1/* GRegex -- regular expression API wrapper around PCRE.
2 *
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>
6 *
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.
11 *
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.
16 *
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
20 */
21
22#ifndef __G_REGEX_H__
23#define __G_REGEX_H__
24
25#if !defined (__GLIB_H_INSIDE__) && !defined (GLIB_COMPILATION)
26#error "Only <glib.h> can be included directly."
27#endif
28
29#include <glib/gerror.h>
30#include <glib/gstring.h>
31
32G_BEGIN_DECLS
33
34/**
35 * GRegexError:
36 * @G_REGEX_ERROR_COMPILE: Compilation of the regular expression failed.
37 * @G_REGEX_ERROR_OPTIMIZE: Optimization of the regular expression failed.
38 * @G_REGEX_ERROR_REPLACE: Replacement failed due to an ill-formed replacement
39 * string.
40 * @G_REGEX_ERROR_MATCH: The match process failed.
41 * @G_REGEX_ERROR_INTERNAL: Internal error of the regular expression engine.
42 * Since 2.16
43 * @G_REGEX_ERROR_STRAY_BACKSLASH: "\\" at end of pattern. Since 2.16
44 * @G_REGEX_ERROR_MISSING_CONTROL_CHAR: "\\c" at end of pattern. Since 2.16
45 * @G_REGEX_ERROR_UNRECOGNIZED_ESCAPE: Unrecognized character follows "\\".
46 * Since 2.16
47 * @G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER: Numbers out of order in "{}"
48 * quantifier. Since 2.16
49 * @G_REGEX_ERROR_QUANTIFIER_TOO_BIG: Number too big in "{}" quantifier.
50 * Since 2.16
51 * @G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS: Missing terminating "]" for
52 * character class. Since 2.16
53 * @G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS: Invalid escape sequence
54 * in character class. Since 2.16
55 * @G_REGEX_ERROR_RANGE_OUT_OF_ORDER: Range out of order in character class.
56 * Since 2.16
57 * @G_REGEX_ERROR_NOTHING_TO_REPEAT: Nothing to repeat. Since 2.16
58 * @G_REGEX_ERROR_UNRECOGNIZED_CHARACTER: Unrecognized character after "(?",
59 * "(?<" or "(?P". Since 2.16
60 * @G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS: POSIX named classes are
61 * supported only within a class. Since 2.16
62 * @G_REGEX_ERROR_UNMATCHED_PARENTHESIS: Missing terminating ")" or ")"
63 * without opening "(". Since 2.16
64 * @G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE: Reference to non-existent
65 * subpattern. Since 2.16
66 * @G_REGEX_ERROR_UNTERMINATED_COMMENT: Missing terminating ")" after comment.
67 * Since 2.16
68 * @G_REGEX_ERROR_EXPRESSION_TOO_LARGE: Regular expression too large.
69 * Since 2.16
70 * @G_REGEX_ERROR_MEMORY_ERROR: Failed to get memory. Since 2.16
71 * @G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND: Lookbehind assertion is not
72 * fixed length. Since 2.16
73 * @G_REGEX_ERROR_MALFORMED_CONDITION: Malformed number or name after "(?(".
74 * Since 2.16
75 * @G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES: Conditional group contains
76 * more than two branches. Since 2.16
77 * @G_REGEX_ERROR_ASSERTION_EXPECTED: Assertion expected after "(?(".
78 * Since 2.16
79 * @G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME: Unknown POSIX class name.
80 * Since 2.16
81 * @G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED: POSIX collating
82 * elements are not supported. Since 2.16
83 * @G_REGEX_ERROR_HEX_CODE_TOO_LARGE: Character value in "\\x{...}" sequence
84 * is too large. Since 2.16
85 * @G_REGEX_ERROR_INVALID_CONDITION: Invalid condition "(?(0)". Since 2.16
86 * @G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND: \\C not allowed in
87 * lookbehind assertion. Since 2.16
88 * @G_REGEX_ERROR_INFINITE_LOOP: Recursive call could loop indefinitely.
89 * Since 2.16
90 * @G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR: Missing terminator
91 * in subpattern name. Since 2.16
92 * @G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME: Two named subpatterns have
93 * the same name. Since 2.16
94 * @G_REGEX_ERROR_MALFORMED_PROPERTY: Malformed "\\P" or "\\p" sequence.
