GnuRegex.h
Go to the documentation of this file.
1 /*
2  * Copyright (C) 1996-2017 The Squid Software Foundation and contributors
3  *
4  * Squid software is distributed under GPLv2+ license and includes
5  * contributions from numerous individuals and organizations.
6  * Please see the COPYING and CONTRIBUTORS files for details.
7  */
8 
9 #ifndef SQUID_REGEXP_LIBRARY_H
10 #define SQUID_REGEXP_LIBRARY_H
11 
12 #if !USE_GNUREGEX /* try the system one by default */
13 
14 /* POSIX says that <sys/types.h> must be included (by the caller) before
15  * <regex.h>. */
16 #if HAVE_SYS_TYPES_H
17 #include <sys/types.h>
18 #endif
19 #if HAVE_REGEX_H
20 #include <regex.h>
21 #endif
22 
23 #else /* USE_GNUREGEX */
24 
25 #ifdef __cplusplus
26 extern "C" {
27 #endif
28 
29 /* Definitions for data structures and routines for the regular
30  * expression library, version 0.12.
31  *
32  * Copyright (C) 1985, 1989, 1990, 1991, 1992, 1993 Free Software Foundation, Inc.
33  *
34  * This program is free software; you can redistribute it and/or modify
35  * it under the terms of the GNU General Public License as published by
36  * the Free Software Foundation; either version 2, or (at your option)
37  * any later version.
38  *
39  * This program is distributed in the hope that it will be useful,
40  * but WITHOUT ANY WARRANTY; without even the implied warranty of
41  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
42  * GNU General Public License for more details.
43  *
44  * You should have received a copy of the GNU General Public License
45  * along with this program; if not, write to the Free Software
46  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA. */
47 
48 /* POSIX says that <sys/types.h> must be included (by the caller) before
49  * <regex.h>. */
50 
51 /* The following bits are used to determine the regexp syntax we
52  * recognize. The set/not-set meanings are chosen so that Emacs syntax
53  * remains the value 0. The bits are given in alphabetical order, and
54  * the definitions shifted by one from the previous bit; thus, when we
55  * add or remove a bit, only one other definition need change. */
56 typedef unsigned reg_syntax_t;
57 
58 /* If this bit is not set, then \ inside a bracket expression is literal.
59  * If set, then such a \ quotes the following character. */
60 #define RE_BACKSLASH_ESCAPE_IN_LISTS (1)
61 
62 /* If this bit is not set, then + and ? are operators, and \+ and \? are
63  * literals.
64  * If set, then \+ and \? are operators and + and ? are literals. */
65 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
66 
67 /* If this bit is set, then character classes are supported. They are:
68  * [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
69  * [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
70  * If not set, then character classes are not supported. */
71 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
72 
73 /* If this bit is set, then ^ and $ are always anchors (outside bracket
74  * expressions, of course).
75  * If this bit is not set, then it depends:
76  * ^ is an anchor if it is at the beginning of a regular
77  * expression or after an open-group or an alternation operator;
78  * $ is an anchor if it is at the end of a regular expression, or
79  * before a close-group or an alternation operator.
80  *
81  * This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
82  * POSIX draft 11.2 says that * etc. in leading positions is undefined.
83  * We already implemented a previous draft which made those constructs
84  * invalid, though, so we haven't changed the code back. */
85 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
86 
87 /* If this bit is set, then special characters are always special
88  * regardless of where they are in the pattern.
89  * If this bit is not set, then special characters are special only in
90  * some contexts; otherwise they are ordinary. Specifically,
91  * * + ? and intervals are only special when not after the beginning,
92  * open-group, or alternation operator. */
93 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
94 
95 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
96  * immediately after an alternation or begin-group operator. */
97 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
98 
99 /* If this bit is set, then . matches newline.
100  * If not set, then it doesn't. */
101 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
102 
103 /* If this bit is set, then . doesn't match NUL.
104  * If not set, then it does. */
