/[pcre]/code/trunk/pcre_compile.c
ViewVC logotype

Contents of /code/trunk/pcre_compile.c

Parent Directory Parent Directory | Revision Log Revision Log


Revision 550 - (show annotations) (download)
Sun Oct 10 16:24:11 2010 UTC (4 years ago) by ph10
File MIME type: text/plain
File size: 233481 byte(s)
Fix problem with (*THEN) not backing up far enough.

1 /*************************************************
2 * Perl-Compatible Regular Expressions *
3 *************************************************/
4
5 /* PCRE is a library of functions to support regular expressions whose syntax
6 and semantics are as close as possible to those of the Perl 5 language.
7
8 Written by Philip Hazel
9 Copyright (c) 1997-2010 University of Cambridge
10
11 -----------------------------------------------------------------------------
12 Redistribution and use in source and binary forms, with or without
13 modification, are permitted provided that the following conditions are met:
14
15 * Redistributions of source code must retain the above copyright notice,
16 this list of conditions and the following disclaimer.
17
18 * Redistributions in binary form must reproduce the above copyright
19 notice, this list of conditions and the following disclaimer in the
20 documentation and/or other materials provided with the distribution.
21
22 * Neither the name of the University of Cambridge nor the names of its
23 contributors may be used to endorse or promote products derived from
24 this software without specific prior written permission.
25
26 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
27 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
30 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
31 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
32 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
33 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
34 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
35 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
36 POSSIBILITY OF SUCH DAMAGE.
37 -----------------------------------------------------------------------------
38 */
39
40
41 /* This module contains the external function pcre_compile(), along with
42 supporting internal functions that are not used by other modules. */
43
44
45 #ifdef HAVE_CONFIG_H
46 #include "config.h"
47 #endif
48
49 #define NLBLOCK cd /* Block containing newline information */
50 #define PSSTART start_pattern /* Field containing processed string start */
51 #define PSEND end_pattern /* Field containing processed string end */
52
53 #include "pcre_internal.h"
54
55
56 /* When PCRE_DEBUG is defined, we need the pcre_printint() function, which is
57 also used by pcretest. PCRE_DEBUG is not defined when building a production
58 library. */
59
60 #ifdef PCRE_DEBUG
61 #include "pcre_printint.src"
62 #endif
63
64
65 /* Macro for setting individual bits in class bitmaps. */
66
67 #define SETBIT(a,b) a[b/8] |= (1 << (b%8))
68
69 /* Maximum length value to check against when making sure that the integer that
70 holds the compiled pattern length does not overflow. We make it a bit less than
71 INT_MAX to allow for adding in group terminating bytes, so that we don't have
72 to check them every time. */
73
74 #define OFLOW_MAX (INT_MAX - 20)
75
76
77 /*************************************************
78 * Code parameters and static tables *
79 *************************************************/
80
81 /* This value specifies the size of stack workspace that is used during the
82 first pre-compile phase that determines how much memory is required. The regex
83 is partly compiled into this space, but the compiled parts are discarded as
84 soon as they can be, so that hopefully there will never be an overrun. The code
85 does, however, check for an overrun. The largest amount I've seen used is 218,
86 so this number is very generous.
87
88 The same workspace is used during the second, actual compile phase for
89 remembering forward references to groups so that they can be filled in at the
90 end. Each entry in this list occupies LINK_SIZE bytes, so even when LINK_SIZE
91 is 4 there is plenty of room. */
92
93 #define COMPILE_WORK_SIZE (4096)
94
95 /* The overrun tests check for a slightly smaller size so that they detect the
96 overrun before it actually does run off the end of the data block. */
97
98 #define WORK_SIZE_CHECK (COMPILE_WORK_SIZE - 100)
99
100
101 /* Table for handling escaped characters in the range '0'-'z'. Positive returns
102 are simple data values; negative values are for special things like \d and so
103 on. Zero means further processing is needed (for things like \x), or the escape
104 is invalid. */
105
106 #ifndef EBCDIC
107
108 /* This is the "normal" table for ASCII systems or for EBCDIC systems running
109 in UTF-8 mode. */
110
111 static const short int escapes[] = {
112 0, 0,
113 0, 0,
114 0, 0,
115 0, 0,
116 0, 0,
117 CHAR_COLON, CHAR_SEMICOLON,
118 CHAR_LESS_THAN_SIGN, CHAR_EQUALS_SIGN,
119 CHAR_GREATER_THAN_SIGN, CHAR_QUESTION_MARK,
120 CHAR_COMMERCIAL_AT, -ESC_A,
121 -ESC_B, -ESC_C,
122 -ESC_D, -ESC_E,
123 0, -ESC_G,
124 -ESC_H, 0,
125 0, -ESC_K,
126 0, 0,
127 -ESC_N, 0,
128 -ESC_P, -ESC_Q,
129 -ESC_R, -ESC_S,
130 0, 0,
131 -ESC_V, -ESC_W,
132 -ESC_X, 0,
133 -ESC_Z, CHAR_LEFT_SQUARE_BRACKET,
134 CHAR_BACKSLASH, CHAR_RIGHT_SQUARE_BRACKET,
135 CHAR_CIRCUMFLEX_ACCENT, CHAR_UNDERSCORE,
136 CHAR_GRAVE_ACCENT, 7,
137 -ESC_b, 0,
138 -ESC_d, ESC_e,
139 ESC_f, 0,
140 -ESC_h, 0,
141 0, -ESC_k,
142 0, 0,
143 ESC_n, 0,
144 -ESC_p, 0,
145 ESC_r, -ESC_s,
146 ESC_tee, 0,
147 -ESC_v, -ESC_w,
148 0, 0,
149 -ESC_z
150 };
151
152 #else
153
154 /* This is the "abnormal" table for EBCDIC systems without UTF-8 support. */
155
156 static const short int escapes[] = {
157 /* 48 */ 0, 0, 0, '.', '<', '(', '+', '|',
158 /* 50 */ '&', 0, 0, 0, 0, 0, 0, 0,
159 /* 58 */ 0, 0, '!', '$', '*', ')', ';', '~',
160 /* 60 */ '-', '/', 0, 0, 0, 0, 0, 0,
161 /* 68 */ 0, 0, '|', ',', '%', '_', '>', '?',
162 /* 70 */ 0, 0, 0, 0, 0, 0, 0, 0,
163 /* 78 */ 0, '`', ':', '#', '@', '\'', '=', '"',
164 /* 80 */ 0, 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0,
165 /* 88 */-ESC_h, 0, 0, '{', 0, 0, 0, 0,
166 /* 90 */ 0, 0, -ESC_k, 'l', 0, ESC_n, 0, -ESC_p,
167 /* 98 */ 0, ESC_r, 0, '}', 0, 0, 0, 0,
168 /* A0 */ 0, '~', -ESC_s, ESC_tee, 0,-ESC_v, -ESC_w, 0,
169 /* A8 */ 0,-ESC_z, 0, 0, 0, '[', 0, 0,
170 /* B0 */ 0, 0, 0, 0, 0, 0, 0, 0,
171 /* B8 */ 0, 0, 0, 0, 0, ']', '=', '-',
172 /* C0 */ '{',-ESC_A, -ESC_B, -ESC_C, -ESC_D,-ESC_E, 0, -ESC_G,
173 /* C8 */-ESC_H, 0, 0, 0, 0, 0, 0, 0,
174 /* D0 */ '}', 0, -ESC_K, 0, 0,-ESC_N, 0, -ESC_P,
175 /* D8 */-ESC_Q,-ESC_R, 0, 0, 0, 0, 0, 0,
176 /* E0 */ '\\', 0, -ESC_S, 0, 0,-ESC_V, -ESC_W, -ESC_X,
177 /* E8 */ 0,-ESC_Z, 0, 0, 0, 0, 0, 0,
178 /* F0 */ 0, 0, 0, 0, 0, 0, 0, 0,
179 /* F8 */ 0, 0, 0, 0, 0, 0, 0, 0
180 };
181 #endif
182
183
184 /* Table of special "verbs" like (*PRUNE). This is a short table, so it is
185 searched linearly. Put all the names into a single string, in order to reduce
186 the number of relocations when a shared library is dynamically linked. The
187 string is built from string macros so that it works in UTF-8 mode on EBCDIC
188 platforms. */
189
190 typedef struct verbitem {
191 int len; /* Length of verb name */
192 int op; /* Op when no arg, or -1 if arg mandatory */
193 int op_arg; /* Op when arg present, or -1 if not allowed */
194 } verbitem;
195
196 static const char verbnames[] =
197 "\0" /* Empty name is a shorthand for MARK */
198 STRING_MARK0
199 STRING_ACCEPT0
200 STRING_COMMIT0
201 STRING_F0
202 STRING_FAIL0
203 STRING_PRUNE0
204 STRING_SKIP0
205 STRING_THEN;
206
207 static const verbitem verbs[] = {
208 { 0, -1, OP_MARK },
209 { 4, -1, OP_MARK },
210 { 6, OP_ACCEPT, -1 },
211 { 6, OP_COMMIT, -1 },
212 { 1, OP_FAIL, -1 },
213 { 4, OP_FAIL, -1 },
214 { 5, OP_PRUNE, OP_PRUNE_ARG },
215 { 4, OP_SKIP, OP_SKIP_ARG },
216 { 4, OP_THEN, OP_THEN_ARG }
217 };
218
219 static const int verbcount = sizeof(verbs)/sizeof(verbitem);
220
221
222 /* Tables of names of POSIX character classes and their lengths. The names are
223 now all in a single string, to reduce the number of relocations when a shared
224 library is dynamically loaded. The list of lengths is terminated by a zero
225 length entry. The first three must be alpha, lower, upper, as this is assumed
226 for handling case independence. */
227
228 static const char posix_names[] =
229 STRING_alpha0 STRING_lower0 STRING_upper0 STRING_alnum0
230 STRING_ascii0 STRING_blank0 STRING_cntrl0 STRING_digit0
231 STRING_graph0 STRING_print0 STRING_punct0 STRING_space0
232 STRING_word0 STRING_xdigit;
233
234 static const uschar posix_name_lengths[] = {
235 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 };
236
237 /* Table of class bit maps for each POSIX class. Each class is formed from a
238 base map, with an optional addition or removal of another map. Then, for some
239 classes, there is some additional tweaking: for [:blank:] the vertical space
240 characters are removed, and for [:alpha:] and [:alnum:] the underscore
241 character is removed. The triples in the table consist of the base map offset,
242 second map offset or -1 if no second map, and a non-negative value for map
243 addition or a negative value for map subtraction (if there are two maps). The
244 absolute value of the third field has these meanings: 0 => no tweaking, 1 =>
245 remove vertical space characters, 2 => remove underscore. */
246
247 static const int posix_class_maps[] = {
248 cbit_word, cbit_digit, -2, /* alpha */
249 cbit_lower, -1, 0, /* lower */
250 cbit_upper, -1, 0, /* upper */
251 cbit_word, -1, 2, /* alnum - word without underscore */
252 cbit_print, cbit_cntrl, 0, /* ascii */
253 cbit_space, -1, 1, /* blank - a GNU extension */
254 cbit_cntrl, -1, 0, /* cntrl */
255 cbit_digit, -1, 0, /* digit */
256 cbit_graph, -1, 0, /* graph */
257 cbit_print, -1, 0, /* print */
258 cbit_punct, -1, 0, /* punct */
259 cbit_space, -1, 0, /* space */
260 cbit_word, -1, 0, /* word - a Perl extension */
261 cbit_xdigit,-1, 0 /* xdigit */
262 };
263
264 /* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class
265 substitutes must be in the order of the names, defined above, and there are
266 both positive and negative cases. NULL means no substitute. */
267
268 #ifdef SUPPORT_UCP
269 static const uschar *substitutes[] = {
270 (uschar *)"\\P{Nd}", /* \D */
271 (uschar *)"\\p{Nd}", /* \d */
272 (uschar *)"\\P{Xsp}", /* \S */ /* NOTE: Xsp is Perl space */
273 (uschar *)"\\p{Xsp}", /* \s */
274 (uschar *)"\\P{Xwd}", /* \W */
275 (uschar *)"\\p{Xwd}" /* \w */
276 };
277
278 static const uschar *posix_substitutes[] = {
279 (uschar *)"\\p{L}", /* alpha */
280 (uschar *)"\\p{Ll}", /* lower */
281 (uschar *)"\\p{Lu}", /* upper */
282 (uschar *)"\\p{Xan}", /* alnum */
283 NULL, /* ascii */
284 (uschar *)"\\h", /* blank */
285 NULL, /* cntrl */
286 (uschar *)"\\p{Nd}", /* digit */
287 NULL, /* graph */
288 NULL, /* print */
289 NULL, /* punct */
290 (uschar *)"\\p{Xps}", /* space */ /* NOTE: Xps is POSIX space */
291 (uschar *)"\\p{Xwd}", /* word */
292 NULL, /* xdigit */
293 /* Negated cases */
294 (uschar *)"\\P{L}", /* ^alpha */
295 (uschar *)"\\P{Ll}", /* ^lower */
296 (uschar *)"\\P{Lu}", /* ^upper */
297 (uschar *)"\\P{Xan}", /* ^alnum */
298 NULL, /* ^ascii */
299 (uschar *)"\\H", /* ^blank */
300 NULL, /* ^cntrl */
301 (uschar *)"\\P{Nd}", /* ^digit */
302 NULL, /* ^graph */
303 NULL, /* ^print */
304 NULL, /* ^punct */
305 (uschar *)"\\P{Xps}", /* ^space */ /* NOTE: Xps is POSIX space */
306 (uschar *)"\\P{Xwd}", /* ^word */
307 NULL /* ^xdigit */
308 };
309 #define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *))
310 #endif
311
312 #define STRING(a) # a
313 #define XSTRING(s) STRING(s)
314
315 /* The texts of compile-time error messages. These are "char *" because they
316 are passed to the outside world. Do not ever re-use any error number, because
317 they are documented. Always add a new error instead. Messages marked DEAD below
318 are no longer used. This used to be a table of strings, but in order to reduce
319 the number of relocations needed when a shared library is loaded dynamically,
320 it is now one long string. We cannot use a table of offsets, because the
321 lengths of inserts such as XSTRING(MAX_NAME_SIZE) are not known. Instead, we
322 simply count through to the one we want - this isn't a performance issue
323 because these strings are used only when there is a compilation error.
