data->type = REPL_TYPE_STRING;
2309 data->text = g_new0 (gchar, 8);
2310 g_unichar_to_utf8 (x, data->text);
2314 data->type = REPL_TYPE_CHANGE_CASE;
2315 data->change_case = CHANGE_CASE_LOWER_SINGLE;
2319 data->type = REPL_TYPE_CHANGE_CASE;
2320 data->change_case = CHANGE_CASE_UPPER_SINGLE;
2324 data->type = REPL_TYPE_CHANGE_CASE;
2325 data->change_case = CHANGE_CASE_LOWER;
2329 data->type = REPL_TYPE_CHANGE_CASE;
2330 data->change_case = CHANGE_CASE_UPPER;
2334 data->type = REPL_TYPE_CHANGE_CASE;
2335 data->change_case = CHANGE_CASE_NONE;
2341 error_detail = _("missing '<' in symbolic reference");
2350 error_detail = _("unfinished symbolic reference");
2357 error_detail = _("zero-length symbolic reference");
2360 if (g_ascii_isdigit (*q))
2365 h = g_ascii_digit_value (*q);
2368 error_detail = _("digit expected");
2377 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2384 if (!g_ascii_isalnum (*r))
2386 error_detail = _("illegal symbolic reference");
2393 data->text = g_strndup (q, p - q);
2394 data->type = REPL_TYPE_SYMBOLIC_REFERENCE;
2399 /* if \0 is followed by a number is an octal number representing a
2400 * character, else it is a numeric reference. */
2401 if (g_ascii_digit_value (*g_utf8_next_char (p)) >= 0)
2404 p = g_utf8_next_char (p);
2417 for (i = 0; i < 3; i++)
2419 h = g_ascii_digit_value (*p);
2429 if (i == 2 && base == 10)
2435 if (base == 8 || i == 3)
2437 data->type = REPL_TYPE_STRING;
2438 data->text = g_new0 (gchar, 8);
2439 g_unichar_to_utf8 (x, data->text);
2443 data->type = REPL_TYPE_NUMERIC_REFERENCE;
2448 error_detail = _("stray final '\\'");
2452 error_detail = _("unknown escape sequence");
2459 /* G_GSSIZE_FORMAT doesn't work with gettext, so we use %lu */
2460 tmp_error = g_error_new (G_REGEX_ERROR,
2461 G_REGEX_ERROR_REPLACE,
2462 _("Error while parsing replacement "
2463 "text \"%s\" at char %lu: %s"),
2465 (gulong)(p - replacement),
2467 g_propagate_error (error, tmp_error);
2473 split_replacement (const gchar *replacement,
2477 InterpolationData *data;
2478 const gchar *p, *start;
2480 start = p = replacement;
2485 data = g_new0 (InterpolationData, 1);
2486 start = p = expand_escape (replacement, p, data, error);
2489 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2490 free_interpolation_data (data);
2494 list = g_list_prepend (list, data);
2499 if (*p == '\\' || *p == '\0')
2503 data = g_new0 (InterpolationData, 1);
2504 data->text = g_strndup (start, p - start);
2505 data->type = REPL_TYPE_STRING;
2506 list = g_list_prepend (list, data);
2512 return g_list_reverse (list);
2515 /* Change the case of c based on change_case. */
2516 #define CHANGE_CASE(c, change_case) \
2517 (((change_case) & CHANGE_CASE_LOWER_MASK) ? \
2518 g_unichar_tolower (c) : \
2519 g_unichar_toupper (c))
2522 string_append (GString *string,
2524 ChangeCase *change_case)
2528 if (text[0] == '\0')
2531 if (*change_case == CHANGE_CASE_NONE)
2533 g_string_append (string, text);
2535 else if (*change_case & CHANGE_CASE_SINGLE_MASK)
2537 c = g_utf8_get_char (text);
2538 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2539 g_string_append (string, g_utf8_next_char (text));
2540 *change_case = CHANGE_CASE_NONE;
2544 while (*text != '\0')
2546 c = g_utf8_get_char (text);
2547 g_string_append_unichar (string, CHANGE_CASE (c, *change_case));
2548 text = g_utf8_next_char (text);
2554 interpolate_replacement (const GMatchInfo *match_info,
2559 InterpolationData *idata;
2561 ChangeCase change_case = CHANGE_CASE_NONE;
2563 for (list = data; list; list = list->next)
2566 switch (idata->type)
2568 case REPL_TYPE_STRING:
2569 string_append (result, idata->text, &change_case);
2571 case REPL_TYPE_CHARACTER:
2572 g_string_append_c (result, CHANGE_CASE (idata->c, change_case));
2573 if (change_case & CHANGE_CASE_SINGLE_MASK)
2574 change_case = CHANGE_CASE_NONE;
2576 case REPL_TYPE_NUMERIC_REFERENCE:
2577 match = g_match_info_fetch (match_info, idata->num);
2580 string_append (result, match, &change_case);
2584 case REPL_TYPE_SYMBOLIC_REFERENCE:
2585 match = g_match_info_fetch_named (match_info, idata->text);
2588 string_append (result, match, &change_case);
2592 case REPL_TYPE_CHANGE_CASE:
2593 change_case = idata->change_case;
2601 /* whether actual match_info is needed for replacement, i.e.