95 * Since 2.16
96 * @G_REGEX_ERROR_UNKNOWN_PROPERTY: Unknown property name after "\\P" or
97 * "\\p". Since 2.16
98 * @G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG: Subpattern name is too long
99 * (maximum 32 characters). Since 2.16
100 * @G_REGEX_ERROR_TOO_MANY_SUBPATTERNS: Too many named subpatterns (maximum
101 * 10,000). Since 2.16
102 * @G_REGEX_ERROR_INVALID_OCTAL_VALUE: Octal value is greater than "\\377".
103 * Since 2.16
104 * @G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE: "DEFINE" group contains more
105 * than one branch. Since 2.16
106 * @G_REGEX_ERROR_DEFINE_REPETION: Repeating a "DEFINE" group is not allowed.
107 * This error is never raised. Since: 2.16 Deprecated: 2.34
108 * @G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS: Inconsistent newline options.
109 * Since 2.16
110 * @G_REGEX_ERROR_MISSING_BACK_REFERENCE: "\\g" is not followed by a braced,
111 * angle-bracketed, or quoted name or number, or by a plain number. Since: 2.16
112 * @G_REGEX_ERROR_INVALID_RELATIVE_REFERENCE: relative reference must not be zero. Since: 2.34
113 * @G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_FORBIDDEN: the backtracing
114 * control verb used does not allow an argument. Since: 2.34
115 * @G_REGEX_ERROR_UNKNOWN_BACKTRACKING_CONTROL_VERB: unknown backtracing
116 * control verb. Since: 2.34
117 * @G_REGEX_ERROR_NUMBER_TOO_BIG: number is too big in escape sequence. Since: 2.34
118 * @G_REGEX_ERROR_MISSING_SUBPATTERN_NAME: Missing subpattern name. Since: 2.34
119 * @G_REGEX_ERROR_MISSING_DIGIT: Missing digit. Since 2.34
120 * @G_REGEX_ERROR_INVALID_DATA_CHARACTER: In JavaScript compatibility mode,
121 * "[" is an invalid data character. Since: 2.34
122 * @G_REGEX_ERROR_EXTRA_SUBPATTERN_NAME: different names for subpatterns of the
123 * same number are not allowed. Since: 2.34
124 * @G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_REQUIRED: the backtracing control
125 * verb requires an argument. Since: 2.34
126 * @G_REGEX_ERROR_INVALID_CONTROL_CHAR: "\\c" must be followed by an ASCII
127 * character. Since: 2.34
128 * @G_REGEX_ERROR_MISSING_NAME: "\\k" is not followed by a braced, angle-bracketed, or
129 * quoted name. Since: 2.34
130 * @G_REGEX_ERROR_NOT_SUPPORTED_IN_CLASS: "\\N" is not supported in a class. Since: 2.34
131 * @G_REGEX_ERROR_TOO_MANY_FORWARD_REFERENCES: too many forward references. Since: 2.34
132 * @G_REGEX_ERROR_NAME_TOO_LONG: the name is too long in "(*MARK)", "(*PRUNE)",
133 * "(*SKIP)", or "(*THEN)". Since: 2.34
134 * @G_REGEX_ERROR_CHARACTER_VALUE_TOO_LARGE: the character value in the \\u sequence is
135 * too large. Since: 2.34
136 *
137 * Error codes returned by regular expressions functions.