324
325 Each substring ends with \0 to insert a null character. This includes the final
326 substring, so that the whole string ends with \0\0, which can be detected when
327 counting through. */
328
329 static const char error_texts[] =
330 "no error\0"
331 "\\ at end of pattern\0"
332 "\\c at end of pattern\0"
333 "unrecognized character follows \\\0"
334 "numbers out of order in {} quantifier\0"
335 /* 5 */
336 "number too big in {} quantifier\0"
337 "missing terminating ] for character class\0"
338 "invalid escape sequence in character class\0"
339 "range out of order in character class\0"
340 "nothing to repeat\0"
341 /* 10 */
342 "operand of unlimited repeat could match the empty string\0" /** DEAD **/
343 "internal error: unexpected repeat\0"
344 "unrecognized character after (? or (?-\0"
345 "POSIX named classes are supported only within a class\0"
346 "missing )\0"
347 /* 15 */
348 "reference to non-existent subpattern\0"
349 "erroffset passed as NULL\0"
350 "unknown option bit(s) set\0"
351 "missing ) after comment\0"
352 "parentheses nested too deeply\0" /** DEAD **/
353 /* 20 */
354 "regular expression is too large\0"
355 "failed to get memory\0"
356 "unmatched parentheses\0"
357 "internal error: code overflow\0"
358 "unrecognized character after (?<\0"
359 /* 25 */
360 "lookbehind assertion is not fixed length\0"
361 "malformed number or name after (?(\0"
362 "conditional group contains more than two branches\0"
363 "assertion expected after (?(\0"
364 "(?R or (?[+-]digits must be followed by )\0"
365 /* 30 */
366 "unknown POSIX class name\0"
367 "POSIX collating elements are not supported\0"
368 "this version of PCRE is not compiled with PCRE_UTF8 support\0"
369 "spare error\0" /** DEAD **/
370 "character value in \\x{...} sequence is too large\0"
371 /* 35 */
372 "invalid condition (?(0)\0"
373 "\\C not allowed in lookbehind assertion\0"
374 "PCRE does not support \\L, \\l, \\N{name}, \\U, or \\u\0"
375 "number after (?C is > 255\0"
376 "closing ) for (?C expected\0"
377 /* 40 */
378 "recursive call could loop indefinitely\0"
379 "unrecognized character after (?P\0"
380 "syntax error in subpattern name (missing terminator)\0"
381 "two named subpatterns have the same name\0"
382 "invalid UTF-8 string\0"
383 /* 45 */
384 "support for \\P, \\p, and \\X has not been compiled\0"
385 "malformed \\P or \\p sequence\0"
386 "unknown property name after \\P or \\p\0"
387 "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)\0"
388 "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")\0"
389 /* 50 */
390 "repeated subpattern is too long\0" /** DEAD **/
391 "octal value is greater than \\377 (not in UTF-8 mode)\0"
392 "internal error: overran compiling workspace\0"
393 "internal error: previously-checked referenced subpattern not found\0"
394 "DEFINE group contains more than one branch\0"
395 /* 55 */
396 "repeating a DEFINE group is not allowed\0"
397 "inconsistent NEWLINE options\0"
398 "\\g is not followed by a braced, angle-bracketed, or quoted name/number or by a plain number\0"
399 "a numbered reference must not be zero\0"
400 "an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)\0"
401 /* 60 */
402 "(*VERB) not recognized\0"
403 "number is too big\0"
404 "subpattern name expected\0"
405 "digit expected after (?+\0"
406 "] is an invalid data character in JavaScript compatibility mode\0"
407 /* 65 */
408 "different names for subpatterns of the same number are not allowed\0"
409 "(*MARK) must have an argument\0"
410 "this version of PCRE is not compiled with PCRE_UCP support\0"
411 ;
412
413 /* Table to identify digits and hex digits. This is used when compiling
414 patterns. Note that the tables in chartables are dependent on the locale, and
415 may mark arbitrary characters as digits - but the PCRE compiling code expects
416 to handle only 0-9, a-z, and A-Z as digits when compiling. That is why we have
417 a private table here. It costs 256 bytes, but it is a lot faster than doing
418 character value tests (at least in some simple cases I timed), and in some
419 applications one wants PCRE to compile efficiently as well as match
420 efficiently.
421
422 For convenience, we use the same bit definitions as in chartables:
423
424 0x04 decimal digit
425 0x08 hexadecimal digit
426
427 Then we can use ctype_digit and ctype_xdigit in the code. */
428
429 #ifndef EBCDIC
430
431 /* This is the "normal" case, for ASCII systems, and EBCDIC systems running in
432 UTF-8 mode. */
433
434 static const unsigned char digitab[] =
435 {
436 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */
437 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */
438 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 */
439 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */
440 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - ' */
441 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ( - / */
442 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 */
443 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00, /* 8 - ? */
444 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* @ - G */
445 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H - O */
446 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* P - W */
447 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* X - _ */
448 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* ` - g */
449 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h - o */
450 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* p - w */
451 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* x -127 */
452 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
453 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
454 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
455 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
456 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
457 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
458 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
459 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
460 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
461 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
462 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
463 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
464 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
465 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
466 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
467 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
468
469 #else
470
471 /* This is the "abnormal" case, for EBCDIC systems not running in UTF-8 mode. */
472
473 static const unsigned char digitab[] =
474 {
475 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 0 */
476 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */
477 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 10 */
478 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */
479 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 32- 39 20 */
480 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */
481 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 30 */
482 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */
483 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 40 */
484 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 72- | */
485 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 50 */
486 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 88- 95 */
487 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 60 */
488 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 104- ? */
489 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 70 */
490 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */
491 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* 128- g 80 */
492 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */
493 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144- p 90 */
494 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */
495 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160- x A0 */
496 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */
497 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 B0 */
498 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
499 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* { - G C0 */
500 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */
501 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* } - P D0 */
502 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */
503 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* \ - X E0 */
504 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */
505 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 F0 */
506 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */
507
508 static const unsigned char ebcdic_chartab[] = { /* chartable partial dup */
509 0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 0- 7 */
510 0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /* 8- 15 */
511 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 16- 23 */
512 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */
513 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 32- 39 */
514 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */
515 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 */
516 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */
517 0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 */
518 0x00,0x00,0x00,0x80,0x00,0x80,0x80,0x80, /* 72- | */
519 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 */
520 0x00,0x00,0x00,0x80,0x80,0x80,0x00,0x00, /* 88- 95 */
521 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 */
522 0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x80, /* 104- ? */
523 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 */
524 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */
525 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* 128- g */
526 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */
527 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* 144- p */
528 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */
529 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* 160- x */
530 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */
531 0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 */
532 0x00,0x00,0x80,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
533 0x80,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* { - G */
534 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */
535 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* } - P */
536 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */
537 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* \ - X */
538 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */
539 0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /* 0 - 7 */
540 0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */
541 #endif
542
543
544 /* Definition to allow mutual recursion */
545
546 static BOOL
547 compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,
548 int *, int *, branch_chain *, compile_data *, int *);
549
550
551
552 /*************************************************
553 * Find an error text *
554 *************************************************/
555
556 /* The error texts are now all in one long string, to save on relocations. As
557 some of the text is of unknown length, we can't use a table of offsets.
558 Instead, just count through the strings. This is not a performance issue
559 because it happens only when there has been a compilation error.
560
561 Argument: the error number
562 Returns: pointer to the error string
563 */
564
565 static const char *
566 find_error_text(int n)
567 {
568 const char *s = error_texts;
569 for (; n > 0; n--)
570 {
571 while (*s++ != 0) {};
572 if (*s == 0) return "Error text not found (please report)";
573 }
574 return s;
575 }
576
577
578 /*************************************************
579 * Handle escapes *
580 *************************************************/
581
582 /* This function is called when a \ has been encountered. It either returns a
583 positive value for a simple escape such as \n, or a negative value which
584 encodes one of the more complicated things such as \d. A backreference to group
585 n is returned as -(ESC_REF + n); ESC_REF is the highest ESC_xxx macro. When
586 UTF-8 is enabled, a positive value greater than 255 may be returned. On entry,
587 ptr is pointing at the \. On exit, it is on the final character of the escape
588 sequence.
589
590 Arguments:
591 ptrptr points to the pattern position pointer
592 errorcodeptr points to the errorcode variable
593 bracount number of previous extracting brackets
594 options the options bits
595 isclass TRUE if inside a character class
596
597 Returns: zero or positive => a data character
598 negative => a special escape sequence
599 on error, errorcodeptr is set
600 */
601
602 static int
603 check_escape(const uschar **ptrptr, int *errorcodeptr, int bracount,
604 int options, BOOL isclass)
605 {
606 BOOL utf8 = (options & PCRE_UTF8) != 0;
607 const uschar *ptr = *ptrptr + 1;
608 int c, i;
609
610 GETCHARINCTEST(c, ptr); /* Get character value, increment pointer */
611 ptr--; /* Set pointer back to the last byte */
612
613 /* If backslash is at the end of the pattern, it's an error. */
614
615 if (c == 0) *errorcodeptr = ERR1;
616
617 /* Non-alphanumerics are literals. For digits or letters, do an initial lookup
618 in a table. A non-zero result is something that can be returned immediately.
619 Otherwise further processing may be required. */
620
621 #ifndef EBCDIC /* ASCII/UTF-8 coding */
622 else if (c < CHAR_0 || c > CHAR_z) {} /* Not alphanumeric */
623 else if ((i = escapes[c - CHAR_0]) != 0) c = i;
624
625 #else /* EBCDIC coding */
626 else if (c < 'a' || (ebcdic_chartab[c] & 0x0E) == 0) {} /* Not alphanumeric */
627 else if ((i = escapes[c - 0x48]) != 0) c = i;
628 #endif
629
630 /* Escapes that need further processing, or are illegal. */
631
632 else
633 {
634 const uschar *oldptr;
635 BOOL braced, negated;
636
637 switch (c)
638 {
639 /* A number of Perl escapes are not handled by PCRE. We give an explicit
640 error. */
641
642 case CHAR_l:
643 case CHAR_L:
644 case CHAR_u:
645 case CHAR_U:
646 *errorcodeptr = ERR37;
647 break;
648
649 /* \g must be followed by one of a number of specific things:
650
651 (1) A number, either plain or braced. If positive, it is an absolute
652 backreference. If negative, it is a relative backreference. This is a Perl
653 5.10 feature.
654
655 (2) Perl 5.10 also supports \g{name} as a reference to a named group. This
656 is part of Perl's movement towards a unified syntax for back references. As
657 this is synonymous with \k{name}, we fudge it up by pretending it really
658 was \k.
659
660 (3) For Oniguruma compatibility we also support \g followed by a name or a
661 number either in angle brackets or in single quotes. However, these are
662 (possibly recursive) subroutine calls, _not_ backreferences. Just return
663 the -ESC_g code (cf \k). */
664
665 case CHAR_g:
666 if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)
667 {
668 c = -ESC_g;
669 break;
670 }
671
672 /* Handle the Perl-compatible cases */
673
674 if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
675 {
676 const uschar *p;
677 for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++)
678 if (*p != CHAR_MINUS && (digitab[*p] & ctype_digit) == 0) break;
679 if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)
680 {
681 c = -ESC_k;
682 break;
683 }
684 braced = TRUE;
685 ptr++;
686 }
687 else braced = FALSE;
688
689 if (ptr[1] == CHAR_MINUS)
690 {
691 negated = TRUE;
692 ptr++;
693 }
694 else negated = FALSE;
695
696 c = 0;
697 while ((digitab[ptr[1]] & ctype_digit) != 0)
698 c = c * 10 + *(++ptr) - CHAR_0;
699
700 if (c < 0) /* Integer overflow */
701 {
702 *errorcodeptr = ERR61;
703 break;
704 }
705
706 if (braced && *(++ptr) != CHAR_RIGHT_CURLY_BRACKET)
707 {
708 *errorcodeptr = ERR57;
709 break;
710 }
711
712 if (c == 0)
713 {
714 *errorcodeptr = ERR58;
715 break;
716 }
717
718 if (negated)
719 {
720 if (c > bracount)
721 {
722 *errorcodeptr = ERR15;
723 break;
724 }
725 c = bracount - (c - 1);
726 }
727
728 c = -(ESC_REF + c);
729 break;
730
731 /* The handling of escape sequences consisting of a string of digits
732 starting with one that is not zero is not straightforward. By experiment,
733 the way Perl works seems to be as follows:
734
735 Outside a character class, the digits are read as a decimal number. If the
736 number is less than 10, or if there are that many previous extracting
737 left brackets, then it is a back reference. Otherwise, up to three octal
738 digits are read to form an escaped byte. Thus \123 is likely to be octal
739 123 (cf \0123, which is octal 012 followed by the literal 3). If the octal
740 value is greater than 377, the least significant 8 bits are taken. Inside a
741 character class, \ followed by a digit is always an octal number. */
742
743 case CHAR_1: case CHAR_2: case CHAR_3: case CHAR_4: case CHAR_5:
744 case CHAR_6: case CHAR_7: case CHAR_8: case CHAR_9:
745
746 if (!isclass)
747 {
748 oldptr = ptr;
749 c -= CHAR_0;
750 while ((digitab[ptr[1]] & ctype_digit) != 0)
751 c = c * 10 + *(++ptr) - CHAR_0;
752 if (c < 0) /* Integer overflow */
753 {
754 *errorcodeptr = ERR61;
755 break;
756 }
757 if (c < 10 || c <= bracount)
758 {
759 c = -(ESC_REF + c);
760 break;
761 }
762 ptr = oldptr; /* Put the pointer back and fall through */
763 }
764
765 /* Handle an octal number following \. If the first digit is 8 or 9, Perl
766 generates a binary zero byte and treats the digit as a following literal.
767 Thus we have to pull back the pointer by one. */
768
769 if ((c = *ptr) >= CHAR_8)
770 {
771 ptr--;
772 c = 0;
773 break;
774 }
775
776 /* \0 always starts an octal number, but we may drop through to here with a
777 larger first octal digit. The original code used just to take the least
778 significant 8 bits of octal numbers (I think this is what early Perls used
779 to do). Nowadays we allow for larger numbers in UTF-8 mode, but no more
780 than 3 octal digits. */
781
782 case CHAR_0:
783 c -= CHAR_0;
784 while(i++ < 2 && ptr[1] >= CHAR_0 && ptr[1] <= CHAR_7)
785 c = c * 8 + *(++ptr) - CHAR_0;
786 if (!utf8 && c > 255) *errorcodeptr = ERR51;
787 break;
788
789 /* \x is complicated. \x{ddd} is a character number which can be greater
790 than 0xff in utf8 mode, but only if the ddd are hex digits. If not, { is
791 treated as a data character. */
792
793 case CHAR_x:
794 if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
795 {
796 const uschar *pt = ptr + 2;
797 int count = 0;
798
799 c = 0;
800 while ((digitab[*pt] & ctype_xdigit) != 0)
801 {
802 register int cc = *pt++;
803 if (c == 0 && cc == CHAR_0) continue; /* Leading zeroes */
804 count++;
805
806 #ifndef EBCDIC /* ASCII/UTF-8 coding */
807 if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */
808 c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
809 #else /* EBCDIC coding */
810 if (cc >= CHAR_a && cc <= CHAR_z) cc += 64; /* Convert to upper case */
811 c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
812 #endif
813 }
814
815 if (*pt == CHAR_RIGHT_CURLY_BRACKET)
816 {
817 if (c < 0 || count > (utf8? 8 : 2)) *errorcodeptr = ERR34;
818 ptr = pt;
819 break;
820 }
821
822 /* If the sequence of hex digits does not end with '}', then we don't
823 recognize this construct; fall through to the normal \x handling. */
824 }
825
826 /* Read just a single-byte hex-defined char */
827
828 c = 0;
829 while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0)
830 {
831 int cc; /* Some compilers don't like */
832 cc = *(++ptr); /* ++ in initializers */
833 #ifndef EBCDIC /* ASCII/UTF-8 coding */
834 if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */
835 c = c * 16 + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
836 #else /* EBCDIC coding */
837 if (cc <= CHAR_z) cc += 64; /* Convert to upper case */
838 c = c * 16 + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
839 #endif
840 }
841 break;
842
843 /* For \c, a following letter is upper-cased; then the 0x40 bit is flipped.
