1 |
/* $Header$ */
|
2 |
|
3 |
#ifndef _REGEX_H_
|
4 |
#define _REGEX_H_ /* never again */
|
5 |
/*
|
6 |
* regular expressions
|
7 |
*
|
8 |
* Copyright (c) 1998, 1999 Henry Spencer. All rights reserved.
|
9 |
*
|
10 |
* Development of this software was funded, in part, by Cray Research Inc.,
|
11 |
* UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
|
12 |
* Corporation, none of whom are responsible for the results. The author
|
13 |
* thanks all of them.
|
14 |
*
|
15 |
* Redistribution and use in source and binary forms -- with or without
|
16 |
* modification -- are permitted for any purpose, provided that
|
17 |
* redistributions in source form retain this entire copyright notice and
|
18 |
* indicate the origin and nature of any modifications.
|
19 |
*
|
20 |
* I'd appreciate being given credit for this package in the documentation
|
21 |
* of software which uses it, but that is not a requirement.
|
22 |
*
|
23 |
* THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
|
24 |
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
|
25 |
* AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
|
26 |
* HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
|
27 |
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
28 |
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
|
29 |
* OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
|
30 |
* WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
|
31 |
* OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
|
32 |
* ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
33 |
*
|
34 |
*
|
35 |
*
|
36 |
* Prototypes etc. marked with "^" within comments get gathered up (and
|
37 |
* possibly edited) by the regfwd program and inserted near the bottom of
|
38 |
* this file.
|
39 |
*
|
40 |
* We offer the option of declaring one wide-character version of the
|
41 |
* RE functions as well as the char versions. To do that, define
|
42 |
* __REG_WIDE_T to the type of wide characters (unfortunately, there
|
43 |
* is no consensus that wchar_t is suitable) and __REG_WIDE_COMPILE and
|
44 |
* __REG_WIDE_EXEC to the names to be used for the compile and execute
|
45 |
* functions (suggestion: re_Xcomp and re_Xexec, where X is a letter
|
46 |
* suggestive of the wide type, e.g. re_ucomp and re_uexec for Unicode).
|
47 |
* For cranky old compilers, it may be necessary to do something like:
|
48 |
* #define __REG_WIDE_COMPILE(a,b,c,d) re_Xcomp(a,b,c,d)
|
49 |
* #define __REG_WIDE_EXEC(a,b,c,d,e,f,g) re_Xexec(a,b,c,d,e,f,g)
|
50 |
* rather than just #defining the names as parameterless macros.
|
51 |
*
|
52 |
* For some specialized purposes, it may be desirable to suppress the
|
53 |
* declarations of the "front end" functions, regcomp() and regexec(),
|
54 |
* or of the char versions of the compile and execute functions. To
|
55 |
* suppress the front-end functions, define __REG_NOFRONT. To suppress
|
56 |
* the char versions, define __REG_NOCHAR.
|
57 |
*
|
58 |
* The right place to do those defines (and some others you may want, see
|
59 |
* below) would be <sys/types.h>. If you don't have control of that file,
|
60 |
* the right place to add your own defines to this file is marked below.
|
61 |
* This is normally done automatically, by the makefile and regmkhdr, based
|
62 |
* on the contents of regcustom.h.
|
63 |
*/
|
64 |
|
65 |
|
66 |
|
67 |
/*
|
68 |
* voodoo for C++
|
69 |
*/
|
70 |
#ifdef __cplusplus
|
71 |
extern "C" {
|
72 |
#endif
|
73 |
|
74 |
|
75 |
|
76 |
/*
|
77 |
* Add your own defines, if needed, here.
|
78 |
*/
|
79 |
|
80 |
|
81 |
|
82 |
/*
|
83 |
* Location where a chunk of regcustom.h is automatically spliced into
|
84 |
* this file (working from its prototype, regproto.h).
|
85 |
*/
|
86 |
/* --- begin --- */
|
87 |
/* ensure certain things don't sneak in from system headers */
|
88 |
#ifdef __REG_WIDE_T
|
89 |
#undef __REG_WIDE_T
|
90 |
#endif
|
91 |
#ifdef __REG_WIDE_COMPILE
|
92 |
#undef __REG_WIDE_COMPILE
|
93 |
#endif
|
94 |
#ifdef __REG_WIDE_EXEC
|
95 |
#undef __REG_WIDE_EXEC
|
96 |
#endif
|
97 |
#ifdef __REG_REGOFF_T
|
98 |
#undef __REG_REGOFF_T
|
99 |
#endif
|
100 |
#ifdef __REG_VOID_T
|
101 |
#undef __REG_VOID_T
|
102 |
#endif
|
103 |
#ifdef __REG_CONST
|
104 |
#undef __REG_CONST
|
105 |
#endif
|
106 |
#ifdef __REG_NOFRONT
|
107 |
#undef __REG_NOFRONT
|
108 |
#endif
|
109 |
#ifdef __REG_NOCHAR
|
110 |
#undef __REG_NOCHAR
|
111 |
#endif
|
112 |
/* interface types */
|
113 |
#define __REG_WIDE_T Tcl_UniChar
|
114 |
#define __REG_REGOFF_T long /* not really right, but good enough... */
|
115 |
#define __REG_VOID_T VOID
|
116 |
#define __REG_CONST CONST
|
117 |
/* names and declarations */
|
118 |
#define __REG_WIDE_COMPILE TclReComp
|
119 |
#define __REG_WIDE_EXEC TclReExec
|
120 |
#define __REG_NOFRONT /* don't want regcomp() and regexec() */
|
121 |
#define __REG_NOCHAR /* or the char versions */
|
122 |
#define regfree TclReFree
|
123 |
#define regerror TclReError
|
124 |
/* --- end --- */
|
125 |
|
126 |
|
127 |
/*
|
128 |
* interface types etc.