138 *
139 * Since: 2.14
140 */
141typedef enum
142{
143 G_REGEX_ERROR_COMPILE,
144 G_REGEX_ERROR_OPTIMIZE,
145 G_REGEX_ERROR_REPLACE,
146 G_REGEX_ERROR_MATCH,
147 G_REGEX_ERROR_INTERNAL,
148
149 /* These are the error codes from PCRE + 100 */
150 G_REGEX_ERROR_STRAY_BACKSLASH = 101,
151 G_REGEX_ERROR_MISSING_CONTROL_CHAR = 102,
152 G_REGEX_ERROR_UNRECOGNIZED_ESCAPE = 103,
153 G_REGEX_ERROR_QUANTIFIERS_OUT_OF_ORDER = 104,
154 G_REGEX_ERROR_QUANTIFIER_TOO_BIG = 105,
155 G_REGEX_ERROR_UNTERMINATED_CHARACTER_CLASS = 106,
156 G_REGEX_ERROR_INVALID_ESCAPE_IN_CHARACTER_CLASS = 107,
157 G_REGEX_ERROR_RANGE_OUT_OF_ORDER = 108,
158 G_REGEX_ERROR_NOTHING_TO_REPEAT = 109,
159 G_REGEX_ERROR_UNRECOGNIZED_CHARACTER = 112,
160 G_REGEX_ERROR_POSIX_NAMED_CLASS_OUTSIDE_CLASS = 113,
161 G_REGEX_ERROR_UNMATCHED_PARENTHESIS = 114,
162 G_REGEX_ERROR_INEXISTENT_SUBPATTERN_REFERENCE = 115,
163 G_REGEX_ERROR_UNTERMINATED_COMMENT = 118,
164 G_REGEX_ERROR_EXPRESSION_TOO_LARGE = 120,
165 G_REGEX_ERROR_MEMORY_ERROR = 121,
166 G_REGEX_ERROR_VARIABLE_LENGTH_LOOKBEHIND = 125,
167 G_REGEX_ERROR_MALFORMED_CONDITION = 126,
168 G_REGEX_ERROR_TOO_MANY_CONDITIONAL_BRANCHES = 127,
169 G_REGEX_ERROR_ASSERTION_EXPECTED = 128,
170 G_REGEX_ERROR_UNKNOWN_POSIX_CLASS_NAME = 130,
171 G_REGEX_ERROR_POSIX_COLLATING_ELEMENTS_NOT_SUPPORTED = 131,
172 G_REGEX_ERROR_HEX_CODE_TOO_LARGE = 134,
173 G_REGEX_ERROR_INVALID_CONDITION = 135,
174 G_REGEX_ERROR_SINGLE_BYTE_MATCH_IN_LOOKBEHIND = 136,
175 G_REGEX_ERROR_INFINITE_LOOP = 140,
176 G_REGEX_ERROR_MISSING_SUBPATTERN_NAME_TERMINATOR = 142,
177 G_REGEX_ERROR_DUPLICATE_SUBPATTERN_NAME = 143,
178 G_REGEX_ERROR_MALFORMED_PROPERTY = 146,
179 G_REGEX_ERROR_UNKNOWN_PROPERTY = 147,
180 G_REGEX_ERROR_SUBPATTERN_NAME_TOO_LONG = 148,
181 G_REGEX_ERROR_TOO_MANY_SUBPATTERNS = 149,
182 G_REGEX_ERROR_INVALID_OCTAL_VALUE = 151,
183 G_REGEX_ERROR_TOO_MANY_BRANCHES_IN_DEFINE = 154,
184 G_REGEX_ERROR_DEFINE_REPETION = 155,
185 G_REGEX_ERROR_INCONSISTENT_NEWLINE_OPTIONS = 156,
186 G_REGEX_ERROR_MISSING_BACK_REFERENCE = 157,
187 G_REGEX_ERROR_INVALID_RELATIVE_REFERENCE = 158,
188 G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_FORBIDDEN = 159,
189 G_REGEX_ERROR_UNKNOWN_BACKTRACKING_CONTROL_VERB = 160,
190 G_REGEX_ERROR_NUMBER_TOO_BIG = 161,
191 G_REGEX_ERROR_MISSING_SUBPATTERN_NAME = 162,
192 G_REGEX_ERROR_MISSING_DIGIT = 163,
193 G_REGEX_ERROR_INVALID_DATA_CHARACTER = 164,
194 G_REGEX_ERROR_EXTRA_SUBPATTERN_NAME = 165,
195 G_REGEX_ERROR_BACKTRACKING_CONTROL_VERB_ARGUMENT_REQUIRED = 166,
196 G_REGEX_ERROR_INVALID_CONTROL_CHAR = 168,
197 G_REGEX_ERROR_MISSING_NAME = 169,
198 G_REGEX_ERROR_NOT_SUPPORTED_IN_CLASS = 171,
199 G_REGEX_ERROR_TOO_MANY_FORWARD_REFERENCES = 172,
200 G_REGEX_ERROR_NAME_TOO_LONG = 175,
201 G_REGEX_ERROR_CHARACTER_VALUE_TOO_LARGE = 176
202} GRegexError;
203
204/**
205 * G_REGEX_ERROR:
206 *
207 * Error domain for regular expressions. Errors in this domain will be
208 * from the #GRegexError enumeration. See #GError for information on
209 * error domains.