105 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
106 
107 /* If this bit is set, nonmatching lists [^...] do not match newline.
108  * If not set, they do. */
109 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
110 
111 /* If this bit is set, either \{...\} or {...} defines an
112  * interval, depending on RE_NO_BK_BRACES.
113  * If not set, \{, \}, {, and } are literals. */
114 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
115 
116 /* If this bit is set, +, ? and | aren't recognized as operators.
117  * If not set, they are. */
118 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
119 
120 /* If this bit is set, newline is an alternation operator.
121  * If not set, newline is literal. */
122 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
123 
124 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
125  * are literals.
126  * If not set, then `\{...\}' defines an interval. */
127 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
128 
129 /* If this bit is set, (...) defines a group, and \( and \) are literals.
130  * If not set, \(...\) defines a group, and ( and ) are literals. */
131 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
132 
133 /* If this bit is set, then <digit> matches <digit>.
134  * If not set, then <digit> is a back-reference. */
135 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
136 
137 /* If this bit is set, then | is an alternation operator, and \| is literal.
138  * If not set, then \| is an alternation operator, and | is literal. */
139 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
140 
141 /* If this bit is set, then an ending range point collating higher
142  * than the starting range point, as in [z-a], is invalid.
143  * If not set, then when ending range point collates higher than the
144  * starting range point, the range is ignored. */
145 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
146 
147 /* If this bit is set, then an unmatched ) is ordinary.
148  * If not set, then an unmatched ) is invalid. */
149 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
150 
151 /* Define combinations of the above bits for the standard possibilities.
152  * (The [[[ comments delimit what gets put into the Texinfo file, so
153  * don't delete them!) */
154 /* [[[begin syntaxes]]] */
155 #define RE_SYNTAX_EMACS 0
156 
157 #define RE_SYNTAX_AWK \
158  (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
159  | RE_NO_BK_PARENS | RE_NO_BK_REFS \
160  | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
161  | RE_UNMATCHED_RIGHT_PAREN_ORD)
162 
163 #define RE_SYNTAX_POSIX_AWK \
164  (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS)
165 
166 #define RE_SYNTAX_GREP \
167  (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
168  | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
169  | RE_NEWLINE_ALT)
170 
171 #define RE_SYNTAX_EGREP \
172  (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
173  | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
174  | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
175  | RE_NO_BK_VBAR)
176 
177 #define RE_SYNTAX_POSIX_EGREP \
178  (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
179 
180 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
181 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
182 
183 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
184 
185 /* Syntax bits common to both basic and extended POSIX regex syntax. */
186 #define _RE_SYNTAX_POSIX_COMMON \
187  (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
188  | RE_INTERVALS | RE_NO_EMPTY_RANGES)
189 
190 #define RE_SYNTAX_POSIX_BASIC \
191  (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
192 
193 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
194  * RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
195  * isn't minimal, since other operators, such as \`, aren't disabled. */
196 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \
197  (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
198 
199 #define RE_SYNTAX_POSIX_EXTENDED \
200  (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
201  | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
202  | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
203  | RE_UNMATCHED_RIGHT_PAREN_ORD)
204 
205 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
206  * replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */
207 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
208  (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