844 This coding is ASCII-specific, but then the whole concept of \cx is
845 ASCII-specific. (However, an EBCDIC equivalent has now been added.) */
846
847 case CHAR_c:
848 c = *(++ptr);
849 if (c == 0)
850 {
851 *errorcodeptr = ERR2;
852 break;
853 }
854
855 #ifndef EBCDIC /* ASCII/UTF-8 coding */
856 if (c >= CHAR_a && c <= CHAR_z) c -= 32;
857 c ^= 0x40;
858 #else /* EBCDIC coding */
859 if (c >= CHAR_a && c <= CHAR_z) c += 64;
860 c ^= 0xC0;
861 #endif
862 break;
863
864 /* PCRE_EXTRA enables extensions to Perl in the matter of escapes. Any
865 other alphanumeric following \ is an error if PCRE_EXTRA was set;
866 otherwise, for Perl compatibility, it is a literal. This code looks a bit
867 odd, but there used to be some cases other than the default, and there may
868 be again in future, so I haven't "optimized" it. */
869
870 default:
871 if ((options & PCRE_EXTRA) != 0) switch(c)
872 {
873 default:
874 *errorcodeptr = ERR3;
875 break;
876 }
877 break;
878 }
879 }
880
881 /* Perl supports \N{name} for character names, as well as plain \N for "not
882 newline". PCRE does not support \N{name}. */
883
884 if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET)
885 *errorcodeptr = ERR37;
886
887 /* If PCRE_UCP is set, we change the values for \d etc. */
888
889 if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w)
890 c -= (ESC_DU - ESC_D);
891
892 /* Set the pointer to the final character before returning. */
893
894 *ptrptr = ptr;
895 return c;
896 }
897
898
899
900 #ifdef SUPPORT_UCP
901 /*************************************************
902 * Handle \P and \p *
903 *************************************************/
904
905 /* This function is called after \P or \p has been encountered, provided that
906 PCRE is compiled with support for Unicode properties. On entry, ptrptr is
907 pointing at the P or p. On exit, it is pointing at the final character of the
908 escape sequence.
909
910 Argument:
911 ptrptr points to the pattern position pointer
912 negptr points to a boolean that is set TRUE for negation else FALSE
913 dptr points to an int that is set to the detailed property value
914 errorcodeptr points to the error code variable
915
916 Returns: type value from ucp_type_table, or -1 for an invalid type
917 */
918
919 static int
920 get_ucp(const uschar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr)
921 {
922 int c, i, bot, top;
923 const uschar *ptr = *ptrptr;
924 char name[32];
925
926 c = *(++ptr);
927 if (c == 0) goto ERROR_RETURN;
928
929 *negptr = FALSE;
930
931 /* \P or \p can be followed by a name in {}, optionally preceded by ^ for
932 negation. */
933
934 if (c == CHAR_LEFT_CURLY_BRACKET)
935 {
936 if (ptr[1] == CHAR_CIRCUMFLEX_ACCENT)
937 {
938 *negptr = TRUE;
939 ptr++;
940 }
941 for (i = 0; i < (int)sizeof(name) - 1; i++)
942 {
943 c = *(++ptr);
944 if (c == 0) goto ERROR_RETURN;
945 if (c == CHAR_RIGHT_CURLY_BRACKET) break;
946 name[i] = c;
947 }
948 if (c != CHAR_RIGHT_CURLY_BRACKET) goto ERROR_RETURN;
949 name[i] = 0;
950 }
951
952 /* Otherwise there is just one following character */
953
954 else
955 {
956 name[0] = c;
957 name[1] = 0;
958 }
959
960 *ptrptr = ptr;
961
962 /* Search for a recognized property name using binary chop */
963
964 bot = 0;
965 top = _pcre_utt_size;
966
967 while (bot < top)
968 {
969 i = (bot + top) >> 1;
970 c = strcmp(name, _pcre_utt_names + _pcre_utt[i].name_offset);
971 if (c == 0)
972 {
973 *dptr = _pcre_utt[i].value;
974 return _pcre_utt[i].type;
975 }
976 if (c > 0) bot = i + 1; else top = i;
977 }
978
979 *errorcodeptr = ERR47;
980 *ptrptr = ptr;
981 return -1;
982
983 ERROR_RETURN:
984 *errorcodeptr = ERR46;
985 *ptrptr = ptr;
986 return -1;
987 }
988 #endif
989
990
991
992
993 /*************************************************
994 * Check for counted repeat *
995 *************************************************/
996
997 /* This function is called when a '{' is encountered in a place where it might
998 start a quantifier. It looks ahead to see if it really is a quantifier or not.
999 It is only a quantifier if it is one of the forms {ddd} {ddd,} or {ddd,ddd}
1000 where the ddds are digits.
1001
1002 Arguments:
1003 p pointer to the first char after '{'
1004
1005 Returns: TRUE or FALSE
1006 */
1007
1008 static BOOL
1009 is_counted_repeat(const uschar *p)
1010 {
1011 if ((digitab[*p++] & ctype_digit) == 0) return FALSE;
1012 while ((digitab[*p] & ctype_digit) != 0) p++;
1013 if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
1014
1015 if (*p++ != CHAR_COMMA) return FALSE;
1016 if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
1017
1018 if ((digitab[*p++] & ctype_digit) == 0) return FALSE;
1019 while ((digitab[*p] & ctype_digit) != 0) p++;
1020
1021 return (*p == CHAR_RIGHT_CURLY_BRACKET);
1022 }
1023
1024
1025
1026 /*************************************************
1027 * Read repeat counts *
1028 *************************************************/
1029
1030 /* Read an item of the form {n,m} and return the values. This is called only
1031 after is_counted_repeat() has confirmed that a repeat-count quantifier exists,
1032 so the syntax is guaranteed to be correct, but we need to check the values.
1033
1034 Arguments:
1035 p pointer to first char after '{'
1036 minp pointer to int for min
1037 maxp pointer to int for max
1038 returned as -1 if no max
1039 errorcodeptr points to error code variable
1040
1041 Returns: pointer to '}' on success;
1042 current ptr on error, with errorcodeptr set non-zero
1043 */
1044
1045 static const uschar *
1046 read_repeat_counts(const uschar *p, int *minp, int *maxp, int *errorcodeptr)
1047 {
1048 int min = 0;
1049 int max = -1;
1050
1051 /* Read the minimum value and do a paranoid check: a negative value indicates
1052 an integer overflow. */
1053
1054 while ((digitab[*p] & ctype_digit) != 0) min = min * 10 + *p++ - CHAR_0;
1055 if (min < 0 || min > 65535)
1056 {
1057 *errorcodeptr = ERR5;
1058 return p;
1059 }
1060
1061 /* Read the maximum value if there is one, and again do a paranoid on its size.
1062 Also, max must not be less than min. */
1063
1064 if (*p == CHAR_RIGHT_CURLY_BRACKET) max = min; else
1065 {
1066 if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
1067 {
1068 max = 0;
1069 while((digitab[*p] & ctype_digit) != 0) max = max * 10 + *p++ - CHAR_0;
1070 if (max < 0 || max > 65535)
1071 {
1072 *errorcodeptr = ERR5;
1073 return p;
1074 }
1075 if (max < min)
1076 {
1077 *errorcodeptr = ERR4;
1078 return p;
1079 }
1080 }
1081 }
1082
1083 /* Fill in the required variables, and pass back the pointer to the terminating
1084 '}'. */
1085
1086 *minp = min;
1087 *maxp = max;
1088 return p;
1089 }
1090
1091
1092
1093 /*************************************************
1094 * Subroutine for finding forward reference *
1095 *************************************************/
1096
1097 /* This recursive function is called only from find_parens() below. The
1098 top-level call starts at the beginning of the pattern. All other calls must
1099 start at a parenthesis. It scans along a pattern's text looking for capturing
1100 subpatterns, and counting them. If it finds a named pattern that matches the
1101 name it is given, it returns its number. Alternatively, if the name is NULL, it
1102 returns when it reaches a given numbered subpattern. We know that if (?P< is
1103 encountered, the name will be terminated by '>' because that is checked in the
1104 first pass. Recursion is used to keep track of subpatterns that reset the
1105 capturing group numbers - the (?| feature.
1106
1107 Arguments:
1108 ptrptr address of the current character pointer (updated)
1109 cd compile background data
1110 name name to seek, or NULL if seeking a numbered subpattern
1111 lorn name length, or subpattern number if name is NULL
1112 xmode TRUE if we are in /x mode
1113 count pointer to the current capturing subpattern number (updated)
1114
1115 Returns: the number of the named subpattern, or -1 if not found
1116 */
1117
1118 static int
1119 find_parens_sub(uschar **ptrptr, compile_data *cd, const uschar *name, int lorn,
1120 BOOL xmode, int *count)
1121 {
1122 uschar *ptr = *ptrptr;
1123 int start_count = *count;
1124 int hwm_count = start_count;
1125 BOOL dup_parens = FALSE;
1126
1127 /* If the first character is a parenthesis, check on the type of group we are
1128 dealing with. The very first call may not start with a parenthesis. */
1129
1130 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1131 {
1132 /* Handle specials such as (*SKIP) or (*UTF8) etc. */
1133
1134 if (ptr[1] == CHAR_ASTERISK) ptr += 2;
1135
1136 /* Handle a normal, unnamed capturing parenthesis. */
1137
1138 else if (ptr[1] != CHAR_QUESTION_MARK)
1139 {
1140 *count += 1;
1141 if (name == NULL && *count == lorn) return *count;
1142 ptr++;
1143 }
1144
1145 /* All cases now have (? at the start. Remember when we are in a group
1146 where the parenthesis numbers are duplicated. */
1147
1148 else if (ptr[2] == CHAR_VERTICAL_LINE)
1149 {
1150 ptr += 3;
1151 dup_parens = TRUE;
1152 }
1153
1154 /* Handle comments; all characters are allowed until a ket is reached. */
1155
1156 else if (ptr[2] == CHAR_NUMBER_SIGN)
1157 {
1158 for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
1159 goto FAIL_EXIT;
1160 }
1161
1162 /* Handle a condition. If it is an assertion, just carry on so that it
1163 is processed as normal. If not, skip to the closing parenthesis of the
1164 condition (there can't be any nested parens). */
1165
1166 else if (ptr[2] == CHAR_LEFT_PARENTHESIS)
1167 {
1168 ptr += 2;
1169 if (ptr[1] != CHAR_QUESTION_MARK)
1170 {
1171 while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
1172 if (*ptr != 0) ptr++;
1173 }
1174 }
1175
1176 /* Start with (? but not a condition. */
1177
1178 else
1179 {
1180 ptr += 2;
1181 if (*ptr == CHAR_P) ptr++; /* Allow optional P */
1182
1183 /* We have to disambiguate (?<! and (?<= from (?<name> for named groups */
1184
1185 if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&
1186 ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)
1187 {
1188 int term;
1189 const uschar *thisname;
1190 *count += 1;
1191 if (name == NULL && *count == lorn) return *count;
1192 term = *ptr++;
1193 if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;
1194 thisname = ptr;
1195 while (*ptr != term) ptr++;
1196 if (name != NULL && lorn == ptr - thisname &&
1197 strncmp((const char *)name, (const char *)thisname, lorn) == 0)
1198 return *count;
1199 term++;
1200 }
1201 }
1202 }
1203
1204 /* Past any initial parenthesis handling, scan for parentheses or vertical
1205 bars. */
1206
1207 for (; *ptr != 0; ptr++)
1208 {
1209 /* Skip over backslashed characters and also entire \Q...\E */
1210
1211 if (*ptr == CHAR_BACKSLASH)
1212 {
1213 if (*(++ptr) == 0) goto FAIL_EXIT;
1214 if (*ptr == CHAR_Q) for (;;)
1215 {
1216 while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};
1217 if (*ptr == 0) goto FAIL_EXIT;
1218 if (*(++ptr) == CHAR_E) break;
1219 }
1220 continue;
1221 }
1222
1223 /* Skip over character classes; this logic must be similar to the way they
1224 are handled for real. If the first character is '^', skip it. Also, if the
1225 first few characters (either before or after ^) are \Q\E or \E we skip them
1226 too. This makes for compatibility with Perl. Note the use of STR macros to
1227 encode "Q\\E" so that it works in UTF-8 on EBCDIC platforms. */
1228
1229 if (*ptr == CHAR_LEFT_SQUARE_BRACKET)
1230 {
1231 BOOL negate_class = FALSE;
1232 for (;;)
1233 {
1234 if (ptr[1] == CHAR_BACKSLASH)
1235 {
1236 if (ptr[2] == CHAR_E)
1237 ptr+= 2;
1238 else if (strncmp((const char *)ptr+2,
1239 STR_Q STR_BACKSLASH STR_E, 3) == 0)
1240 ptr += 4;
1241 else
1242 break;
1243 }
1244 else if (!negate_class && ptr[1] == CHAR_CIRCUMFLEX_ACCENT)
1245 {
1246 negate_class = TRUE;
1247 ptr++;
1248 }
1249 else break;
1250 }
1251
1252 /* If the next character is ']', it is a data character that must be
1253 skipped, except in JavaScript compatibility mode. */
1254
1255 if (ptr[1] == CHAR_RIGHT_SQUARE_BRACKET &&
1256 (cd->external_options & PCRE_JAVASCRIPT_COMPAT) == 0)
1257 ptr++;
1258
1259 while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)
1260 {
1261 if (*ptr == 0) return -1;
1262 if (*ptr == CHAR_BACKSLASH)
1263 {
1264 if (*(++ptr) == 0) goto FAIL_EXIT;
1265 if (*ptr == CHAR_Q) for (;;)
1266 {
1267 while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};
1268 if (*ptr == 0) goto FAIL_EXIT;
1269 if (*(++ptr) == CHAR_E) break;
1270 }
1271 continue;
1272 }
1273 }
1274 continue;
1275 }
1276
1277 /* Skip comments in /x mode */
1278
1279 if (xmode && *ptr == CHAR_NUMBER_SIGN)
1280 {
1281 while (*(++ptr) != 0 && *ptr != CHAR_NL) {};
1282 if (*ptr == 0) goto FAIL_EXIT;
1283 continue;
1284 }
1285
1286 /* Check for the special metacharacters */
1287
1288 if (*ptr == CHAR_LEFT_PARENTHESIS)
1289 {
1290 int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, count);
1291 if (rc > 0) return rc;
1292 if (*ptr == 0) goto FAIL_EXIT;
1293 }
1294
1295 else if (*ptr == CHAR_RIGHT_PARENTHESIS)
1296 {
1297 if (dup_parens && *count < hwm_count) *count = hwm_count;
1298 goto FAIL_EXIT;
1299 }
1300
1301 else if (*ptr == CHAR_VERTICAL_LINE && dup_parens)
1302 {
1303 if (*count > hwm_count) hwm_count = *count;
1304 *count = start_count;
1305 }
1306 }
1307
1308 FAIL_EXIT:
1309 *ptrptr = ptr;
1310 return -1;
1311 }
1312
1313
1314
1315
1316 /*************************************************
1317 * Find forward referenced subpattern *
1318 *************************************************/
1319
1320 /* This function scans along a pattern's text looking for capturing
1321 subpatterns, and counting them. If it finds a named pattern that matches the
1322 name it is given, it returns its number. Alternatively, if the name is NULL, it
1323 returns when it reaches a given numbered subpattern. This is used for forward
1324 references to subpatterns. We used to be able to start this scan from the
1325 current compiling point, using the current count value from cd->bracount, and
1326 do it all in a single loop, but the addition of the possibility of duplicate
1327 subpattern numbers means that we have to scan from the very start, in order to
1328 take account of such duplicates, and to use a recursive function to keep track
1329 of the different types of group.