210 *
211 * Since: 2.14
212 */
213#define G_REGEX_ERROR g_regex_error_quark ()
214
215GLIB_AVAILABLE_IN_ALL
216GQuark g_regex_error_quark (void);
217
218/**
219 * GRegexCompileFlags:
220 * @G_REGEX_CASELESS: Letters in the pattern match both upper- and
221 * lowercase letters. This option can be changed within a pattern
222 * by a "(?i)" option setting.
223 * @G_REGEX_MULTILINE: By default, GRegex treats the strings as consisting
224 * of a single line of characters (even if it actually contains
225 * newlines). The "start of line" metacharacter ("^") matches only
226 * at the start of the string, while the "end of line" metacharacter
227 * ("$") matches only at the end of the string, or before a terminating
228 * newline (unless #G_REGEX_DOLLAR_ENDONLY is set). When
229 * #G_REGEX_MULTILINE is set, the "start of line" and "end of line"
230 * constructs match immediately following or immediately before any
231 * newline in the string, respectively, as well as at the very start
232 * and end. This can be changed within a pattern by a "(?m)" option
233 * setting.
234 * @G_REGEX_DOTALL: A dot metacharater (".") in the pattern matches all
235 * characters, including newlines. Without it, newlines are excluded.
236 * This option can be changed within a pattern by a ("?s") option setting.
237 * @G_REGEX_EXTENDED: Whitespace data characters in the pattern are
238 * totally ignored except when escaped or inside a character class.
239 * Whitespace does not include the VT character (code 11). In addition,
240 * characters between an unescaped "#" outside a character class and
241 * the next newline character, inclusive, are also ignored. This can
242 * be changed within a pattern by a "(?x)" option setting.
243 * @G_REGEX_ANCHORED: The pattern is forced to be "anchored", that is,
244 * it is constrained to match only at the first matching point in the
245 * string that is being searched. This effect can also be achieved by
246 * appropriate constructs in the pattern itself such as the "^"
247 * metacharater.
248 * @G_REGEX_DOLLAR_ENDONLY: A dollar metacharacter ("$") in the pattern
249 * matches only at the end of the string. Without this option, a
250 * dollar also matches immediately before the final character if
251 * it is a newline (but not before any other newlines). This option
252 * is ignored if #G_REGEX_MULTILINE is set.
253 * @G_REGEX_UNGREEDY: Inverts the "greediness" of the quantifiers so that
254 * they are not greedy by default, but become greedy if followed by "?".
255 * It can also be set by a "(?U)" option setting within the pattern.
256 * @G_REGEX_RAW: Usually strings must be valid UTF-8 strings, using this
257 * flag they are considered as a raw sequence of bytes.
258 * @G_REGEX_NO_AUTO_CAPTURE: Disables the use of numbered capturing
259 * parentheses in the pattern. Any opening parenthesis that is not
260 * followed by "?" behaves as if it were followed by "?:" but named
261 * parentheses can still be used for capturing (and they acquire numbers
262 * in the usual way).
263 * @G_REGEX_OPTIMIZE: Optimize the regular expression. If the pattern will
264 * be used many times, then it may be worth the effort to optimize it
265 * to improve the speed of matches.
266 * @G_REGEX_FIRSTLINE: Limits an unanchored pattern to match before (or at) the
267 * first newline. Since: 2.34
268 * @G_REGEX_DUPNAMES: Names used to identify capturing subpatterns need not
269 * be unique. This can be helpful for certain types of pattern when it
270 * is known that only one instance of the named subpattern can ever be
271 * matched.
272 * @G_REGEX_NEWLINE_CR: Usually any newline character or character sequence is
273 * recognized. If this option is set, the only recognized newline character
274 * is '\r'.
275 * @G_REGEX_NEWLINE_LF: Usually any newline character or character sequence is
276 * recognized. If this option is set, the only recognized newline character
277 * is '\n'.
