]> git.saurik.com Git - wxWidgets.git/blame - src/regex/regex.h
I changed something, but I forgot what.
[wxWidgets.git] / src / regex / regex.h
CommitLineData
652bbafa
VZ
1#ifndef _REGEX_H_
2#define _REGEX_H_ /* never again */
d37acbdf
RN
3/*
4 * regular expressions
5 *
3ca4086b
VS
6 * Copyright (c) 1998, 1999 Henry Spencer. All rights reserved.
7 *
d37acbdf
RN
8 * Development of this software was funded, in part, by Cray Research Inc.,
9 * UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
10 * Corporation, none of whom are responsible for the results. The author
3ca4086b
VS
11 * thanks all of them.
12 *
d37acbdf
RN
13 * Redistribution and use in source and binary forms -- with or without
14 * modification -- are permitted for any purpose, provided that
15 * redistributions in source form retain this entire copyright notice and
16 * indicate the origin and nature of any modifications.
17 *
18 * I'd appreciate being given credit for this package in the documentation
19 * of software which uses it, but that is not a requirement.
3ca4086b 20 *
d37acbdf
RN
21 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
22 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
23 * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
24 * HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
25 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
26 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
27 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
28 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
29 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
30 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31 *
3ca4086b
VS
32 *
33 *
34 * Prototypes etc. marked with "^" within comments get gathered up (and
35 * possibly edited) by the regfwd program and inserted near the bottom of
36 * this file.
37 *
38 * We offer the option of declaring one wide-character version of the
39 * RE functions as well as the char versions. To do that, define
40 * __REG_WIDE_T to the type of wide characters (unfortunately, there
41 * is no consensus that wchar_t is suitable) and __REG_WIDE_COMPILE and
42 * __REG_WIDE_EXEC to the names to be used for the compile and execute
43 * functions (suggestion: re_Xcomp and re_Xexec, where X is a letter
44 * suggestive of the wide type, e.g. re_ucomp and re_uexec for Unicode).
45 * For cranky old compilers, it may be necessary to do something like:
46 * #define __REG_WIDE_COMPILE(a,b,c,d) re_Xcomp(a,b,c,d)
47 * #define __REG_WIDE_EXEC(a,b,c,d,e,f,g) re_Xexec(a,b,c,d,e,f,g)
48 * rather than just #defining the names as parameterless macros.
49 *
50 * For some specialized purposes, it may be desirable to suppress the
51 * declarations of the "front end" functions, regcomp() and regexec(),
52 * or of the char versions of the compile and execute functions. To
53 * suppress the front-end functions, define __REG_NOFRONT. To suppress
54 * the char versions, define __REG_NOCHAR.
55 *
56 * The right place to do those defines (and some others you may want, see
57 * below) would be <sys/types.h>. If you don't have control of that file,
58 * the right place to add your own defines to this file is marked below.
59 * This is normally done automatically, by the makefile and regmkhdr, based
60 * on the contents of regcustom.h.
d37acbdf
RN
61 */
62
3ca4086b
VS
63
64
d37acbdf 65/*
3ca4086b 66 * voodoo for C++
d37acbdf 67 */
652bbafa
VZ
68#ifdef __cplusplus
69extern "C" {
70#endif
71
3ca4086b
VS
72
73
74/*
75 * Add your own defines, if needed, here.
76 */
77
78
79
80/*
81 * Location where a chunk of regcustom.h is automatically spliced into
82 * this file (working from its prototype, regproto.h).
83 */
84/* --- begin --- */
85/* ensure certain things don't sneak in from system headers */
86#ifdef __REG_WIDE_T
87#undef __REG_WIDE_T
88#endif
89#ifdef __REG_WIDE_COMPILE
90#undef __REG_WIDE_COMPILE
91#endif
92#ifdef __REG_WIDE_EXEC
93#undef __REG_WIDE_EXEC
94#endif
95#ifdef __REG_REGOFF_T
96#undef __REG_REGOFF_T
97#endif
98#ifdef __REG_VOID_T
99#undef __REG_VOID_T
100#endif
101#ifdef __REG_CONST
102#undef __REG_CONST
103#endif
104#ifdef __REG_NOFRONT
105#undef __REG_NOFRONT
106#endif
107#ifdef __REG_NOCHAR
108#undef __REG_NOCHAR
109#endif
1701c4d4
VS
110#if wxUSE_UNICODE
111# define __REG_WIDE_T wxChar
112# define __REG_WIDE_COMPILE re_comp
113# define __REG_WIDE_EXEC re_exec
114# define __REG_NOCHAR /* don't want the char versions */
115#endif
116#define __REG_NOFRONT /* don't want regcomp() and regexec() */
117#define _ANSI_ARGS_(x) x
fbeb9303
VZ
118
119/* under FreeBSD re_comp and re_exec are declared (differently) in unistd.h */
120#ifdef __FreeBSD__
121#define re_comp wx_re_comp
122#define re_exec wx_re_exec
123#endif
3ca4086b
VS
124/* --- end --- */
125
d37acbdf 126
d37acbdf
RN
127/*
128 * interface types etc.