1330
1331 Arguments:
1332 cd compile background data
1333 name name to seek, or NULL if seeking a numbered subpattern
1334 lorn name length, or subpattern number if name is NULL
1335 xmode TRUE if we are in /x mode
1336
1337 Returns: the number of the found subpattern, or -1 if not found
1338 */
1339
1340 static int
1341 find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode)
1342 {
1343 uschar *ptr = (uschar *)cd->start_pattern;
1344 int count = 0;
1345 int rc;
1346
1347 /* If the pattern does not start with an opening parenthesis, the first call
1348 to find_parens_sub() will scan right to the end (if necessary). However, if it
1349 does start with a parenthesis, find_parens_sub() will return when it hits the
1350 matching closing parens. That is why we have to have a loop. */
1351
1352 for (;;)
1353 {
1354 rc = find_parens_sub(&ptr, cd, name, lorn, xmode, &count);
1355 if (rc > 0 || *ptr++ == 0) break;
1356 }
1357
1358 return rc;
1359 }
1360
1361
1362
1363
1364 /*************************************************
1365 * Find first significant op code *
1366 *************************************************/
1367
1368 /* This is called by several functions that scan a compiled expression looking
1369 for a fixed first character, or an anchoring op code etc. It skips over things
1370 that do not influence this. For some calls, a change of option is important.
1371 For some calls, it makes sense to skip negative forward and all backward
1372 assertions, and also the \b assertion; for others it does not.
1373
1374 Arguments:
1375 code pointer to the start of the group
1376 options pointer to external options
1377 optbit the option bit whose changing is significant, or
1378 zero if none are
1379 skipassert TRUE if certain assertions are to be skipped
1380
1381 Returns: pointer to the first significant opcode
1382 */
1383
1384 static const uschar*
1385 first_significant_code(const uschar *code, int *options, int optbit,
1386 BOOL skipassert)
1387 {
1388 for (;;)
1389 {
1390 switch ((int)*code)
1391 {
1392 case OP_OPT:
1393 if (optbit > 0 && ((int)code[1] & optbit) != (*options & optbit))
1394 *options = (int)code[1];
1395 code += 2;
1396 break;
1397
1398 case OP_ASSERT_NOT:
1399 case OP_ASSERTBACK:
1400 case OP_ASSERTBACK_NOT:
1401 if (!skipassert) return code;
1402 do code += GET(code, 1); while (*code == OP_ALT);
1403 code += _pcre_OP_lengths[*code];
1404 break;
1405
1406 case OP_WORD_BOUNDARY:
1407 case OP_NOT_WORD_BOUNDARY:
1408 if (!skipassert) return code;
1409 /* Fall through */
1410
1411 case OP_CALLOUT:
1412 case OP_CREF:
1413 case OP_NCREF:
1414 case OP_RREF:
1415 case OP_NRREF:
1416 case OP_DEF:
1417 code += _pcre_OP_lengths[*code];
1418 break;
1419
1420 default:
1421 return code;
1422 }
1423 }
1424 /* Control never reaches here */
1425 }
1426
1427
1428
1429
1430 /*************************************************
1431 * Find the fixed length of a branch *
1432 *************************************************/
1433
1434 /* Scan a branch and compute the fixed length of subject that will match it,
1435 if the length is fixed. This is needed for dealing with backward assertions.
1436 In UTF8 mode, the result is in characters rather than bytes. The branch is
1437 temporarily terminated with OP_END when this function is called.
1438
1439 This function is called when a backward assertion is encountered, so that if it
1440 fails, the error message can point to the correct place in the pattern.
1441 However, we cannot do this when the assertion contains subroutine calls,
1442 because they can be forward references. We solve this by remembering this case
1443 and doing the check at the end; a flag specifies which mode we are running in.
1444
1445 Arguments:
1446 code points to the start of the pattern (the bracket)
1447 options the compiling options
1448 atend TRUE if called when the pattern is complete
1449 cd the "compile data" structure
1450
1451 Returns: the fixed length,
1452 or -1 if there is no fixed length,
1453 or -2 if \C was encountered
1454 or -3 if an OP_RECURSE item was encountered and atend is FALSE
1455 */
1456
1457 static int
1458 find_fixedlength(uschar *code, int options, BOOL atend, compile_data *cd)
1459 {
1460 int length = -1;
1461
1462 register int branchlength = 0;
1463 register uschar *cc = code + 1 + LINK_SIZE;
1464
1465 /* Scan along the opcodes for this branch. If we get to the end of the
1466 branch, check the length against that of the other branches. */
1467
1468 for (;;)
1469 {
1470 int d;
1471 uschar *ce, *cs;
1472 register int op = *cc;
1473 switch (op)
1474 {
1475 case OP_CBRA:
1476 case OP_BRA:
1477 case OP_ONCE:
1478 case OP_COND:
1479 d = find_fixedlength(cc + ((op == OP_CBRA)? 2:0), options, atend, cd);
1480 if (d < 0) return d;
1481 branchlength += d;
1482 do cc += GET(cc, 1); while (*cc == OP_ALT);
1483 cc += 1 + LINK_SIZE;
1484 break;
1485
1486 /* Reached end of a branch; if it's a ket it is the end of a nested
1487 call. If it's ALT it is an alternation in a nested call. If it is
1488 END it's the end of the outer call. All can be handled by the same code. */
1489
1490 case OP_ALT:
1491 case OP_KET:
1492 case OP_KETRMAX:
1493 case OP_KETRMIN:
1494 case OP_END:
1495 if (length < 0) length = branchlength;
1496 else if (length != branchlength) return -1;
1497 if (*cc != OP_ALT) return length;
1498 cc += 1 + LINK_SIZE;
1499 branchlength = 0;
1500 break;
1501
1502 /* A true recursion implies not fixed length, but a subroutine call may
1503 be OK. If the subroutine is a forward reference, we can't deal with
1504 it until the end of the pattern, so return -3. */
1505
1506 case OP_RECURSE:
1507 if (!atend) return -3;
1508 cs = ce = (uschar *)cd->start_code + GET(cc, 1); /* Start subpattern */
1509 do ce += GET(ce, 1); while (*ce == OP_ALT); /* End subpattern */
1510 if (cc > cs && cc < ce) return -1; /* Recursion */
1511 d = find_fixedlength(cs + 2, options, atend, cd);
1512 if (d < 0) return d;
1513 branchlength += d;
1514 cc += 1 + LINK_SIZE;
1515 break;
1516
1517 /* Skip over assertive subpatterns */
1518
1519 case OP_ASSERT:
1520 case OP_ASSERT_NOT:
1521 case OP_ASSERTBACK:
1522 case OP_ASSERTBACK_NOT:
1523 do cc += GET(cc, 1); while (*cc == OP_ALT);
1524 /* Fall through */
1525
1526 /* Skip over things that don't match chars */
1527
1528 case OP_REVERSE:
1529 case OP_CREF:
1530 case OP_NCREF:
1531 case OP_RREF:
1532 case OP_NRREF:
1533 case OP_DEF:
1534 case OP_OPT:
1535 case OP_CALLOUT:
1536 case OP_SOD:
1537 case OP_SOM:
1538 case OP_SET_SOM:
1539 case OP_EOD:
1540 case OP_EODN:
1541 case OP_CIRC:
1542 case OP_DOLL:
1543 case OP_NOT_WORD_BOUNDARY:
1544 case OP_WORD_BOUNDARY:
1545 cc += _pcre_OP_lengths[*cc];
1546 break;
1547
1548 /* Handle literal characters */
1549
1550 case OP_CHAR:
1551 case OP_CHARNC:
1552 case OP_NOT:
1553 branchlength++;
1554 cc += 2;
1555 #ifdef SUPPORT_UTF8
1556 if ((options & PCRE_UTF8) != 0 && cc[-1] >= 0xc0)
1557 cc += _pcre_utf8_table4[cc[-1] & 0x3f];
1558 #endif
1559 break;
1560
1561 /* Handle exact repetitions. The count is already in characters, but we
1562 need to skip over a multibyte character in UTF8 mode. */
1563
1564 case OP_EXACT:
1565 branchlength += GET2(cc,1);
1566 cc += 4;
1567 #ifdef SUPPORT_UTF8
1568 if ((options & PCRE_UTF8) != 0 && cc[-1] >= 0xc0)
1569 cc += _pcre_utf8_table4[cc[-1] & 0x3f];
1570 #endif
1571 break;
1572
1573 case OP_TYPEEXACT:
1574 branchlength += GET2(cc,1);
1575 if (cc[3] == OP_PROP || cc[3] == OP_NOTPROP) cc += 2;
1576 cc += 4;
1577 break;
1578
1579 /* Handle single-char matchers */
1580
1581 case OP_PROP:
1582 case OP_NOTPROP:
1583 cc += 2;
1584 /* Fall through */
1585
1586 case OP_NOT_DIGIT:
1587 case OP_DIGIT:
1588 case OP_NOT_WHITESPACE:
1589 case OP_WHITESPACE:
1590 case OP_NOT_WORDCHAR:
1591 case OP_WORDCHAR:
1592 case OP_ANY:
1593 case OP_ALLANY:
1594 branchlength++;
1595 cc++;
1596 break;
1597
1598 /* The single-byte matcher isn't allowed */
1599
1600 case OP_ANYBYTE:
1601 return -2;
1602
1603 /* Check a class for variable quantification */
1604
1605 #ifdef SUPPORT_UTF8
1606 case OP_XCLASS:
1607 cc += GET(cc, 1) - 33;
1608 /* Fall through */
1609 #endif
1610
1611 case OP_CLASS:
1612 case OP_NCLASS:
1613 cc += 33;
1614
1615 switch (*cc)
1616 {
1617 case OP_CRSTAR:
1618 case OP_CRMINSTAR:
1619 case OP_CRQUERY:
1620 case OP_CRMINQUERY:
1621 return -1;
1622
1623 case OP_CRRANGE:
1624 case OP_CRMINRANGE:
1625 if (GET2(cc,1) != GET2(cc,3)) return -1;
1626 branchlength += GET2(cc,1);
1627 cc += 5;
1628 break;
1629
1630 default:
1631 branchlength++;
1632 }
1633 break;
1634
1635 /* Anything else is variable length */
1636
1637 default:
1638 return -1;
1639 }
1640 }
1641 /* Control never gets here */
1642 }
1643
1644
1645
1646
1647 /*************************************************
1648 * Scan compiled regex for specific bracket *
1649 *************************************************/
1650
1651 /* This little function scans through a compiled pattern until it finds a
1652 capturing bracket with the given number, or, if the number is negative, an
1653 instance of OP_REVERSE for a lookbehind. The function is global in the C sense
1654 so that it can be called from pcre_study() when finding the minimum matching
1655 length.
1656
1657 Arguments:
1658 code points to start of expression
1659 utf8 TRUE in UTF-8 mode
1660 number the required bracket number or negative to find a lookbehind
1661
1662 Returns: pointer to the opcode for the bracket, or NULL if not found
1663 */
1664
1665 const uschar *
1666 _pcre_find_bracket(const uschar *code, BOOL utf8, int number)
1667 {
1668 for (;;)
1669 {
1670 register int c = *code;
1671 if (c == OP_END) return NULL;
1672
1673 /* XCLASS is used for classes that cannot be represented just by a bit
1674 map. This includes negated single high-valued characters. The length in
1675 the table is zero; the actual length is stored in the compiled code. */
1676
1677 if (c == OP_XCLASS) code += GET(code, 1);
1678
1679 /* Handle recursion */
1680
1681 else if (c == OP_REVERSE)
1682 {
1683 if (number < 0) return (uschar *)code;
1684 code += _pcre_OP_lengths[c];
1685 }
1686
1687 /* Handle capturing bracket */
1688
1689 else if (c == OP_CBRA)
1690 {
1691 int n = GET2(code, 1+LINK_SIZE);
1692 if (n == number) return (uschar *)code;
1693 code += _pcre_OP_lengths[c];
1694 }
1695
1696 /* Otherwise, we can get the item's length from the table, except that for
1697 repeated character types, we have to test for \p and \P, which have an extra
1698 two bytes of parameters, and for MARK/PRUNE/SKIP/THEN with an argument, we
1699 must add in its length. */
1700
1701 else
1702 {
1703 switch(c)
1704 {
1705 case OP_TYPESTAR:
1706 case OP_TYPEMINSTAR:
1707 case OP_TYPEPLUS:
1708 case OP_TYPEMINPLUS:
1709 case OP_TYPEQUERY:
1710 case OP_TYPEMINQUERY:
1711 case OP_TYPEPOSSTAR:
1712 case OP_TYPEPOSPLUS:
1713 case OP_TYPEPOSQUERY:
1714 if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1715 break;
1716
1717 case OP_TYPEUPTO:
1718 case OP_TYPEMINUPTO:
1719 case OP_TYPEEXACT:
1720 case OP_TYPEPOSUPTO:
1721 if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
1722 break;
1723
1724 case OP_MARK:
1725 case OP_PRUNE_ARG:
1726 case OP_SKIP_ARG:
1727 code += code[1];
1728 break;
1729
1730 case OP_THEN_ARG:
1731 code += code[1+LINK_SIZE];
1732 break;
1733 }
1734
1735 /* Add in the fixed length from the table */
1736
1737 code += _pcre_OP_lengths[c];
1738
1739 /* In UTF-8 mode, opcodes that are followed by a character may be followed by
1740 a multi-byte character. The length in the table is a minimum, so we have to
1741 arrange to skip the extra bytes. */
1742
1743 #ifdef SUPPORT_UTF8
1744 if (utf8) switch(c)
1745 {
1746 case OP_CHAR:
1747 case OP_CHARNC:
1748 case OP_EXACT:
1749 case OP_UPTO:
1750 case OP_MINUPTO:
1751 case OP_POSUPTO:
1752 case OP_STAR:
1753 case OP_MINSTAR:
1754 case OP_POSSTAR:
1755 case OP_PLUS:
1756 case OP_MINPLUS:
1757 case OP_POSPLUS:
1758 case OP_QUERY:
1759 case OP_MINQUERY:
1760 case OP_POSQUERY:
1761 if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];
1762 break;
1763 }
1764 #else
1765 (void)(utf8); /* Keep compiler happy by referencing function argument */
1766 #endif
1767 }
1768 }
1769 }
1770
1771
1772
1773 /*************************************************
1774 * Scan compiled regex for recursion reference *
1775 *************************************************/
1776
1777 /* This little function scans through a compiled pattern until it finds an
1778 instance of OP_RECURSE.