278 * @G_REGEX_NEWLINE_CRLF: Usually any newline character or character sequence is
279 * recognized. If this option is set, the only recognized newline character
280 * sequence is '\r\n'.
281 * @G_REGEX_NEWLINE_ANYCRLF: Usually any newline character or character sequence
282 * is recognized. If this option is set, the only recognized newline character
283 * sequences are '\r', '\n', and '\r\n'. Since: 2.34
284 * @G_REGEX_BSR_ANYCRLF: Usually any newline character or character sequence
285 * is recognised. If this option is set, then "\R" only recognizes the newline
286 * characters '\r', '\n' and '\r\n'. Since: 2.34
287 * @G_REGEX_JAVASCRIPT_COMPAT: Changes behaviour so that it is compatible with
288 * JavaScript rather than PCRE. Since: 2.34
289 *
290 * Flags specifying compile-time options.
291 *
292 * Since: 2.14
293 */
294/* Remember to update G_REGEX_COMPILE_MASK in gregex.c after
295 * adding a new flag.
296 */
297typedef enum
298{
299 G_REGEX_CASELESS = 1 << 0,
300 G_REGEX_MULTILINE = 1 << 1,
301 G_REGEX_DOTALL = 1 << 2,
302 G_REGEX_EXTENDED = 1 << 3,
303 G_REGEX_ANCHORED = 1 << 4,
304 G_REGEX_DOLLAR_ENDONLY = 1 << 5,
305 G_REGEX_UNGREEDY = 1 << 9,
306 G_REGEX_RAW = 1 << 11,
307 G_REGEX_NO_AUTO_CAPTURE = 1 << 12,
308 G_REGEX_OPTIMIZE = 1 << 13,
309 G_REGEX_FIRSTLINE = 1 << 18,
310 G_REGEX_DUPNAMES = 1 << 19,
311 G_REGEX_NEWLINE_CR = 1 << 20,
312 G_REGEX_NEWLINE_LF = 1 << 21,
313 G_REGEX_NEWLINE_CRLF = G_REGEX_NEWLINE_CR | G_REGEX_NEWLINE_LF,
314 G_REGEX_NEWLINE_ANYCRLF = G_REGEX_NEWLINE_CR | 1 << 22,
315 G_REGEX_BSR_ANYCRLF = 1 << 23,
316 G_REGEX_JAVASCRIPT_COMPAT = 1 << 25
317} GRegexCompileFlags;
318
319/**
320 * GRegexMatchFlags:
321 * @G_REGEX_MATCH_ANCHORED: The pattern is forced to be "anchored", that is,
322 * it is constrained to match only at the first matching point in the
323 * string that is being searched. This effect can also be achieved by
324 * appropriate constructs in the pattern itself such as the "^"
325 * metacharater.
326 * @G_REGEX_MATCH_NOTBOL: Specifies that first character of the string is
327 * not the beginning of a line, so the circumflex metacharacter should
328 * not match before it. Setting this without #G_REGEX_MULTILINE (at
329 * compile time) causes circumflex never to match. This option affects
330 * only the behaviour of the circumflex metacharacter, it does not
331 * affect "\A".
332 * @G_REGEX_MATCH_NOTEOL: Specifies that the end of the subject string is
333 * not the end of a line, so the dollar metacharacter should not match
334 * it nor (except in multiline mode) a newline immediately before it.
335 * Setting this without #G_REGEX_MULTILINE (at compile time) causes
336 * dollar never to match. This option affects only the behaviour of
337 * the dollar metacharacter, it does not affect "\Z" or "\z".
338 * @G_REGEX_MATCH_NOTEMPTY: An empty string is not considered to be a valid
339 * match if this option is set. If there are alternatives in the pattern,
340 * they are tried. If all the alternatives match the empty string, the
341 * entire match fails. For example, if the pattern "a?b?" is applied to
342 * a string not beginning with "a" or "b", it matches the empty string
343 * at the start of the string. With this flag set, this match is not
344 * valid, so GRegex searches further into the string for occurrences
345 * of "a" or "b".
346 * @G_REGEX_MATCH_PARTIAL: Turns on the partial matching feature, for more
347 * documentation on partial matching see g_match_info_is_partial_match().
348 * @G_REGEX_MATCH_NEWLINE_CR: Overrides the newline definition set when
349 * creating a new #GRegex, setting the '\r' character as line terminator.