2602 * whether there are references
2605 interpolation_list_needs_match (GList *list)
2607 while (list != NULL)
2609 InterpolationData *data = list->data;
2611 if (data->type == REPL_TYPE_SYMBOLIC_REFERENCE ||
2612 data->type == REPL_TYPE_NUMERIC_REFERENCE)
2625 * @regex: a #GRegex structure
2626 * @string: (array length=string_len): the string to perform matches against
2627 * @string_len: the length of @string, or -1 if @string is nul-terminated
2628 * @start_position: starting index of the string to match
2629 * @replacement: text to replace each match with
2630 * @match_options: options for the match
2631 * @error: location to store the error occurring, or %NULL to ignore errors
2633 * Replaces all occurrences of the pattern in @regex with the
2634 * replacement text. Backreferences of the form '\number' or
2635 * '\g<number>' in the replacement text are interpolated by the
2636 * number-th captured subexpression of the match, '\g<name>' refers
2637 * to the captured subexpression with the given name. '\0' refers to the
2638 * complete match, but '\0' followed by a number is the octal representation
2639 * of a character. To include a literal '\' in the replacement, write '\\'.
2640 * There are also escapes that changes the case of the following text:
2643 * <varlistentry><term>\l</term>
2645 * <para>Convert to lower case the next character</para>
2648 * <varlistentry><term>\u</term>
2650 * <para>Convert to upper case the next character</para>
2653 * <varlistentry><term>\L</term>
2655 * <para>Convert to lower case till \E</para>
2658 * <varlistentry><term>\U</term>
2660 * <para>Convert to upper case till \E</para>
2663 * <varlistentry><term>\E</term>
2665 * <para>End case modification</para>
2670 * If you do not need to use backreferences use g_regex_replace_literal().
2672 * The @replacement string must be UTF-8 encoded even if #G_REGEX_RAW was
2673 * passed to g_regex_new(). If you want to use not UTF-8 encoded stings
2674 * you can use g_regex_replace_literal().
2676 * Setting @start_position differs from just passing over a shortened
2677 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern that
2678 * begins with any kind of lookbehind assertion, such as "\b".
2680 * Returns: a newly allocated string containing the replacements
2685 g_regex_replace (const GRegex *regex,
2686 const gchar *string,
2688 gint start_position,
2689 const gchar *replacement,
2690 GRegexMatchFlags match_options,
2695 GError *tmp_error = NULL;
2697 g_return_val_if_fail (regex != NULL, NULL);
2698 g_return_val_if_fail (string != NULL, NULL);
2699 g_return_val_if_fail (start_position >= 0, NULL);
2700 g_return_val_if_fail (replacement != NULL, NULL);
2701 g_return_val_if_fail (error == NULL || *error == NULL, NULL);
2702 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2704 list = split_replacement (replacement, &tmp_error);
2705 if (tmp_error != NULL)
2707 g_propagate_error (error, tmp_error);
2711 result = g_regex_replace_eval (regex,
2712 string, string_len, start_position,
2714 interpolate_replacement,
2717 if (tmp_error != NULL)
2718 g_propagate_error (error, tmp_error);
2720 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2726 literal_replacement (const GMatchInfo *match_info,
2730 g_string_append (result, data);
2735 * g_regex_replace_literal:
2736 * @regex: a #GRegex structure
2737 * @string: (array length=string_len): the string to perform matches against
2738 * @string_len: the length of @string, or -1 if @string is nul-terminated
2739 * @start_position: starting index of the string to match
2740 * @replacement: text to replace each match with
2741 * @match_options: options for the match
2742 * @error: location to store the error occurring, or %NULL to ignore errors
2744 * Replaces all occurrences of the pattern in @regex with the
2745 * replacement text. @replacement is replaced literally, to
2746 * include backreferences use g_regex_replace().