1779
1780 Arguments:
1781 code points to start of expression
1782 utf8 TRUE in UTF-8 mode
1783
1784 Returns: pointer to the opcode for OP_RECURSE, or NULL if not found
1785 */
1786
1787 static const uschar *
1788 find_recurse(const uschar *code, BOOL utf8)
1789 {
1790 for (;;)
1791 {
1792 register int c = *code;
1793 if (c == OP_END) return NULL;
1794 if (c == OP_RECURSE) return code;
1795
1796 /* XCLASS is used for classes that cannot be represented just by a bit
1797 map. This includes negated single high-valued characters. The length in
1798 the table is zero; the actual length is stored in the compiled code. */
1799
1800 if (c == OP_XCLASS) code += GET(code, 1);
1801
1802 /* Otherwise, we can get the item's length from the table, except that for
1803 repeated character types, we have to test for \p and \P, which have an extra
1804 two bytes of parameters, and for MARK/PRUNE/SKIP/THEN with an argument, we
1805 must add in its length. */
1806
1807 else
1808 {
1809 switch(c)
1810 {
1811 case OP_TYPESTAR:
1812 case OP_TYPEMINSTAR:
1813 case OP_TYPEPLUS:
1814 case OP_TYPEMINPLUS:
1815 case OP_TYPEQUERY:
1816 case OP_TYPEMINQUERY:
1817 case OP_TYPEPOSSTAR:
1818 case OP_TYPEPOSPLUS:
1819 case OP_TYPEPOSQUERY:
1820 if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1821 break;
1822
1823 case OP_TYPEPOSUPTO:
1824 case OP_TYPEUPTO:
1825 case OP_TYPEMINUPTO:
1826 case OP_TYPEEXACT:
1827 if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
1828 break;
1829
1830 case OP_MARK:
1831 case OP_PRUNE_ARG:
1832 case OP_SKIP_ARG:
1833 code += code[1];
1834 break;
1835
1836 case OP_THEN_ARG:
1837 code += code[1+LINK_SIZE];
1838 break;
1839 }
1840
1841 /* Add in the fixed length from the table */
1842
1843 code += _pcre_OP_lengths[c];
1844
1845 /* In UTF-8 mode, opcodes that are followed by a character may be followed
1846 by a multi-byte character. The length in the table is a minimum, so we have
1847 to arrange to skip the extra bytes. */
1848
1849 #ifdef SUPPORT_UTF8
1850 if (utf8) switch(c)
1851 {
1852 case OP_CHAR:
1853 case OP_CHARNC:
1854 case OP_EXACT:
1855 case OP_UPTO:
1856 case OP_MINUPTO:
1857 case OP_POSUPTO:
1858 case OP_STAR:
1859 case OP_MINSTAR:
1860 case OP_POSSTAR:
1861 case OP_PLUS:
1862 case OP_MINPLUS:
1863 case OP_POSPLUS:
1864 case OP_QUERY:
1865 case OP_MINQUERY:
1866 case OP_POSQUERY:
1867 if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];
1868 break;
1869 }
1870 #else
1871 (void)(utf8); /* Keep compiler happy by referencing function argument */
1872 #endif
1873 }
1874 }
1875 }
1876
1877
1878
1879 /*************************************************
1880 * Scan compiled branch for non-emptiness *
1881 *************************************************/
1882
1883 /* This function scans through a branch of a compiled pattern to see whether it
1884 can match the empty string or not. It is called from could_be_empty()
1885 below and from compile_branch() when checking for an unlimited repeat of a
1886 group that can match nothing. Note that first_significant_code() skips over
1887 backward and negative forward assertions when its final argument is TRUE. If we
1888 hit an unclosed bracket, we return "empty" - this means we've struck an inner
1889 bracket whose current branch will already have been scanned.
1890
1891 Arguments:
1892 code points to start of search
1893 endcode points to where to stop
1894 utf8 TRUE if in UTF8 mode
1895 cd contains pointers to tables etc.
1896
1897 Returns: TRUE if what is matched could be empty
1898 */
1899
1900 static BOOL
1901 could_be_empty_branch(const uschar *code, const uschar *endcode, BOOL utf8,
1902 compile_data *cd)
1903 {
1904 register int c;
1905 for (code = first_significant_code(code + _pcre_OP_lengths[*code], NULL, 0, TRUE);
1906 code < endcode;
1907 code = first_significant_code(code + _pcre_OP_lengths[c], NULL, 0, TRUE))
1908 {
1909 const uschar *ccode;
1910
1911 c = *code;
1912
1913 /* Skip over forward assertions; the other assertions are skipped by
1914 first_significant_code() with a TRUE final argument. */
1915
1916 if (c == OP_ASSERT)
1917 {
1918 do code += GET(code, 1); while (*code == OP_ALT);
1919 c = *code;
1920 continue;
1921 }
1922
1923 /* Groups with zero repeats can of course be empty; skip them. */
1924
1925 if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO)
1926 {
1927 code += _pcre_OP_lengths[c];
1928 do code += GET(code, 1); while (*code == OP_ALT);
1929 c = *code;
1930 continue;
1931 }
1932
1933 /* For a recursion/subroutine call, if its end has been reached, which
1934 implies a subroutine call, we can scan it. */
1935
1936 if (c == OP_RECURSE)
1937 {
1938 BOOL empty_branch = FALSE;
1939 const uschar *scode = cd->start_code + GET(code, 1);
1940 if (GET(scode, 1) == 0) return TRUE; /* Unclosed */
1941 do
1942 {
1943 if (could_be_empty_branch(scode, endcode, utf8, cd))
1944 {
1945 empty_branch = TRUE;
1946 break;
1947 }
1948 scode += GET(scode, 1);
1949 }
1950 while (*scode == OP_ALT);
1951 if (!empty_branch) return FALSE; /* All branches are non-empty */
1952 continue;
1953 }
1954
1955 /* For other groups, scan the branches. */
1956
1957 if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE || c == OP_COND)
1958 {
1959 BOOL empty_branch;
1960 if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */
1961
1962 /* If a conditional group has only one branch, there is a second, implied,
1963 empty branch, so just skip over the conditional, because it could be empty.
1964 Otherwise, scan the individual branches of the group. */
1965
1966 if (c == OP_COND && code[GET(code, 1)] != OP_ALT)
1967 code += GET(code, 1);
1968 else
1969 {
1970 empty_branch = FALSE;
1971 do
1972 {
1973 if (!empty_branch && could_be_empty_branch(code, endcode, utf8, cd))
1974 empty_branch = TRUE;
1975 code += GET(code, 1);
1976 }
1977 while (*code == OP_ALT);
1978 if (!empty_branch) return FALSE; /* All branches are non-empty */
1979 }
1980
1981 c = *code;
1982 continue;
1983 }
1984
1985 /* Handle the other opcodes */
1986
1987 switch (c)
1988 {
1989 /* Check for quantifiers after a class. XCLASS is used for classes that
1990 cannot be represented just by a bit map. This includes negated single
1991 high-valued characters. The length in _pcre_OP_lengths[] is zero; the
1992 actual length is stored in the compiled code, so we must update "code"
1993 here. */
1994
1995 #ifdef SUPPORT_UTF8
1996 case OP_XCLASS:
1997 ccode = code += GET(code, 1);
1998 goto CHECK_CLASS_REPEAT;
1999 #endif
2000
2001 case OP_CLASS:
2002 case OP_NCLASS:
2003 ccode = code + 33;
2004
2005 #ifdef SUPPORT_UTF8
2006 CHECK_CLASS_REPEAT:
2007 #endif
2008
2009 switch (*ccode)
2010 {
2011 case OP_CRSTAR: /* These could be empty; continue */
2012 case OP_CRMINSTAR:
2013 case OP_CRQUERY:
2014 case OP_CRMINQUERY:
2015 break;
2016
2017 default: /* Non-repeat => class must match */
2018 case OP_CRPLUS: /* These repeats aren't empty */
2019 case OP_CRMINPLUS:
2020 return FALSE;
2021
2022 case OP_CRRANGE:
2023 case OP_CRMINRANGE:
2024 if (GET2(ccode, 1) > 0) return FALSE; /* Minimum > 0 */
2025 break;
2026 }
2027 break;
2028
2029 /* Opcodes that must match a character */
2030
2031 case OP_PROP:
2032 case OP_NOTPROP:
2033 case OP_EXTUNI:
2034 case OP_NOT_DIGIT:
2035 case OP_DIGIT:
2036 case OP_NOT_WHITESPACE:
2037 case OP_WHITESPACE:
2038 case OP_NOT_WORDCHAR:
2039 case OP_WORDCHAR:
2040 case OP_ANY:
2041 case OP_ALLANY:
2042 case OP_ANYBYTE:
2043 case OP_CHAR:
2044 case OP_CHARNC:
2045 case OP_NOT:
2046 case OP_PLUS:
2047 case OP_MINPLUS:
2048 case OP_POSPLUS:
2049 case OP_EXACT:
2050 case OP_NOTPLUS:
2051 case OP_NOTMINPLUS:
2052 case OP_NOTPOSPLUS:
2053 case OP_NOTEXACT:
2054 case OP_TYPEPLUS:
2055 case OP_TYPEMINPLUS:
2056 case OP_TYPEPOSPLUS:
2057 case OP_TYPEEXACT:
2058 return FALSE;
2059
2060 /* These are going to continue, as they may be empty, but we have to
2061 fudge the length for the \p and \P cases. */
2062
2063 case OP_TYPESTAR:
2064 case OP_TYPEMINSTAR:
2065 case OP_TYPEPOSSTAR:
2066 case OP_TYPEQUERY:
2067 case OP_TYPEMINQUERY:
2068 case OP_TYPEPOSQUERY:
2069 if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
2070 break;
2071
2072 /* Same for these */
2073
2074 case OP_TYPEUPTO:
2075 case OP_TYPEMINUPTO:
2076 case OP_TYPEPOSUPTO:
2077 if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
2078 break;
2079
2080 /* End of branch */
2081
2082 case OP_KET:
2083 case OP_KETRMAX:
2084 case OP_KETRMIN:
2085 case OP_ALT:
2086 return TRUE;
2087
2088 /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
2089 MINUPTO, and POSUPTO may be followed by a multibyte character */
2090
2091 #ifdef SUPPORT_UTF8
2092 case OP_STAR:
2093 case OP_MINSTAR:
2094 case OP_POSSTAR:
2095 case OP_QUERY:
2096 case OP_MINQUERY:
2097 case OP_POSQUERY:
2098 if (utf8 && code[1] >= 0xc0) code += _pcre_utf8_table4[code[1] & 0x3f];
2099 break;
2100
2101 case OP_UPTO:
2102 case OP_MINUPTO:
2103 case OP_POSUPTO:
2104 if (utf8 && code[3] >= 0xc0) code += _pcre_utf8_table4[code[3] & 0x3f];
2105 break;
2106 #endif
2107
2108 /* MARK, and PRUNE/SKIP/THEN with an argument must skip over the argument
2109 string. */
2110
2111 case OP_MARK:
2112 case OP_PRUNE_ARG:
2113 case OP_SKIP_ARG:
2114 code += code[1];
2115 break;
2116
2117 case OP_THEN_ARG:
2118 code += code[1+LINK_SIZE];
2119 break;
2120
2121 /* None of the remaining opcodes are required to match a character. */
2122
2123 default:
2124 break;
2125 }
2126 }
2127
2128 return TRUE;
2129 }
2130
2131
2132
2133 /*************************************************
2134 * Scan compiled regex for non-emptiness *
2135 *************************************************/
2136
2137 /* This function is called to check for left recursive calls. We want to check
2138 the current branch of the current pattern to see if it could match the empty
2139 string. If it could, we must look outwards for branches at other levels,
2140 stopping when we pass beyond the bracket which is the subject of the recursion.
2141
2142 Arguments:
2143 code points to start of the recursion
2144 endcode points to where to stop (current RECURSE item)
2145 bcptr points to the chain of current (unclosed) branch starts
2146 utf8 TRUE if in UTF-8 mode
2147 cd pointers to tables etc
2148
2149 Returns: TRUE if what is matched could be empty
2150 */
2151
2152 static BOOL
2153 could_be_empty(const uschar *code, const uschar *endcode, branch_chain *bcptr,
2154 BOOL utf8, compile_data *cd)
2155 {
2156 while (bcptr != NULL && bcptr->current_branch >= code)
2157 {
2158 if (!could_be_empty_branch(bcptr->current_branch, endcode, utf8, cd))
2159 return FALSE;
2160 bcptr = bcptr->outer;
2161 }
2162 return TRUE;
2163 }
2164
2165
2166
2167 /*************************************************
2168 * Check for POSIX class syntax *
2169 *************************************************/
2170
2171 /* This function is called when the sequence "[:" or "[." or "[=" is
2172 encountered in a character class. It checks whether this is followed by a
2173 sequence of characters terminated by a matching ":]" or ".]" or "=]". If we
2174 reach an unescaped ']' without the special preceding character, return FALSE.
2175
2176 Originally, this function only recognized a sequence of letters between the
2177 terminators, but it seems that Perl recognizes any sequence of characters,
2178 though of course unknown POSIX names are subsequently rejected. Perl gives an
2179 "Unknown POSIX class" error for [:f\oo:] for example, where previously PCRE
2180 didn't consider this to be a POSIX class. Likewise for [:1234:].
2181
2182 The problem in trying to be exactly like Perl is in the handling of escapes. We
2183 have to be sure that [abc[:x\]pqr] is *not* treated as containing a POSIX
2184 class, but [abc[:x\]pqr:]] is (so that an error can be generated). The code
2185 below handles the special case of \], but does not try to do any other escape
2186 processing. This makes it different from Perl for cases such as [:l\ower:]
2187 where Perl recognizes it as the POSIX class "lower" but PCRE does not recognize
2188 "l\ower". This is a lesser evil that not diagnosing bad classes when Perl does,
2189 I think.
2190
2191 Arguments:
2192 ptr pointer to the initial [
2193 endptr where to return the end pointer
2194
2195 Returns: TRUE or FALSE
2196 */
2197
2198 static BOOL
2199 check_posix_syntax(const uschar *ptr, const uschar **endptr)
2200 {
2201 int terminator; /* Don't combine these lines; the Solaris cc */
2202 terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */
2203 for (++ptr; *ptr != 0; ptr++)
2204 {
2205 if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) ptr++; else
2206 {
2207 if (*ptr == CHAR_RIGHT_SQUARE_BRACKET) return FALSE;
2208 if (*ptr == terminator && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
2209 {
2210 *endptr = ptr;
2211 return TRUE;
2212 }
2213 }
2214 }
2215 return FALSE;
2216 }
2217
2218
2219
2220
2221 /*************************************************
2222 * Check POSIX class name *
2223 *************************************************/
2224
2225 /* This function is called to check the name given in a POSIX-style class entry
2226 such as [:alnum:].