350 * @G_REGEX_MATCH_NEWLINE_LF: Overrides the newline definition set when
351 * creating a new #GRegex, setting the '\n' character as line terminator.
352 * @G_REGEX_MATCH_NEWLINE_CRLF: Overrides the newline definition set when
353 * creating a new #GRegex, setting the '\r\n' characters sequence as line terminator.
354 * @G_REGEX_MATCH_NEWLINE_ANY: Overrides the newline definition set when
355 * creating a new #GRegex, any Unicode newline sequence
356 * is recognised as a newline. These are '\r', '\n' and '\rn', and the
357 * single characters U+000B LINE TABULATION, U+000C FORM FEED (FF),
358 * U+0085 NEXT LINE (NEL), U+2028 LINE SEPARATOR and
359 * U+2029 PARAGRAPH SEPARATOR.
360 * @G_REGEX_MATCH_NEWLINE_ANYCRLF: Overrides the newline definition set when
361 * creating a new #GRegex; any '\r', '\n', or '\r\n' character sequence
362 * is recognized as a newline. Since: 2.34
363 * @G_REGEX_MATCH_BSR_ANYCRLF: Overrides the newline definition for "\R" set when
364 * creating a new #GRegex; only '\r', '\n', or '\r\n' character sequences
365 * are recognized as a newline by "\R". Since: 2.34
366 * @G_REGEX_MATCH_BSR_ANY: Overrides the newline definition for "\R" set when
367 * creating a new #GRegex; any Unicode newline character or character sequence
368 * are recognized as a newline by "\R". These are '\r', '\n' and '\rn', and the
369 * single characters U+000B LINE TABULATION, U+000C FORM FEED (FF),
370 * U+0085 NEXT LINE (NEL), U+2028 LINE SEPARATOR and
371 * U+2029 PARAGRAPH SEPARATOR. Since: 2.34
372 * @G_REGEX_MATCH_PARTIAL_SOFT: An alias for #G_REGEX_MATCH_PARTIAL. Since: 2.34
373 * @G_REGEX_MATCH_PARTIAL_HARD: Turns on the partial matching feature. In contrast to
374 * to #G_REGEX_MATCH_PARTIAL_SOFT, this stops matching as soon as a partial match
375 * is found, without continuing to search for a possible complete match. See
376 * g_match_info_is_partial_match() for more information. Since: 2.34
377 * @G_REGEX_MATCH_NOTEMPTY_ATSTART: Like #G_REGEX_MATCH_NOTEMPTY, but only applied to
378 * the start of the matched string. For anchored
379 * patterns this can only happen for pattern containing "\K". Since: 2.34
380 *
381 * Flags specifying match-time options.
382 *
383 * Since: 2.14
384 */
385/* Remember to update G_REGEX_MATCH_MASK in gregex.c after
386 * adding a new flag. */
387typedef enum
388{
389 G_REGEX_MATCH_ANCHORED = 1 << 4,
390 G_REGEX_MATCH_NOTBOL = 1 << 7,
391 G_REGEX_MATCH_NOTEOL = 1 << 8,
392 G_REGEX_MATCH_NOTEMPTY = 1 << 10,
393 G_REGEX_MATCH_PARTIAL = 1 << 15,
394 G_REGEX_MATCH_NEWLINE_CR = 1 << 20,
395 G_REGEX_MATCH_NEWLINE_LF = 1 << 21,
396 G_REGEX_MATCH_NEWLINE_CRLF = G_REGEX_MATCH_NEWLINE_CR | G_REGEX_MATCH_NEWLINE_LF,
397 G_REGEX_MATCH_NEWLINE_ANY = 1 << 22,
398 G_REGEX_MATCH_NEWLINE_ANYCRLF = G_REGEX_MATCH_NEWLINE_CR | G_REGEX_MATCH_NEWLINE_ANY,
399 G_REGEX_MATCH_BSR_ANYCRLF = 1 << 23,
400 G_REGEX_MATCH_BSR_ANY = 1 << 24,
401 G_REGEX_MATCH_PARTIAL_SOFT = G_REGEX_MATCH_PARTIAL,
402 G_REGEX_MATCH_PARTIAL_HARD = 1 << 27,
403 G_REGEX_MATCH_NOTEMPTY_ATSTART = 1 << 28
404} GRegexMatchFlags;
405
406/**
407 * GRegex:
408 *
409 * A GRegex is the "compiled" form of a regular expression pattern.