2748 * Setting @start_position differs from just passing over a
2749 * shortened string and setting #G_REGEX_MATCH_NOTBOL in the
2750 * case of a pattern that begins with any kind of lookbehind
2751 * assertion, such as "\b".
2753 * Returns: a newly allocated string containing the replacements
2758 g_regex_replace_literal (const GRegex *regex,
2759 const gchar *string,
2761 gint start_position,
2762 const gchar *replacement,
2763 GRegexMatchFlags match_options,
2766 g_return_val_if_fail (replacement != NULL, NULL);
2767 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2769 return g_regex_replace_eval (regex,
2770 string, string_len, start_position,
2772 literal_replacement,
2773 (gpointer)replacement,
2778 * g_regex_replace_eval:
2779 * @regex: a #GRegex structure from g_regex_new()
2780 * @string: (array length=string_len): string to perform matches against
2781 * @string_len: the length of @string, or -1 if @string is nul-terminated
2782 * @start_position: starting index of the string to match
2783 * @match_options: options for the match
2784 * @eval: a function to call for each match
2785 * @user_data: user data to pass to the function
2786 * @error: location to store the error occurring, or %NULL to ignore errors
2788 * Replaces occurrences of the pattern in regex with the output of
2789 * @eval for that occurrence.
2791 * Setting @start_position differs from just passing over a shortened
2792 * string and setting #G_REGEX_MATCH_NOTBOL in the case of a pattern
2793 * that begins with any kind of lookbehind assertion, such as "\b".
2795 * The following example uses g_regex_replace_eval() to replace multiple
2799 * eval_cb (const GMatchInfo *info,
2806 * match = g_match_info_fetch (info, 0);
2807 * r = g_hash_table_lookup ((GHashTable *)data, match);
2808 * g_string_append (res, r);
2820 * h = g_hash_table_new (g_str_hash, g_str_equal);
2822 * g_hash_table_insert (h, "1", "ONE");
2823 * g_hash_table_insert (h, "2", "TWO");
2824 * g_hash_table_insert (h, "3", "THREE");
2825 * g_hash_table_insert (h, "4", "FOUR");
2827 * reg = g_regex_new ("1|2|3|4", 0, 0, NULL);
2828 * res = g_regex_replace_eval (reg, text, -1, 0, 0, eval_cb, h, NULL);
2829 * g_hash_table_destroy (h);
2834 * Returns: a newly allocated string containing the replacements
2839 g_regex_replace_eval (const GRegex *regex,
2840 const gchar *string,
2842 gint start_position,
2843 GRegexMatchFlags match_options,
2844 GRegexEvalCallback eval,
2848 GMatchInfo *match_info;
2851 gboolean done = FALSE;
2852 GError *tmp_error = NULL;
2854 g_return_val_if_fail (regex != NULL, NULL);
2855 g_return_val_if_fail (string != NULL, NULL);
2856 g_return_val_if_fail (start_position >= 0, NULL);
2857 g_return_val_if_fail (eval != NULL, NULL);
2858 g_return_val_if_fail ((match_options & ~G_REGEX_MATCH_MASK) == 0, NULL);
2861 string_len = strlen (string);
2863 result = g_string_sized_new (string_len);
2865 /* run down the string making matches. */
2866 g_regex_match_full (regex, string, string_len, start_position,
2867 match_options, &match_info, &tmp_error);
2868 while (!done && g_match_info_matches (match_info))
2870 g_string_append_len (result,
2872 match_info->offsets[0] - str_pos);
2873 done = (*eval) (match_info, result, user_data);
2874 str_pos = match_info->offsets[1];
2875 g_match_info_next (match_info, &tmp_error);
2877 g_match_info_free (match_info);
2878 if (tmp_error != NULL)
2880 g_propagate_error (error, tmp_error);
2881 g_string_free (result, TRUE);
2885 g_string_append_len (result, string + str_pos, string_len - str_pos);
2886 return g_string_free (result, FALSE);
2890 * g_regex_check_replacement:
2891 * @replacement: the replacement string
2892 * @has_references: (out) (allow-none): location to store information about
2893 * references in @replacement or %NULL
2894 * @error: location to store error
2896 * Checks whether @replacement is a valid replacement string
2897 * (see g_regex_replace()), i.e. that all escape sequences in
2900 * If @has_references is not %NULL then @replacement is checked
2901 * for pattern references. For instance, replacement text 'foo\n'
2902 * does not contain references and may be evaluated without information
2903 * about actual match, but '\0\1' (whole match followed by first
2904 * subpattern) requires valid #GMatchInfo object.