209  | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
210  | RE_NO_BK_PARENS | RE_NO_BK_REFS \
211  | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
212 /* [[[end syntaxes]]] */
213 
214 /* Maximum number of duplicates an interval can allow. Some systems
215  * (erroneously) define this in other header files, but we want our
216  * value, so remove any previous define. */
217 #ifdef RE_DUP_MAX
218 #undef RE_DUP_MAX
219 #endif
220 #define RE_DUP_MAX ((1 << 15) - 1)
221 
222 /* POSIX `cflags' bits (i.e., information for `regcomp'). */
223 
224 /* If this bit is set, then use extended regular expression syntax.
225  * If not set, then use basic regular expression syntax. */
226 #define REG_EXTENDED 1
227 
228 /* If this bit is set, then ignore case when matching.
229  * If not set, then case is significant. */
230 #define REG_ICASE (REG_EXTENDED << 1)
231 
232 /* If this bit is set, then anchors do not match at newline
233  * characters in the string.
234  * If not set, then anchors do match at newlines. */
235 #define REG_NEWLINE (REG_ICASE << 1)
236 
237 /* If this bit is set, then report only success or fail in regexec.
238  * If not set, then returns differ between not matching and errors. */
239 #define REG_NOSUB (REG_NEWLINE << 1)
240 
241 /* POSIX `eflags' bits (i.e., information for regexec). */
242 
243 /* If this bit is set, then the beginning-of-line operator doesn't match
244  * the beginning of the string (presumably because it's not the
245  * beginning of a line).
246  * If not set, then the beginning-of-line operator does match the
247  * beginning of the string. */
248 #define REG_NOTBOL 1
249 
250 /* Like REG_NOTBOL, except for the end-of-line. */
251 #define REG_NOTEOL (1 << 1)
252 
253 /* If any error codes are removed, changed, or added, update the
254  * `re_error_msg' table in regex.c. */
255 typedef enum {
256  REG_NOERROR = 0, /* Success. */
257  REG_NOMATCH, /* Didn't find a match (for regexec). */
258 
259  /* POSIX regcomp return error codes. (In the order listed in the
260  * standard.) */
261  REG_BADPAT, /* Invalid pattern. */
262  REG_ECOLLATE, /* Not implemented. */
263  REG_ECTYPE, /* Invalid character class name. */
264  REG_EESCAPE, /* Trailing backslash. */
265  REG_ESUBREG, /* Invalid back reference. */
266  REG_EBRACK, /* Unmatched left bracket. */
267  REG_EPAREN, /* Parenthesis imbalance. */
268  REG_EBRACE, /* Unmatched \{. */
269  REG_BADBR, /* Invalid contents of \{\}. */
270  REG_ERANGE, /* Invalid range end. */
271  REG_ESPACE, /* Ran out of memory. */
272  REG_BADRPT, /* No preceding re for repetition op. */
273 
274  /* Error codes we've added. */
275  REG_EEND, /* Premature end. */
276  REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
277  REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
278 } reg_errcode_t;
279 
280 /* This data structure represents a compiled pattern. Before calling
281  * the pattern compiler, the fields `buffer', `allocated', `fastmap',
282  * `translate', and `no_sub' can be set. After the pattern has been
283  * compiled, the `re_nsub' field is available. All other fields are
284  * private to the regex routines. */
285 
287  /* [[[begin pattern_buffer]]] */
288  /* Space that holds the compiled pattern. It is declared as
289  * `unsigned char *' because its elements are
290  * sometimes used as array indexes. */
291  unsigned char *buffer;
292 
293  /* Number of bytes to which `buffer' points. */
294  unsigned long allocated;
295 
296  /* Number of bytes actually used in `buffer'. */
297  unsigned long used;
298 
299  /* Syntax setting with which the pattern was compiled. */
300  reg_syntax_t syntax;
301 
302  /* Pointer to a fastmap, if any, otherwise zero. re_search uses
303  * the fastmap, if there is one, to skip over impossible
304  * starting points for matches. */
305  char *fastmap;
306 
307  /* Either a translate table to apply to all characters before
308  * comparing them, or zero for no translation. The translation
309  * is applied to a pattern when it is compiled and to a string
310  * when it is matched. */
311  char *translate;
312 
313  /* Number of subexpressions found by the compiler. */
314  size_t re_nsub;
315 
316  /* Zero if this pattern cannot match the empty string, one else.
317  * Well, in truth it's used only in `re_search_2', to see
318  * whether or not we should use the fastmap, so we don't set
319  * this absolutely perfectly; see `re_compile_fastmap' (the
320  * `duplicate' case). */