129 */
130
131/*
132 * regoff_t has to be large enough to hold either off_t or ssize_t,
3ca4086b
VS
133 * and must be signed; it's only a guess that long is suitable, so we
134 * offer <sys/types.h> an override.
d37acbdf 135 */
3ca4086b
VS
136#ifdef __REG_REGOFF_T
137typedef __REG_REGOFF_T regoff_t;
138#else
d37acbdf 139typedef long regoff_t;
3ca4086b
VS
140#endif
141
142/*
143 * For benefit of old compilers, we offer <sys/types.h> the option of
144 * overriding the `void' type used to declare nonexistent return types.
145 */
146#ifdef __REG_VOID_T
147typedef __REG_VOID_T re_void;
148#else
149typedef void re_void;
150#endif
151
152/*
153 * Also for benefit of old compilers, <sys/types.h> can supply a macro
154 * which expands to a substitute for `const'.
155 */
156#ifndef __REG_CONST
157#define __REG_CONST const
158#endif
159
160
d37acbdf
RN
161
162/*
163 * other interface types
164 */
165
166/* the biggie, a compiled RE (or rather, a front end to same) */
3ca4086b
VS
167typedef struct {
168 int re_magic; /* magic number */
169 size_t re_nsub; /* number of subexpressions */
170 long re_info; /* information about RE */
171# define REG_UBACKREF 000001
172# define REG_ULOOKAHEAD 000002
173# define REG_UBOUNDS 000004
174# define REG_UBRACES 000010
175# define REG_UBSALNUM 000020
176# define REG_UPBOTCH 000040
177# define REG_UBBS 000100
178# define REG_UNONPOSIX 000200
179# define REG_UUNSPEC 000400
180# define REG_UUNPORT 001000
181# define REG_ULOCALE 002000
182# define REG_UEMPTYMATCH 004000
183# define REG_UIMPOSSIBLE 010000
184# define REG_USHORTEST 020000
185 int re_csize; /* sizeof(character) */
186 char *re_endp; /* backward compatibility kludge */
d37acbdf 187 /* the rest is opaque pointers to hidden innards */
3ca4086b
VS
188 char *re_guts; /* `char *' is more portable than `void *' */
189 char *re_fns;
652bbafa 190} regex_t;
d37acbdf
RN
191
192/* result reporting (may acquire more fields later) */
3ca4086b
VS
193typedef struct {
194 regoff_t rm_so; /* start of substring */
195 regoff_t rm_eo; /* end of substring */
652bbafa
VZ
196} regmatch_t;
197
d37acbdf 198/* supplementary control and reporting */
3ca4086b
VS
199typedef struct {
200 regmatch_t rm_extend; /* see REG_EXPECT */
d37acbdf 201} rm_detail_t;
652bbafa 202
652bbafa
VZ
203
204
d37acbdf 205/*
3ca4086b
VS
206 * compilation
207 ^ #ifndef __REG_NOCHAR
208 ^ int re_comp(regex_t *, __REG_CONST char *, size_t, int);
209 ^ #endif
210 ^ #ifndef __REG_NOFRONT
211 ^ int regcomp(regex_t *, __REG_CONST char *, int);
212 ^ #endif
213 ^ #ifdef __REG_WIDE_T
214 ^ int __REG_WIDE_COMPILE(regex_t *, __REG_CONST __REG_WIDE_T *, size_t, int);
215 ^ #endif
216 */
217#define REG_BASIC 000000 /* BREs (convenience) */
218#define REG_EXTENDED 000001 /* EREs */
219#define REG_ADVF 000002 /* advanced features in EREs */
220#define REG_ADVANCED 000003 /* AREs (which are also EREs) */
221#define REG_QUOTE 000004 /* no special characters, none */
222#define REG_NOSPEC REG_QUOTE /* historical synonym */
223#define REG_ICASE 000010 /* ignore case */
224#define REG_NOSUB 000020 /* don't care about subexpressions */
225#define REG_EXPANDED 000040 /* expanded format, white space & comments */
226#define REG_NLSTOP 000100 /* \n doesn't match . or [^ ] */
227#define REG_NLANCH 000200 /* ^ matches after \n, $ before */
228#define REG_NEWLINE 000300 /* newlines are line terminators */
229#define REG_PEND 000400 /* ugh -- backward-compatibility hack */
230#define REG_EXPECT 001000 /* report details on partial/limited matches */
231#define REG_BOSONLY 002000 /* temporary kludge for BOS-only matches */
232#define REG_DUMP 004000 /* none of your business :-) */
233#define REG_FAKE 010000 /* none of your business :-) */
234#define REG_PROGRESS 020000 /* none of your business :-) */
235
236
237
238/*
239 * execution
240 ^ #ifndef __REG_NOCHAR
241 ^ int re_exec(regex_t *, __REG_CONST char *, size_t,
242 ^ rm_detail_t *, size_t, regmatch_t [], int);
243 ^ #endif
244 ^ #ifndef __REG_NOFRONT
245 ^ int regexec(regex_t *, __REG_CONST char *, size_t, regmatch_t [], int);
246 ^ #endif
247 ^ #ifdef __REG_WIDE_T
248 ^ int __REG_WIDE_EXEC(regex_t *, __REG_CONST __REG_WIDE_T *, size_t,
249 ^ rm_detail_t *, size_t, regmatch_t [], int);
250 ^ #endif
d37acbdf 251 */
3ca4086b
VS
252#define REG_NOTBOL 0001 /* BOS is not BOL */
253#define REG_NOTEOL 0002 /* EOS is not EOL */
254#define REG_STARTEND 0004 /* backward compatibility kludge */
255#define REG_FTRACE 0010 /* none of your business */
256#define REG_MTRACE 0020 /* none of your business */
257#define REG_SMALL 0040 /* none of your business */
652bbafa
VZ
258
259
d37acbdf
RN
260
261/*
3ca4086b
VS
262 * misc generics (may be more functions here eventually)
263 ^ re_void regfree(regex_t *);
d37acbdf 264 */
3ca4086b 265
d37acbdf
RN
266
267
268/*
269 * error reporting
270 * Be careful if modifying the list of error codes -- the table used by
271 * regerror() is generated automatically from this file!