2227
2228 Arguments:
2229 ptr points to the first letter
2230 len the length of the name
2231
2232 Returns: a value representing the name, or -1 if unknown
2233 */
2234
2235 static int
2236 check_posix_name(const uschar *ptr, int len)
2237 {
2238 const char *pn = posix_names;
2239 register int yield = 0;
2240 while (posix_name_lengths[yield] != 0)
2241 {
2242 if (len == posix_name_lengths[yield] &&
2243 strncmp((const char *)ptr, pn, len) == 0) return yield;
2244 pn += posix_name_lengths[yield] + 1;
2245 yield++;
2246 }
2247 return -1;
2248 }
2249
2250
2251 /*************************************************
2252 * Adjust OP_RECURSE items in repeated group *
2253 *************************************************/
2254
2255 /* OP_RECURSE items contain an offset from the start of the regex to the group
2256 that is referenced. This means that groups can be replicated for fixed
2257 repetition simply by copying (because the recursion is allowed to refer to
2258 earlier groups that are outside the current group). However, when a group is
2259 optional (i.e. the minimum quantifier is zero), OP_BRAZERO or OP_SKIPZERO is
2260 inserted before it, after it has been compiled. This means that any OP_RECURSE
2261 items within it that refer to the group itself or any contained groups have to
2262 have their offsets adjusted. That one of the jobs of this function. Before it
2263 is called, the partially compiled regex must be temporarily terminated with
2264 OP_END.
2265
2266 This function has been extended with the possibility of forward references for
2267 recursions and subroutine calls. It must also check the list of such references
2268 for the group we are dealing with. If it finds that one of the recursions in
2269 the current group is on this list, it adjusts the offset in the list, not the
2270 value in the reference (which is a group number).
2271
2272 Arguments:
2273 group points to the start of the group
2274 adjust the amount by which the group is to be moved
2275 utf8 TRUE in UTF-8 mode
2276 cd contains pointers to tables etc.
2277 save_hwm the hwm forward reference pointer at the start of the group
2278
2279 Returns: nothing
2280 */
2281
2282 static void
2283 adjust_recurse(uschar *group, int adjust, BOOL utf8, compile_data *cd,
2284 uschar *save_hwm)
2285 {
2286 uschar *ptr = group;
2287
2288 while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL)
2289 {
2290 int offset;
2291 uschar *hc;
2292
2293 /* See if this recursion is on the forward reference list. If so, adjust the
2294 reference. */
2295
2296 for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
2297 {
2298 offset = GET(hc, 0);
2299 if (cd->start_code + offset == ptr + 1)
2300 {
2301 PUT(hc, 0, offset + adjust);
2302 break;
2303 }
2304 }
2305
2306 /* Otherwise, adjust the recursion offset if it's after the start of this
2307 group. */
2308
2309 if (hc >= cd->hwm)
2310 {
2311 offset = GET(ptr, 1);
2312 if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
2313 }
2314
2315 ptr += 1 + LINK_SIZE;
2316 }
2317 }
2318
2319
2320
2321 /*************************************************
2322 * Insert an automatic callout point *
2323 *************************************************/
2324
2325 /* This function is called when the PCRE_AUTO_CALLOUT option is set, to insert
2326 callout points before each pattern item.
2327
2328 Arguments:
2329 code current code pointer
2330 ptr current pattern pointer
2331 cd pointers to tables etc
2332
2333 Returns: new code pointer
2334 */
2335
2336 static uschar *
2337 auto_callout(uschar *code, const uschar *ptr, compile_data *cd)
2338 {
2339 *code++ = OP_CALLOUT;
2340 *code++ = 255;
2341 PUT(code, 0, (int)(ptr - cd->start_pattern)); /* Pattern offset */
2342 PUT(code, LINK_SIZE, 0); /* Default length */
2343 return code + 2*LINK_SIZE;
2344 }
2345
2346
2347
2348 /*************************************************
2349 * Complete a callout item *
2350 *************************************************/
2351
2352 /* A callout item contains the length of the next item in the pattern, which
2353 we can't fill in till after we have reached the relevant point. This is used
2354 for both automatic and manual callouts.
2355
2356 Arguments:
2357 previous_callout points to previous callout item
2358 ptr current pattern pointer
2359 cd pointers to tables etc
2360
2361 Returns: nothing
2362 */
2363
2364 static void
2365 complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd)
2366 {
2367 int length = (int)(ptr - cd->start_pattern - GET(previous_callout, 2));
2368 PUT(previous_callout, 2 + LINK_SIZE, length);
2369 }
2370
2371
2372
2373 #ifdef SUPPORT_UCP
2374 /*************************************************
2375 * Get othercase range *
2376 *************************************************/
2377
2378 /* This function is passed the start and end of a class range, in UTF-8 mode
2379 with UCP support. It searches up the characters, looking for internal ranges of
2380 characters in the "other" case. Each call returns the next one, updating the
2381 start address.
2382
2383 Arguments:
2384 cptr points to starting character value; updated
2385 d end value
2386 ocptr where to put start of othercase range
2387 odptr where to put end of othercase range
2388
2389 Yield: TRUE when range returned; FALSE when no more
2390 */
2391
2392 static BOOL
2393 get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr,
2394 unsigned int *odptr)
2395 {
2396 unsigned int c, othercase, next;
2397
2398 for (c = *cptr; c <= d; c++)
2399 { if ((othercase = UCD_OTHERCASE(c)) != c) break; }
2400
2401 if (c > d) return FALSE;
2402
2403 *ocptr = othercase;
2404 next = othercase + 1;
2405
2406 for (++c; c <= d; c++)
2407 {
2408 if (UCD_OTHERCASE(c) != next) break;
2409 next++;
2410 }
2411
2412 *odptr = next - 1;
2413 *cptr = c;
2414
2415 return TRUE;
2416 }
2417
2418
2419
2420 /*************************************************
2421 * Check a character and a property *
2422 *************************************************/
2423
2424 /* This function is called by check_auto_possessive() when a property item
2425 is adjacent to a fixed character.
2426
2427 Arguments:
2428 c the character
2429 ptype the property type
2430 pdata the data for the type
2431 negated TRUE if it's a negated property (\P or \p{^)
2432
2433 Returns: TRUE if auto-possessifying is OK
2434 */
2435
2436 static BOOL
2437 check_char_prop(int c, int ptype, int pdata, BOOL negated)
2438 {
2439 const ucd_record *prop = GET_UCD(c);
2440 switch(ptype)
2441 {
2442 case PT_LAMP:
2443 return (prop->chartype == ucp_Lu ||
2444 prop->chartype == ucp_Ll ||
2445 prop->chartype == ucp_Lt) == negated;
2446
2447 case PT_GC:
2448 return (pdata == _pcre_ucp_gentype[prop->chartype]) == negated;
2449
2450 case PT_PC:
2451 return (pdata == prop->chartype) == negated;
2452
2453 case PT_SC:
2454 return (pdata == prop->script) == negated;
2455
2456 /* These are specials */
2457
2458 case PT_ALNUM:
2459 return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||
2460 _pcre_ucp_gentype[prop->chartype] == ucp_N) == negated;
2461
2462 case PT_SPACE: /* Perl space */
2463 return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
2464 c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2465 == negated;
2466
2467 case PT_PXSPACE: /* POSIX space */
2468 return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
2469 c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2470 c == CHAR_FF || c == CHAR_CR)
2471 == negated;
2472
2473 case PT_WORD:
2474 return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||
2475 _pcre_ucp_gentype[prop->chartype] == ucp_N ||
2476 c == CHAR_UNDERSCORE) == negated;
2477 }
2478 return FALSE;
2479 }
2480 #endif /* SUPPORT_UCP */
2481
2482
2483
2484 /*************************************************
2485 * Check if auto-possessifying is possible *
2486 *************************************************/
2487
2488 /* This function is called for unlimited repeats of certain items, to see
2489 whether the next thing could possibly match the repeated item. If not, it makes
2490 sense to automatically possessify the repeated item.
2491
2492 Arguments:
2493 previous pointer to the repeated opcode
2494 utf8 TRUE in UTF-8 mode
2495 ptr next character in pattern
2496 options options bits
2497 cd contains pointers to tables etc.
2498
2499 Returns: TRUE if possessifying is wanted
2500 */
2501
2502 static BOOL
2503 check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr,
2504 int options, compile_data *cd)
2505 {
2506 int c, next;
2507 int op_code = *previous++;
2508
2509 /* Skip whitespace and comments in extended mode */
2510
2511 if ((options & PCRE_EXTENDED) != 0)
2512 {
2513 for (;;)
2514 {
2515 while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
2516 if (*ptr == CHAR_NUMBER_SIGN)
2517 {
2518 while (*(++ptr) != 0)
2519 if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2520 }
2521 else break;
2522 }
2523 }
2524
2525 /* If the next item is one that we can handle, get its value. A non-negative
2526 value is a character, a negative value is an escape value. */
2527
2528 if (*ptr == CHAR_BACKSLASH)
2529 {
2530 int temperrorcode = 0;
2531 next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE);
2532 if (temperrorcode != 0) return FALSE;
2533 ptr++; /* Point after the escape sequence */
2534 }
2535
2536 else if ((cd->ctypes[*ptr] & ctype_meta) == 0)
2537 {
2538 #ifdef SUPPORT_UTF8
2539 if (utf8) { GETCHARINC(next, ptr); } else
2540 #endif
2541 next = *ptr++;
2542 }
2543
2544 else return FALSE;
2545
2546 /* Skip whitespace and comments in extended mode */
2547
2548 if ((options & PCRE_EXTENDED) != 0)
2549 {
2550 for (;;)
2551 {
2552 while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
2553 if (*ptr == CHAR_NUMBER_SIGN)
2554 {
2555 while (*(++ptr) != 0)
2556 if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2557 }
2558 else break;
2559 }
2560 }
2561
2562 /* If the next thing is itself optional, we have to give up. */
2563
2564 if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
2565 strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
2566 return FALSE;
2567
2568 /* Now compare the next item with the previous opcode. First, handle cases when
2569 the next item is a character. */
2570
2571 if (next >= 0) switch(op_code)
2572 {
2573 case OP_CHAR:
2574 #ifdef SUPPORT_UTF8
2575 GETCHARTEST(c, previous);
2576 #else
2577 c = *previous;
2578 #endif
2579 return c != next;
2580
2581 /* For CHARNC (caseless character) we must check the other case. If we have
2582 Unicode property support, we can use it to test the other case of
2583 high-valued characters. */
2584
2585 case OP_CHARNC:
2586 #ifdef SUPPORT_UTF8
2587 GETCHARTEST(c, previous);
2588 #else
2589 c = *previous;
2590 #endif
2591 if (c == next) return FALSE;
2592 #ifdef SUPPORT_UTF8
2593 if (utf8)
2594 {
2595 unsigned int othercase;
2596 if (next < 128) othercase = cd->fcc[next]; else
2597 #ifdef SUPPORT_UCP
2598 othercase = UCD_OTHERCASE((unsigned int)next);
2599 #else
2600 othercase = NOTACHAR;
2601 #endif
2602 return (unsigned int)c != othercase;
2603 }
2604 else
2605 #endif /* SUPPORT_UTF8 */
2606 return (c != cd->fcc[next]); /* Non-UTF-8 mode */
2607
2608 /* For OP_NOT, its data is always a single-byte character. */
2609
2610 case OP_NOT:
2611 if ((c = *previous) == next) return TRUE;
2612 if ((options & PCRE_CASELESS) == 0) return FALSE;
2613 #ifdef SUPPORT_UTF8
2614 if (utf8)
2615 {
2616 unsigned int othercase;
2617 if (next < 128) othercase = cd->fcc[next]; else
2618 #ifdef SUPPORT_UCP
2619 othercase = UCD_OTHERCASE(next);
2620 #else
2621 othercase = NOTACHAR;
2622 #endif
2623 return (unsigned int)c == othercase;
2624 }
2625 else
2626 #endif /* SUPPORT_UTF8 */
2627 return (c == cd->fcc[next]); /* Non-UTF-8 mode */
2628
2629 /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
2630 When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
2631
2632 case OP_DIGIT:
2633 return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;
2634
2635 case OP_NOT_DIGIT:
2636 return next <= 127 && (cd->ctypes[next] & ctype_digit) != 0;
2637
2638 case OP_WHITESPACE:
2639 return next > 127 || (cd->ctypes[next] & ctype_space) == 0;
2640
2641 case OP_NOT_WHITESPACE:
2642 return next <= 127 && (cd->ctypes[next] & ctype_space) != 0;
2643
2644 case OP_WORDCHAR:
2645 return next > 127 || (cd->ctypes[next] & ctype_word) == 0;
2646
2647 case OP_NOT_WORDCHAR:
2648 return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;
2649
2650 case OP_HSPACE:
2651 case OP_NOT_HSPACE:
2652 switch(next)
2653 {
2654 case 0x09:
2655 case 0x20:
2656 case 0xa0:
2657 case 0x1680:
2658 case 0x180e:
2659 case 0x2000:
2660 case 0x2001:
2661 case 0x2002:
2662 case 0x2003:
2663 case 0x2004:
2664 case 0x2005:
2665 case 0x2006:
2666 case 0x2007:
2667 case 0x2008:
2668 case 0x2009:
2669 case 0x200A:
2670 case 0x202f:
2671 case 0x205f:
2672 case 0x3000:
2673 return op_code == OP_NOT_HSPACE;
2674 default:
2675 return op_code != OP_NOT_HSPACE;
2676 }
2677
2678 case OP_ANYNL:
2679 case OP_VSPACE:
2680 case OP_NOT_VSPACE:
2681 switch(next)
2682 {
2683 case 0x0a:
2684 case 0x0b:
2685 case 0x0c:
2686 case 0x0d:
2687 case 0x85:
2688 case 0x2028:
2689 case 0x2029:
2690 return op_code == OP_NOT_VSPACE;
2691 default:
2692 return op_code != OP_NOT_VSPACE;
2693 }
2694
2695 #ifdef SUPPORT_UCP
2696 case OP_PROP:
2697 return check_char_prop(next, previous[0], previous[1], FALSE);
2698
2699 case OP_NOTPROP:
2700 return check_char_prop(next, previous[0], previous[1], TRUE);
2701 #endif
2702
2703 default:
2704 return FALSE;
2705 }
2706
2707
2708 /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
2709 is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
2710 generated only when PCRE_UCP is *not* set, that is, when only ASCII
2711 characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are
2712 replaced by OP_PROP codes when PCRE_UCP is set. */
2713
2714 switch(op_code)
2715 {
2716 case OP_CHAR:
2717 case OP_CHARNC:
2718 #ifdef SUPPORT_UTF8
2719 GETCHARTEST(c, previous);
2720 #else
2721 c = *previous;
2722 #endif
2723 switch(-next)
2724 {
2725 case ESC_d:
2726 return c > 127 || (cd->ctypes[c] & ctype_digit) == 0;
2727
2728 case ESC_D:
2729 return c <= 127 && (cd->ctypes[c] & ctype_digit) != 0;
2730
2731 case ESC_s:
2732 return c > 127 || (cd->ctypes[c] & ctype_space) == 0;
2733
2734 case ESC_S:
2735 return c <= 127 && (cd->ctypes[c] & ctype_space) != 0;
2736
2737 case ESC_w:
2738 return c > 127 || (cd->ctypes[c] & ctype_word) == 0;
2739
2740 case ESC_W:
2741 return c <= 127 && (cd->ctypes[c] & ctype_word) != 0;
2742
2743 case ESC_h:
2744 case ESC_H:
2745 switch(c)
2746 {
2747 case 0x09:
2748 case 0x20:
2749 case 0xa0:
2750 case 0x1680:
2751 case 0x180e:
2752 case 0x2000:
2753 case 0x2001:
2754 case 0x2002:
2755 case 0x2003:
2756 case 0x2004:
2757 case 0x2005:
2758 case 0x2006:
2759 case 0x2007:
2760 case 0x2008:
2761 case 0x2009:
2762 case 0x200A:
2763 case 0x202f:
2764 case 0x205f:
2765 case 0x3000:
2766 return -next != ESC_h;
2767 default:
2768 return -next == ESC_h;
2769 }
2770
2771 case ESC_v:
2772 case ESC_V:
2773 switch(c)
2774 {
2775 case 0x0a:
2776 case 0x0b:
2777 case 0x0c:
2778 case 0x0d:
2779 case 0x85:
2780 case 0x2028:
2781 case 0x2029:
2782 return -next != ESC_v;
2783 default:
2784 return -next == ESC_v;
2785 }
2786
2787 /* When PCRE_UCP is set, these values get generated for \d etc. Find
2788 their substitutions and process them. The result will always be either
2789 -ESC_p or -ESC_P. Then fall through to process those values. */
2790
2791 #ifdef SUPPORT_UCP
2792 case ESC_du:
2793 case ESC_DU:
2794 case ESC_wu:
2795 case ESC_WU:
2796 case ESC_su:
2797 case ESC_SU:
2798 {
2799 int temperrorcode = 0;
2800 ptr = substitutes[-next - ESC_DU];
2801 next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);
2802 if (temperrorcode != 0) return FALSE;
2803 ptr++; /* For compatibility */
2804 }
2805 /* Fall through */
2806
2807 case ESC_p:
2808 case ESC_P:
2809 {
2810 int ptype, pdata, errorcodeptr;
2811 BOOL negated;
2812
2813 ptr--; /* Make ptr point at the p or P */
2814 ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);
2815 if (ptype < 0) return FALSE;
2816 ptr++; /* Point past the final curly ket */
2817
2818 /* If the property item is optional, we have to give up. (When generated
2819 from \d etc by PCRE_UCP, this test will have been applied much earlier,
2820 to the original \d etc. At this point, ptr will point to a zero byte. */
2821
2822 if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
2823 strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
2824 return FALSE;
2825
2826 /* Do the property check. */
2827
2828 return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);
2829 }
2830 #endif
2831
2832 default:
2833 return FALSE;
2834 }
2835
2836 /* In principle, support for Unicode properties should be integrated here as
2837 well. It means re-organizing the above code so as to get hold of the property
2838 values before switching on the op-code. However, I wonder how many patterns
2839 combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set,
2840 these op-codes are never generated.) */
2841
2842 case OP_DIGIT:
2843 return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
2844 next == -ESC_h || next == -ESC_v || next == -ESC_R;
2845
2846 case OP_NOT_DIGIT:
2847 return next == -ESC_d;
2848
2849 case OP_WHITESPACE:
2850 return next == -ESC_S || next == -ESC_d || next == -ESC_w || next == -ESC_R;
2851
2852 case OP_NOT_WHITESPACE:
2853 return next == -ESC_s || next == -ESC_h || next == -ESC_v;
2854
2855 case OP_HSPACE:
2856 return next == -ESC_S || next == -ESC_H || next == -ESC_d ||
2857 next == -ESC_w || next == -ESC_v || next == -ESC_R;
2858
2859 case OP_NOT_HSPACE:
2860 return next == -ESC_h;
2861
2862 /* Can't have \S in here because VT matches \S (Perl anomaly) */
2863 case OP_ANYNL:
2864 case OP_VSPACE:
2865 return next == -ESC_V || next == -ESC_d || next == -ESC_w;
2866
2867 case OP_NOT_VSPACE:
2868 return next == -ESC_v || next == -ESC_R;
2869
2870 case OP_WORDCHAR:
2871 return next == -ESC_W || next == -ESC_s || next == -ESC_h ||
2872 next == -ESC_v || next == -ESC_R;
2873
2874 case OP_NOT_WORDCHAR:
2875 return next == -ESC_w || next == -ESC_d;
2876
2877 default:
2878 return FALSE;
2879 }
2880
2881 /* Control does not reach here */
2882 }
2883
2884
2885
2886 /*************************************************
2887 * Compile one branch *
2888 *************************************************/
2889
2890 /* Scan the pattern, compiling it into the a vector. If the options are
2891 changed during the branch, the pointer is used to change the external options
2892 bits. This function is used during the pre-compile phase when we are trying
2893 to find out the amount of memory needed, as well as during the real compile
2894 phase. The value of lengthptr distinguishes the two phases.
2895
2896 Arguments:
2897 optionsptr pointer to the option bits
2898 codeptr points to the pointer to the current code point
2899 ptrptr points to the current pattern pointer
2900 errorcodeptr points to error code variable
2901 firstbyteptr set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)
2902 reqbyteptr set to the last literal character required, else < 0
2903 bcptr points to current branch chain
2904 cd contains pointers to tables etc.
2905 lengthptr NULL during the real compile phase
2906 points to length accumulator during pre-compile phase
2907
2908 Returns: TRUE on success
2909 FALSE, with *errorcodeptr set non-zero on error
2910 */
2911
2912 static BOOL
2913 compile_branch(int *optionsptr, uschar **codeptr, const uschar **ptrptr,
2914 int *errorcodeptr, int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr,
2915 compile_data *cd, int *lengthptr)
2916 {
2917 int repeat_type, op_type;
2918 int repeat_min = 0, repeat_max = 0; /* To please picky compilers */
2919 int bravalue = 0;
2920 int greedy_default, greedy_non_default;
2921 int firstbyte, reqbyte;
2922 int zeroreqbyte, zerofirstbyte;
2923 int req_caseopt, reqvary, tempreqvary;
2924 int options = *optionsptr;
2925 int after_manual_callout = 0;
2926 int length_prevgroup = 0;
2927 register int c;
2928 register uschar *code = *codeptr;
2929 uschar *last_code = code;
2930 uschar *orig_code = code;
2931 uschar *tempcode;
2932 BOOL inescq = FALSE;
2933 BOOL groupsetfirstbyte = FALSE;
2934 const uschar *ptr = *ptrptr;
2935 const uschar *tempptr;
2936 const uschar *nestptr = NULL;
2937 uschar *previous = NULL;
2938 uschar *previous_callout = NULL;
2939 uschar *save_hwm = NULL;
2940 uschar classbits[32];
2941
2942 #ifdef SUPPORT_UTF8
2943 BOOL class_utf8;
2944 BOOL utf8 = (options & PCRE_UTF8) != 0;
2945 uschar *class_utf8data;
2946 uschar *class_utf8data_base;
2947 uschar utf8_char[6];
2948 #else
2949 BOOL utf8 = FALSE;
2950 uschar *utf8_char = NULL;
2951 #endif
2952
2953 #ifdef PCRE_DEBUG
2954 if (lengthptr != NULL) DPRINTF((">> start branch\n"));
2955 #endif
2956
2957 /* Set up the default and non-default settings for greediness */
2958
2959 greedy_default = ((options & PCRE_UNGREEDY) != 0);
2960 greedy_non_default = greedy_default ^ 1;
2961
2962 /* Initialize no first byte, no required byte. REQ_UNSET means "no char
2963 matching encountered yet". It gets changed to REQ_NONE if we hit something that
2964 matches a non-fixed char first char; reqbyte just remains unset if we never
2965 find one.
2966
2967 When we hit a repeat whose minimum is zero, we may have to adjust these values
2968 to take the zero repeat into account. This is implemented by setting them to
2969 zerofirstbyte and zeroreqbyte when such a repeat is encountered. The individual
2970 item types that can be repeated set these backoff variables appropriately. */
2971
2972 firstbyte = reqbyte = zerofirstbyte = zeroreqbyte = REQ_UNSET;
2973
2974 /* The variable req_caseopt contains either the REQ_CASELESS value or zero,
2975 according to the current setting of the caseless flag. REQ_CASELESS is a bit
2976 value > 255. It is added into the firstbyte or reqbyte variables to record the
2977 case status of the value. This is used only for ASCII characters. */
2978
2979 req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;
2980
2981 /* Switch on next character until the end of the branch */
2982
2983 for (;; ptr++)
2984 {
2985 BOOL negate_class;
2986 BOOL should_flip_negation;
2987 BOOL possessive_quantifier;
2988 BOOL is_quantifier;
2989 BOOL is_recurse;
2990 BOOL reset_bracount;
2991 int class_charcount;
2992 int class_lastchar;
2993 int newoptions;
2994 int recno;
2995 int refsign;
2996 int skipbytes;
2997 int subreqbyte;
2998 int subfirstbyte;
2999 int terminator;
3000 int mclength;
3001 uschar mcbuffer[8];
3002
3003 /* Get next byte in the pattern */
3004
3005 c = *ptr;
3006
3007 /* If we are at the end of a nested substitution, revert to the outer level
3008 string. Nesting only happens one level deep. */
3009
3010 if (c == 0 && nestptr != NULL)
3011 {
3012 ptr = nestptr;
3013 nestptr = NULL;
3014 c = *ptr;
3015 }
3016
3017 /* If we are in the pre-compile phase, accumulate the length used for the
3018 previous cycle of this loop. */
3019
3020 if (lengthptr != NULL)
3021 {
3022 #ifdef PCRE_DEBUG
3023 if (code > cd->hwm) cd->hwm = code; /* High water info */
3024 #endif
3025 if (code > cd->start_workspace + WORK_SIZE_CHECK) /* Check for overrun */
3026 {
3027 *errorcodeptr = ERR52;
3028 goto FAILED;
3029 }
3030
3031 /* There is at least one situation where code goes backwards: this is the
3032 case of a zero quantifier after a class (e.g. [ab]{0}). At compile time,
3033 the class is simply eliminated. However, it is created first, so we have to
3034 allow memory for it. Therefore, don't ever reduce the length at this point.
3035 */
3036
3037 if (code < last_code) code = last_code;
3038
3039 /* Paranoid check for integer overflow */
3040
3041 if (OFLOW_MAX - *lengthptr < code - last_code)
3042 {
3043 *errorcodeptr = ERR20;
3044 goto FAILED;
3045 }
3046
3047 *lengthptr += (int)(code - last_code);
3048 DPRINTF(("length=%d added %d c=%c\n", *lengthptr, code - last_code, c));
3049
3050 /* If "previous" is set and it is not at the start of the work space, move
3051 it back to there, in order to avoid filling up the work space. Otherwise,
3052 if "previous" is NULL, reset the current code pointer to the start. */
3053
3054 if (previous != NULL)
3055 {
3056 if (previous > orig_code)
3057 {
3058 memmove(orig_code, previous, code - previous);
3059 code -= previous - orig_code;
3060 previous = orig_code;
3061 }
3062 }
3063 else code = orig_code;
3064
3065 /* Remember where this code item starts so we can pick up the length
3066 next time round. */
3067
3068 last_code = code;
3069 }
3070
3071 /* In the real compile phase, just check the workspace used by the forward
3072 reference list. */
3073
3074 else if (cd->hwm > cd->start_workspace + WORK_SIZE_CHECK)
3075 {
3076 *errorcodeptr = ERR52;
3077 goto FAILED;
3078 }
3079
3080 /* If in \Q...\E, check for the end; if not, we have a literal */
3081
3082 if (inescq && c != 0)
3083 {
3084 if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)
3085 {
3086 inescq = FALSE;
3087 ptr++;
3088 continue;
3089 }
3090 else
3091 {
3092 if (previous_callout != NULL)
3093 {
3094 if (lengthptr == NULL) /* Don't attempt in pre-compile phase */
3095 complete_callout(previous_callout, ptr, cd);
3096 previous_callout = NULL;
3097 }
3098 if ((options & PCRE_AUTO_CALLOUT) != 0)
3099 {
3100 previous_callout = code;
3101 code = auto_callout(code, ptr, cd);
3102 }
3103 goto NORMAL_CHAR;
3104 }
3105 }
3106
3107 /* Fill in length of a previous callout, except when the next thing is
3108 a quantifier. */
3109
3110 is_quantifier =
3111 c == CHAR_ASTERISK || c == CHAR_PLUS || c == CHAR_QUESTION_MARK ||
3112 (c == CHAR_LEFT_CURLY_BRACKET && is_counted_repeat(ptr+1));
3113
3114 if (!is_quantifier && previous_callout != NULL &&
3115 after_manual_callout-- <= 0)
3116 {
3117 if (lengthptr == NULL) /* Don't attempt in pre-compile phase */
3118 complete_callout(previous_callout, ptr, cd);
3119 previous_callout = NULL;
3120 }
3121
3122 /* In extended mode, skip white space and comments */
3123
3124 if ((options & PCRE_EXTENDED) != 0)
3125 {
3126 if ((cd->ctypes[c] & ctype_space) != 0) continue;
3127 if (c == CHAR_NUMBER_SIGN)
3128 {
3129 while (*(++ptr) != 0)
3130 {
3131 if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
3132 }
3133 if (*ptr != 0) continue;
3134
3135 /* Else fall through to handle end of string */
3136 c = 0;
3137 }
3138 }
3139
3140 /* No auto callout for quantifiers. */
3141
3142 if ((options & PCRE_AUTO_CALLOUT) != 0 && !is_quantifier)
3143 {
3144 previous_callout = code;
3145 code = auto_callout(code, ptr, cd);
3146 }
3147
3148 switch(c)
3149 {
3150 /* ===================================================================*/
3151 case 0: /* The branch terminates at string end */
3152 case CHAR_VERTICAL_LINE: /* or | or ) */
3153 case CHAR_RIGHT_PARENTHESIS:
3154 *firstbyteptr = firstbyte;
3155 *reqbyteptr = reqbyte;
3156 *codeptr = code;
3157 *ptrptr = ptr;
3158 if (lengthptr != NULL)
3159 {
3160 if (OFLOW_MAX - *lengthptr < code - last_code)
3161 {
3162 *errorcodeptr = ERR20;
3163 goto FAILED;
3164 }
3165 *lengthptr += (int)(code - last_code); /* To include callout length */
3166 DPRINTF((">> end branch\n"));
3167 }
3168 return TRUE;
3169
3170
3171 /* ===================================================================*/
3172 /* Handle single-character metacharacters. In multiline mode, ^ disables
3173 the setting of any following char as a first character. */
3174
3175 case CHAR_CIRCUMFLEX_ACCENT:
3176 if ((options & PCRE_MULTILINE) != 0)
3177 {
3178 if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
3179 }
3180 previous = NULL;
3181 *code++ = OP_CIRC;
3182 break;
3183
3184 case CHAR_DOLLAR_SIGN:
3185 previous = NULL;
3186 *code++ = OP_DOLL;
3187 break;
3188
3189 /* There can never be a first char if '.' is first, whatever happens about
3190 repeats. The value of reqbyte doesn't change either. */
3191
3192 case CHAR_DOT:
3193 if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
3194 zerofirstbyte = firstbyte;
3195 zeroreqbyte = reqbyte;
3196 previous = code;
3197 *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
3198 break;
3199
3200
3201 /* ===================================================================*/
3202 /* Character classes. If the included characters are all < 256, we build a
3203 32-byte bitmap of the permitted characters, except in the special case
3204 where there is only one such character. For negated classes, we build the
3205 map as usual, then invert it at the end. However, we use a different opcode
3206 so that data characters > 255 can be handled correctly.