|
129 |
*/
|
130 |
|
131 |
/*
|
132 |
* regoff_t has to be large enough to hold either off_t or ssize_t,
|
133 |
* and must be signed; it's only a guess that long is suitable, so we
|
134 |
* offer <sys/types.h> an override.
|
135 |
*/
|
136 |
#ifdef __REG_REGOFF_T
|
137 |
typedef __REG_REGOFF_T regoff_t;
|
138 |
#else
|
139 |
typedef long regoff_t;
|
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
|
147 |
typedef __REG_VOID_T re_void;
|
148 |
#else
|
149 |
typedef 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 |
|
161 |
|
162 |
/*
|
163 |
* other interface types
|
164 |
*/
|
165 |
|
166 |
/* the biggie, a compiled RE (or rather, a front end to same) */
|
167 |
typedef 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 */
|
187 |
/* the rest is opaque pointers to hidden innards */
|
188 |
char *re_guts; /* `char *' is more portable than `void *' */
|
189 |
char *re_fns;
|
190 |
} regex_t;
|
191 |
|
192 |
/* result reporting (may acquire more fields later) */
|
193 |
typedef struct {
|
194 |
regoff_t rm_so; /* start of substring */
|
195 |
regoff_t rm_eo; /* end of substring */
|
196 |
} regmatch_t;
|
197 |
|
198 |
/* supplementary control and reporting */
|
199 |
typedef struct {
|
200 |
regmatch_t rm_extend; /* see REG_EXPECT */
|
201 |
} rm_detail_t;
|
202 |
|
203 |
|
204 |
|
205 |
/*
|
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
|
251 |
*/
|
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 */
|
258 |
|
259 |
|
260 |
|
261 |
/*
|
262 |
* misc generics (may be more functions here eventually)
|
263 |
^ re_void regfree(regex_t *);
|
264 |
*/
|
265 |
|
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!
|
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);
|
278 |
*/
|
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 */
|
297 |
/* two specials for debugging and testing */
|
298 |
#define REG_ATOI 101 /* convert error-code name to number */
|
299 |
#define REG_ITOA 102 /* convert error-code number to name */
|
300 |
|
301 |
|
302 |
|
303 |
/*
|
304 |
* the prototypes, as possibly munched by regfwd
|
305 |
*/
|
306 |
/* =====^!^===== begin forwards =====^!^===== */
|
307 |
/* automatically gathered by fwd; do not hand-edit */
|
308 |
/* === regproto.h === */
|
309 |
#ifndef __REG_NOCHAR
|
310 |
int re_comp _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, int));
|
311 |
#endif
|
312 |
#ifndef __REG_NOFRONT
|
313 |
int regcomp _ANSI_ARGS_((regex_t *, __REG_CONST char *, int));
|
314 |
#endif
|
315 |
#ifdef __REG_WIDE_T
|
316 |
int __REG_WIDE_COMPILE _ANSI_ARGS_((regex_t *, __REG_CONST __REG_WIDE_T *, size_t, int));
|
317 |
#endif
|
318 |
#ifndef __REG_NOCHAR
|
319 |
int 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
|
322 |
int regexec _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, regmatch_t [], int));
|
323 |
#endif
|
324 |
#ifdef __REG_WIDE_T
|
325 |
int __REG_WIDE_EXEC _ANSI_ARGS_((regex_t *, __REG_CONST __REG_WIDE_T *, size_t, rm_detail_t *, size_t, regmatch_t [], int));
|
326 |
#endif
|
327 |
re_void regfree _ANSI_ARGS_((regex_t *));
|
328 |
extern 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 =====^!^===== */
|
331 |
|
332 |
|
333 |
|
334 |
/*
|
335 |
* more C++ voodoo
|
336 |
*/
|
337 |
#ifdef __cplusplus
|
338 |
}
|
339 |
#endif
|
340 |
|
341 |
|
342 |
|
343 |
#endif
|
344 |
|
345 |
/* End of regex.h */
|