2906 * Returns: whether @replacement is a valid replacement string
2911 g_regex_check_replacement (const gchar *replacement,
2912 gboolean *has_references,
2918 list = split_replacement (replacement, &tmp);
2922 g_propagate_error (error, tmp);
2927 *has_references = interpolation_list_needs_match (list);
2929 g_list_free_full (list, (GDestroyNotify) free_interpolation_data);
2935 * g_regex_escape_nul:
2936 * @string: the string to escape
2937 * @length: the length of @string
2939 * Escapes the nul characters in @string to "\x00". It can be used
2940 * to compile a regex with embedded nul characters.
2942 * For completeness, @length can be -1 for a nul-terminated string.
2943 * In this case the output string will be of course equal to @string.
2945 * Returns: a newly-allocated escaped string
2950 g_regex_escape_nul (const gchar *string,
2954 const gchar *p, *piece_start, *end;
2957 g_return_val_if_fail (string != NULL, NULL);
2960 return g_strdup (string);
2962 end = string + length;
2963 p = piece_start = string;
2964 escaped = g_string_sized_new (length + 1);
2972 if (p != piece_start)
2974 /* copy the previous piece. */
2975 g_string_append_len (escaped, piece_start, p - piece_start);
2977 if ((backslashes & 1) == 0)
2978 g_string_append_c (escaped, '\\');
2979 g_string_append_c (escaped, 'x');
2980 g_string_append_c (escaped, '0');
2981 g_string_append_c (escaped, '0');
2991 p = g_utf8_next_char (p);
2996 if (piece_start < end)
2997 g_string_append_len (escaped, piece_start, end - piece_start);
2999 return g_string_free (escaped, FALSE);
3003 * g_regex_escape_string:
3004 * @string: (array length=length): the string to escape
3005 * @length: the length of @string, or -1 if @string is nul-terminated
3007 * Escapes the special characters used for regular expressions
3008 * in @string, for instance "a.b*c" becomes "a\.b\*c". This
3009 * function is useful to dynamically generate regular expressions.
3011 * @string can contain nul characters that are replaced with "\0",
3012 * in this case remember to specify the correct length of @string
3015 * Returns: a newly-allocated escaped string
3020 g_regex_escape_string (const gchar *string,
3024 const char *p, *piece_start, *end;
3026 g_return_val_if_fail (string != NULL, NULL);
3029 length = strlen (string);
3031 end = string + length;
3032 p = piece_start = string;
3033 escaped = g_string_sized_new (length + 1);
3054 if (p != piece_start)
3055 /* copy the previous piece. */
3056 g_string_append_len (escaped, piece_start, p - piece_start);
3057 g_string_append_c (escaped, '\\');
3059 g_string_append_c (escaped, '0');
3061 g_string_append_c (escaped, *p);
3065 p = g_utf8_next_char (p);
3070 if (piece_start < end)
3071 g_string_append_len (escaped, piece_start, end - piece_start);
3073 return g_string_free (escaped, FALSE);