3ca4086b
VS
272 *
273 * Note that there is no wide-char variant of regerror at this time; what
274 * kind of character is used for error reports is independent of what kind
275 * is used in matching.
276 *
277 ^ extern size_t regerror(int, __REG_CONST regex_t *, char *, size_t);
d37acbdf 278 */
3ca4086b
VS
279#define REG_OKAY 0 /* no errors detected */
280#define REG_NOMATCH 1 /* failed to match */
281#define REG_BADPAT 2 /* invalid regexp */
282#define REG_ECOLLATE 3 /* invalid collating element */
283#define REG_ECTYPE 4 /* invalid character class */
284#define REG_EESCAPE 5 /* invalid escape \ sequence */
285#define REG_ESUBREG 6 /* invalid backreference number */
286#define REG_EBRACK 7 /* brackets [] not balanced */
287#define REG_EPAREN 8 /* parentheses () not balanced */
288#define REG_EBRACE 9 /* braces {} not balanced */
289#define REG_BADBR 10 /* invalid repetition count(s) */
290#define REG_ERANGE 11 /* invalid character range */
291#define REG_ESPACE 12 /* out of memory */
292#define REG_BADRPT 13 /* quantifier operand invalid */
293#define REG_ASSERT 15 /* "can't happen" -- you found a bug */
294#define REG_INVARG 16 /* invalid argument to regex function */
295#define REG_MIXED 17 /* character widths of regex and string differ */
296#define REG_BADOPT 18 /* invalid embedded option */
d37acbdf 297/* two specials for debugging and testing */
3ca4086b
VS
298#define REG_ATOI 101 /* convert error-code name to number */
299#define REG_ITOA 102 /* convert error-code number to name */
d37acbdf
RN
300
301
302
303/*
3ca4086b 304 * the prototypes, as possibly munched by regfwd
d37acbdf 305 */
3ca4086b
VS
306/* =====^!^===== begin forwards =====^!^===== */
307/* automatically gathered by fwd; do not hand-edit */
308/* === regproto.h === */
309#ifndef __REG_NOCHAR
310int re_comp _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, int));
311#endif
312#ifndef __REG_NOFRONT
313int regcomp _ANSI_ARGS_((regex_t *, __REG_CONST char *, int));
314#endif
315#ifdef __REG_WIDE_T
316int __REG_WIDE_COMPILE _ANSI_ARGS_((regex_t *, __REG_CONST __REG_WIDE_T *, size_t, int));
317#endif
318#ifndef __REG_NOCHAR
319int re_exec _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, rm_detail_t *, size_t, regmatch_t [], int));
320#endif
321#ifndef __REG_NOFRONT
322int regexec _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, regmatch_t [], int));
323#endif
324#ifdef __REG_WIDE_T
325int __REG_WIDE_EXEC _ANSI_ARGS_((regex_t *, __REG_CONST __REG_WIDE_T *, size_t, rm_detail_t *, size_t, regmatch_t [], int));
326#endif
327re_void regfree _ANSI_ARGS_((regex_t *));
328extern size_t regerror _ANSI_ARGS_((int, __REG_CONST regex_t *, char *, size_t));
329/* automatically gathered by fwd; do not hand-edit */
330/* =====^!^===== end forwards =====^!^===== */
652bbafa 331
3ca4086b
VS
332
333
334/*
335 * more C++ voodoo
336 */
652bbafa
VZ
337#ifdef __cplusplus
338}
339#endif
d37acbdf 340
3ca4086b
VS
341
342
343#endif