321  unsigned can_be_null:1;
322 
323  /* If REGS_UNALLOCATED, allocate space in the `regs' structure
324  * for `max (RE_NREGS, re_nsub + 1)' groups.
325  * If REGS_REALLOCATE, reallocate space if necessary.
326  * If REGS_FIXED, use what's there. */
327 #define REGS_UNALLOCATED 0
328 #define REGS_REALLOCATE 1
329 #define REGS_FIXED 2
330  unsigned regs_allocated:2;
331 
332  /* Set to zero when `regex_compile' compiles a pattern; set to one
333  * by `re_compile_fastmap' if it updates the fastmap. */
334  unsigned fastmap_accurate:1;
335 
336  /* If set, `re_match_2' does not return information about
337  * subexpressions. */
338  unsigned no_sub:1;
339 
340  /* If set, a beginning-of-line anchor doesn't match at the
341  * beginning of the string. */
342  unsigned not_bol:1;
343 
344  /* Similarly for an end-of-line anchor. */
345  unsigned not_eol:1;
346 
347  /* If true, an anchor at a newline matches. */
348  unsigned newline_anchor:1;
349 
350  /* [[[end pattern_buffer]]] */
351 };
352 
353 typedef struct re_pattern_buffer regex_t;
354 
355 /* search.c (search_buffer) in Emacs needs this one opcode value. It is
356  * defined both in `regex.c' and here. */
357 #define RE_EXACTN_VALUE 1
358 
359 /* Type for byte offsets within the string. POSIX mandates this. */
360 typedef int regoff_t;
361 
362 /* This is the structure we store register match data in. See
363  * regex.texinfo for a full description of what registers match. */
364 struct re_registers {
365  unsigned num_regs;
366  regoff_t *start;
367  regoff_t *end;
368 };
369 
370 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
371  * `re_match_2' returns information about at least this many registers
372  * the first time a `regs' structure is passed. */
373 #ifndef RE_NREGS
374 #define RE_NREGS 30
375 #endif
376 
377 /* POSIX specification for registers. Aside from the different names than
378  * `re_registers', POSIX uses an array of structures, instead of a
379  * structure of arrays. */
380 typedef struct {
381  regoff_t rm_so; /* Byte offset from string's start to substring's start. */
382  regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
383 } regmatch_t;
384 
385 /* Declarations for routines. */
386 
387 /* To avoid duplicating every routine declaration -- once with a
388  * prototype (if we are ANSI), and once without (if we aren't) -- we
389  * use the following macro to declare argument types. This
390  * unfortunately clutters up the declarations a bit, but I think it's
391  * worth it. */
392 
393 /* POSIX compatibility. */
394 extern int regcomp(regex_t * preg, const char *pattern, int cflags);
395 extern int regexec(const regex_t * preg, const char *string, size_t nmatch, regmatch_t pmatch[], int eflags);
396 extern size_t regerror(int errcode, const regex_t * preg, char *errbuf, size_t errbuf_size);
397 extern void regfree(regex_t * preg);
398 
399 #ifdef __cplusplus
400 }
401 #endif
402 
403 #endif /* USE_GNUREGEX */
404 #endif /* SQUID_REGEXP_LIBRARY_H */
405 
406 /*
407  * Local variables:
408  * make-backup-files: t
409  * version-control: t
410  * trim-versions-without-asking: nil
411  * End:
412  */
413 
regoff_t * start
Definition: GnuRegex.h:366
unsigned long used
Definition: GnuRegex.h:297
regoff_t * end
Definition: GnuRegex.h:367
void regfree(regex_t *preg)
Definition: GnuRegex.c:4283
reg_errcode_t
Definition: GnuRegex.h:255
regoff_t rm_eo
Definition: GnuRegex.h:382
unsigned not_bol
Definition: GnuRegex.h:342
unsigned can_be_null
Definition: GnuRegex.h:321
unsigned newline_anchor
Definition: GnuRegex.h:348
unsigned long allocated
Definition: GnuRegex.h:294
int regoff_t
Definition: GnuRegex.h:360
unsigned not_eol
Definition: GnuRegex.h:345
unsigned char * buffer
Definition: GnuRegex.h:291
unsigned regs_allocated
Definition: GnuRegex.h:330
int regexec(const regex_t *preg, const char *string, size_t nmatch, regmatch_t pmatch[], int eflags)
unsigned fastmap_accurate
Definition: GnuRegex.h:334
size_t regerror(int errcode, const regex_t *preg, char *errbuf, size_t errbuf_size)
Definition: GnuRegex.c:4248
unsigned num_regs
Definition: GnuRegex.h:365
regoff_t rm_so
Definition: GnuRegex.h:381
reg_syntax_t syntax
Definition: GnuRegex.h:300
unsigned no_sub
Definition: GnuRegex.h:338
char * translate
Definition: GnuRegex.h:311
unsigned reg_syntax_t
Definition: GnuRegex.h:56
int regcomp(regex_t *preg, const char *pattern, int cflags)
Definition: GnuRegex.c:4120

 

Introduction

Documentation

Support

Miscellaneous

Web Site Translations

Mirrors