410 * This structure is opaque and its fields cannot be accessed directly.
411 *
412 * Since: 2.14
413 */
414typedef struct _GRegex GRegex;
415
416
417/**
418 * GMatchInfo:
419 *
420 * A GMatchInfo is an opaque struct used to return information about
421 * matches.
422 */
423typedef struct _GMatchInfo GMatchInfo;
424
425/**
426 * GRegexEvalCallback:
427 * @match_info: the #GMatchInfo generated by the match.
428 * Use g_match_info_get_regex() and g_match_info_get_string() if you
429 * need the #GRegex or the matched string.
430 * @result: a #GString containing the new string
431 * @user_data: user data passed to g_regex_replace_eval()
432 *
433 * Specifies the type of the function passed to g_regex_replace_eval().
434 * It is called for each occurrence of the pattern in the string passed
435 * to g_regex_replace_eval(), and it should append the replacement to
436 * @result.
437 *
438 * Returns: %FALSE to continue the replacement process, %TRUE to stop it
439 *
440 * Since: 2.14
441 */
442typedef gboolean (*GRegexEvalCallback) (const GMatchInfo *match_info,
443 GString *result,
444 gpointer user_data);
445
446
447GLIB_AVAILABLE_IN_ALL
448GRegex *g_regex_new (const gchar *pattern,
449 GRegexCompileFlags compile_options,
450 GRegexMatchFlags match_options,
451 GError **error);
452GLIB_AVAILABLE_IN_ALL
453GRegex *g_regex_ref (GRegex *regex);
454GLIB_AVAILABLE_IN_ALL
455void g_regex_unref (GRegex *regex);
456GLIB_AVAILABLE_IN_ALL
457const gchar *g_regex_get_pattern (const GRegex *regex);
458GLIB_AVAILABLE_IN_ALL
459gint g_regex_get_max_backref (const GRegex *regex);
460GLIB_AVAILABLE_IN_ALL
461gint g_regex_get_capture_count (const GRegex *regex);
462GLIB_AVAILABLE_IN_ALL
463gboolean g_regex_get_has_cr_or_lf (const GRegex *regex);
464GLIB_AVAILABLE_IN_2_38
465gint g_regex_get_max_lookbehind (const GRegex *regex);
466GLIB_AVAILABLE_IN_ALL
467gint g_regex_get_string_number (const GRegex *regex,
468 const gchar *name);
469GLIB_AVAILABLE_IN_ALL
470gchar *g_regex_escape_string (const gchar *string,
471 gint length);
472GLIB_AVAILABLE_IN_ALL
473gchar *g_regex_escape_nul (const gchar *string,
474 gint length);
475
476GLIB_AVAILABLE_IN_ALL
477GRegexCompileFlags g_regex_get_compile_flags (const GRegex *regex);
478GLIB_AVAILABLE_IN_ALL
479GRegexMatchFlags g_regex_get_match_flags (const GRegex *regex);
480
481/* Matching. */
482GLIB_AVAILABLE_IN_ALL
483gboolean g_regex_match_simple (const gchar *pattern,
484 const gchar *string,
485 GRegexCompileFlags compile_options,
486 GRegexMatchFlags match_options);
487GLIB_AVAILABLE_IN_ALL
488gboolean g_regex_match (const GRegex *regex,
489 const gchar *string,
490 GRegexMatchFlags match_options,
491 GMatchInfo **match_info);
492GLIB_AVAILABLE_IN_ALL
493gboolean g_regex_match_full (const GRegex *regex,
494 const gchar *string,
495 gssize string_len,
496 gint start_position,
497 GRegexMatchFlags match_options,
498 GMatchInfo **match_info,
499 GError **error);
500GLIB_AVAILABLE_IN_ALL
501gboolean g_regex_match_all (const GRegex *regex,
502 const gchar *string,
503 GRegexMatchFlags match_options,
504 GMatchInfo **match_info);
505GLIB_AVAILABLE_IN_ALL
506gboolean g_regex_match_all_full (const GRegex *regex,
507 const gchar *string,
508 gssize string_len,
509 gint start_position,
510 GRegexMatchFlags match_options,
511 GMatchInfo **match_info,
512 GError **error);
513
514/* String splitting. */
515GLIB_AVAILABLE_IN_ALL
516gchar **g_regex_split_simple (const gchar *pattern,
517 const gchar *string,
518 GRegexCompileFlags compile_options,
519 GRegexMatchFlags match_options);
520GLIB_AVAILABLE_IN_ALL
521gchar **g_regex_split (const GRegex *regex,
522 const gchar *string,
523 GRegexMatchFlags match_options);
524GLIB_AVAILABLE_IN_ALL
525gchar **g_regex_split_full (const GRegex *regex,
526 const gchar *string,
527 gssize string_len,
528 gint start_position,
529 GRegexMatchFlags match_options,
530 gint max_tokens,
531 GError **error);
532
533/* String replacement. */
534GLIB_AVAILABLE_IN_ALL
535gchar *g_regex_replace (const GRegex *regex,
536 const gchar *string,
537 gssize string_len,
538 gint start_position,
539 const gchar *replacement,
540 GRegexMatchFlags match_options,
541 GError **error);
542GLIB_AVAILABLE_IN_ALL
543gchar *g_regex_replace_literal (const GRegex *regex,
544 const gchar *string,
545 gssize string_len,
546 gint start_position,
547 const gchar *replacement,
548 GRegexMatchFlags match_options,
549 GError **error);
550GLIB_AVAILABLE_IN_ALL
551gchar *g_regex_replace_eval (const GRegex *regex,
552 const gchar *string,
553 gssize string_len,
554 gint start_position,
555 GRegexMatchFlags match_options,
556 GRegexEvalCallback eval,
557 gpointer user_data,
558 GError **error);
559GLIB_AVAILABLE_IN_ALL
560gboolean g_regex_check_replacement (const gchar *replacement,
561 gboolean *has_references,
562 GError **error);
563
564/* Match info */
565GLIB_AVAILABLE_IN_ALL
566GRegex *g_match_info_get_regex (const GMatchInfo *match_info);
567GLIB_AVAILABLE_IN_ALL
568const gchar *g_match_info_get_string (const GMatchInfo *match_info);
569
570GLIB_AVAILABLE_IN_ALL
571GMatchInfo *g_match_info_ref (GMatchInfo *match_info);
572GLIB_AVAILABLE_IN_ALL
573void g_match_info_unref (GMatchInfo *match_info);
574GLIB_AVAILABLE_IN_ALL
575void g_match_info_free (GMatchInfo *match_info);
576GLIB_AVAILABLE_IN_ALL
577gboolean g_match_info_next (GMatchInfo *match_info,
578 GError **error);
579GLIB_AVAILABLE_IN_ALL
580gboolean g_match_info_matches (const GMatchInfo *match_info);
581GLIB_AVAILABLE_IN_ALL
582gint g_match_info_get_match_count (const GMatchInfo *match_info);
583GLIB_AVAILABLE_IN_ALL
584gboolean g_match_info_is_partial_match (const GMatchInfo *match_info);
585GLIB_AVAILABLE_IN_ALL
586gchar *g_match_info_expand_references(const GMatchInfo *match_info,
587 const gchar *string_to_expand,
588 GError **error);
589GLIB_AVAILABLE_IN_ALL
590gchar *g_match_info_fetch (const GMatchInfo *match_info,
591 gint match_num);
592GLIB_AVAILABLE_IN_ALL
593gboolean g_match_info_fetch_pos (const GMatchInfo *match_info,
594 gint match_num,
595 gint *start_pos,
596 gint *end_pos);
597GLIB_AVAILABLE_IN_ALL
598gchar *g_match_info_fetch_named (const GMatchInfo *match_info,
599 const gchar *name);
600GLIB_AVAILABLE_IN_ALL
601gboolean g_match_info_fetch_named_pos (const GMatchInfo *match_info,
602 const gchar *name,
603 gint *start_pos,
604 gint *end_pos);
605GLIB_AVAILABLE_IN_ALL
606gchar **g_match_info_fetch_all (const GMatchInfo *match_info);
607
608G_END_DECLS
609
610#endif /* __G_REGEX_H__ */
611