3207
3208 If the class contains characters outside the 0-255 range, a different
3209 opcode is compiled. It may optionally have a bit map for characters < 256,
3210 but those above are are explicitly listed afterwards. A flag byte tells
3211 whether the bitmap is present, and whether this is a negated class or not.
3212
3213 In JavaScript compatibility mode, an isolated ']' causes an error. In
3214 default (Perl) mode, it is treated as a data character. */
3215
3216 case CHAR_RIGHT_SQUARE_BRACKET:
3217 if ((cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
3218 {
3219 *errorcodeptr = ERR64;
3220 goto FAILED;
3221 }
3222 goto NORMAL_CHAR;
3223
3224 case CHAR_LEFT_SQUARE_BRACKET:
3225 previous = code;
3226
3227 /* PCRE supports POSIX class stuff inside a class. Perl gives an error if
3228 they are encountered at the top level, so we'll do that too. */
3229
3230 if ((ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT ||
3231 ptr[1] == CHAR_EQUALS_SIGN) &&
3232 check_posix_syntax(ptr, &tempptr))
3233 {
3234 *errorcodeptr = (ptr[1] == CHAR_COLON)? ERR13 : ERR31;
3235 goto FAILED;
3236 }
3237
3238 /* If the first character is '^', set the negation flag and skip it. Also,
3239 if the first few characters (either before or after ^) are \Q\E or \E we
3240 skip them too. This makes for compatibility with Perl. */
3241
3242 negate_class = FALSE;
3243 for (;;)
3244 {
3245 c = *(++ptr);
3246 if (c == CHAR_BACKSLASH)
3247 {
3248 if (ptr[1] == CHAR_E)
3249 ptr++;
3250 else if (strncmp((const char *)ptr+1,
3251 STR_Q STR_BACKSLASH STR_E, 3) == 0)
3252 ptr += 3;
3253 else
3254 break;
3255 }
3256 else if (!negate_class && c == CHAR_CIRCUMFLEX_ACCENT)
3257 negate_class = TRUE;
3258 else break;
3259 }
3260
3261 /* Empty classes are allowed in JavaScript compatibility mode. Otherwise,
3262 an initial ']' is taken as a data character -- the code below handles
3263 that. In JS mode, [] must always fail, so generate OP_FAIL, whereas
3264 [^] must match any character, so generate OP_ALLANY. */
3265
3266 if (c == CHAR_RIGHT_SQUARE_BRACKET &&
3267 (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
3268 {
3269 *code++ = negate_class? OP_ALLANY : OP_FAIL;
3270 if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
3271 zerofirstbyte = firstbyte;
3272 break;
3273 }
3274
3275 /* If a class contains a negative special such as \S, we need to flip the
3276 negation flag at the end, so that support for characters > 255 works
3277 correctly (they are all included in the class). */
3278
3279 should_flip_negation = FALSE;
3280
3281 /* Keep a count of chars with values < 256 so that we can optimize the case
3282 of just a single character (as long as it's < 256). However, For higher
3283 valued UTF-8 characters, we don't yet do any optimization. */
3284
3285 class_charcount = 0;
3286 class_lastchar = -1;
3287
3288 /* Initialize the 32-char bit map to all zeros. We build the map in a
3289 temporary bit of memory, in case the class contains only 1 character (less
3290 than 256), because in that case the compiled code doesn't use the bit map.
3291 */
3292
3293 memset(classbits, 0, 32 * sizeof(uschar));
3294
3295 #ifdef SUPPORT_UTF8
3296 class_utf8 = FALSE; /* No chars >= 256 */
3297 class_utf8data = code + LINK_SIZE + 2; /* For UTF-8 items */
3298 class_utf8data_base = class_utf8data; /* For resetting in pass 1 */
3299 #endif
3300
3301 /* Process characters until ] is reached. By writing this as a "do" it
3302 means that an initial ] is taken as a data character. At the start of the
3303 loop, c contains the first byte of the character. */
3304
3305 if (c != 0) do
3306 {
3307 const uschar *oldptr;
3308
3309 #ifdef SUPPORT_UTF8
3310 if (utf8 && c > 127)
3311 { /* Braces are required because the */
3312 GETCHARLEN(c, ptr, ptr); /* macro generates multiple statements */
3313 }
3314
3315 /* In the pre-compile phase, accumulate the length of any UTF-8 extra
3316 data and reset the pointer. This is so that very large classes that
3317 contain a zillion UTF-8 characters no longer overwrite the work space
3318 (which is on the stack). */
3319
3320 if (lengthptr != NULL)
3321 {
3322 *lengthptr += class_utf8data - class_utf8data_base;
3323 class_utf8data = class_utf8data_base;
3324 }
3325
3326 #endif
3327
3328 /* Inside \Q...\E everything is literal except \E */
3329
3330 if (inescq)
3331 {
3332 if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E) /* If we are at \E */
3333 {
3334 inescq = FALSE; /* Reset literal state */
3335 ptr++; /* Skip the 'E' */
3336 continue; /* Carry on with next */
3337 }
3338 goto CHECK_RANGE; /* Could be range if \E follows */
3339 }
3340
3341 /* Handle POSIX class names. Perl allows a negation extension of the
3342 form [:^name:]. A square bracket that doesn't match the syntax is
3343 treated as a literal. We also recognize the POSIX constructions
3344 [.ch.] and [=ch=] ("collating elements") and fault them, as Perl
3345 5.6 and 5.8 do. */
3346
3347 if (c == CHAR_LEFT_SQUARE_BRACKET &&
3348 (ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT ||
3349 ptr[1] == CHAR_EQUALS_SIGN) && check_posix_syntax(ptr, &tempptr))
3350 {
3351 BOOL local_negate = FALSE;
3352 int posix_class, taboffset, tabopt;
3353 register const uschar *cbits = cd->cbits;
3354 uschar pbits[32];
3355
3356 if (ptr[1] != CHAR_COLON)
3357 {
3358 *errorcodeptr = ERR31;
3359 goto FAILED;
3360 }
3361
3362 ptr += 2;
3363 if (*ptr == CHAR_CIRCUMFLEX_ACCENT)
3364 {
3365 local_negate = TRUE;
3366 should_flip_negation = TRUE; /* Note negative special */
3367 ptr++;
3368 }
3369
3370 posix_class = check_posix_name(ptr, (int)(tempptr - ptr));
3371 if (posix_class < 0)
3372 {
3373 *errorcodeptr = ERR30;
3374 goto FAILED;
3375 }
3376
3377 /* If matching is caseless, upper and lower are converted to
3378 alpha. This relies on the fact that the class table starts with
3379 alpha, lower, upper as the first 3 entries. */
3380
3381 if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
3382 posix_class = 0;
3383
3384 /* When PCRE_UCP is set, some of the POSIX classes are converted to
3385 different escape sequences that use Unicode properties. */
3386
3387 #ifdef SUPPORT_UCP
3388 if ((options & PCRE_UCP) != 0)
3389 {
3390 int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);
3391 if (posix_substitutes[pc] != NULL)
3392 {
3393 nestptr = tempptr + 1;
3394 ptr = posix_substitutes[pc] - 1;
3395 continue;
3396 }
3397 }
3398 #endif
3399 /* In the non-UCP case, we build the bit map for the POSIX class in a
3400 chunk of local store because we may be adding and subtracting from it,
3401 and we don't want to subtract bits that may be in the main map already.
3402 At the end we or the result into the bit map that is being built. */
3403
3404 posix_class *= 3;
3405
3406 /* Copy in the first table (always present) */
3407
3408 memcpy(pbits, cbits + posix_class_maps[posix_class],
3409 32 * sizeof(uschar));
3410
3411 /* If there is a second table, add or remove it as required. */
3412
3413 taboffset = posix_class_maps[posix_class + 1];
3414 tabopt = posix_class_maps[posix_class + 2];
3415
3416 if (taboffset >= 0)
3417 {
3418 if (tabopt >= 0)
3419 for (c = 0; c < 32; c++) pbits[c] |= cbits[c + taboffset];
3420 else
3421 for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset];
3422 }
3423
3424 /* Not see if we need to remove any special characters. An option
3425 value of 1 removes vertical space and 2 removes underscore. */
3426
3427 if (tabopt < 0) tabopt = -tabopt;
3428 if (tabopt == 1) pbits[1] &= ~0x3c;
3429 else if (tabopt == 2) pbits[11] &= 0x7f;
3430
3431 /* Add the POSIX table or its complement into the main table that is
3432 being built and we are done. */
3433
3434 if (local_negate)
3435 for (c = 0; c < 32; c++) classbits[c] |= ~pbits[c];
3436 else
3437 for (c = 0; c < 32; c++) classbits[c] |= pbits[c];
3438
3439 ptr = tempptr + 1;
3440 class_charcount = 10; /* Set > 1; assumes more than 1 per class */
3441 continue; /* End of POSIX syntax handling */
3442 }
3443
3444 /* Backslash may introduce a single character, or it may introduce one
3445 of the specials, which just set a flag. The sequence \b is a special
3446 case. Inside a class (and only there) it is treated as backspace. We
3447 assume that other escapes have more than one character in them, so set
3448 class_charcount bigger than one. Unrecognized escapes fall through and
3449 are either treated as literal characters (by default), or are faulted if
3450 PCRE_EXTRA is set. */
3451
3452 if (c == CHAR_BACKSLASH)
3453 {
3454 c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);
3455 if (*errorcodeptr != 0) goto FAILED;
3456
3457 if (-c == ESC_b) c = CHAR_BS; /* \b is backspace in a class */
3458 else if (-c == ESC_Q) /* Handle start of quoted string */
3459 {
3460 if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
3461 {
3462 ptr += 2; /* avoid empty string */
3463 }
3464 else inescq = TRUE;
3465 continue;
3466 }
3467 else if (-c == ESC_E) continue; /* Ignore orphan \E */
3468
3469 if (c < 0)
3470 {
3471 register const uschar *cbits = cd->cbits;
3472 class_charcount += 2; /* Greater than 1 is what matters */
3473
3474 switch (-c)
3475 {
3476 #ifdef SUPPORT_UCP
3477 case ESC_du: /* These are the values given for \d etc */
3478 case ESC_DU: /* when PCRE_UCP is set. We replace the */
3479 case ESC_wu: /* escape sequence with an appropriate \p */
3480 case ESC_WU: /* or \P to test Unicode properties instead */
3481 case ESC_su: /* of the default ASCII testing. */
3482 case ESC_SU:
3483 nestptr = ptr;
3484 ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */
3485 class_charcount -= 2; /* Undo! */
3486 continue;
3487 #endif
3488 case ESC_d:
3489 for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit];
3490 continue;
3491
3492 case ESC_D:
3493 should_flip_negation = TRUE;
3494 for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit];
3495 continue;
3496
3497 case ESC_w:
3498 for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_word];
3499 continue;
3500
3501 case ESC_W:
3502 should_flip_negation = TRUE;
3503 for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];
3504 continue;
3505
3506 case ESC_s:
3507 for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space];
3508 classbits[1] &= ~0x08; /* Perl 5.004 onwards omits VT from \s */
3509 continue;
3510
3511 case ESC_S:
3512 should_flip_negation = TRUE;
3513 for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];
3514 classbits[1] |= 0x08; /* Perl 5.004 onwards omits VT from \s */
3515 continue;
3516
3517 case ESC_h:
3518 SETBIT(classbits, 0x09); /* VT */
3519 SETBIT(classbits, 0x20); /* SPACE */
3520 SETBIT(classbits, 0xa0); /* NSBP */
3521 #ifdef SUPPORT_UTF8
3522 if (utf8)
3523 {
3524 class_utf8 = TRUE;
3525 *class_utf8data++ = XCL_SINGLE;
3526 class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);
3527 *class_utf8data++ = XCL_SINGLE;
3528 class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);
3529 *class_utf8data++ = XCL_RANGE;
3530 class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);
3531 class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);
3532 *class_utf8data++ = XCL_SINGLE;
3533 class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);
3534 *class_utf8data++ = XCL_SINGLE;
3535 class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);
3536 *class_utf8data++ = XCL_SINGLE;
3537 class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);
3538 }
3539 #endif
3540 continue;
3541
3542 case ESC_H:
3543 for (c = 0; c < 32; c++)
3544 {
3545 int x = 0xff;
3546 switch (c)
3547 {
3548 case 0x09/8: x ^= 1 << (0x09%8); break;
3549 case 0x20/8: x ^= 1 << (0x20%8); break;
3550 case 0xa0/8: x ^= 1 << (0xa0%8); break;
3551 default: break;
3552 }
3553 classbits[c] |= x;
3554 }
3555
3556 #ifdef SUPPORT_UTF8
3557 if (utf8)
3558 {
3559 class_utf8 = TRUE;
3560 *class_utf8data++ = XCL_RANGE;
3561 class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
3562 class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);
3563 *class_utf8data++ = XCL_RANGE;
3564 class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);
3565 class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);
3566 *class_utf8data++ = XCL_RANGE;
3567 class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);
3568 class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);
3569 *class_utf8data++ = XCL_RANGE;
3570 class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);
3571 class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);
3572 *class_utf8data++ = XCL_RANGE;
3573 class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);
3574 class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);
3575 *class_utf8data++ = XCL_RANGE;
3576 class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);
3577 class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);
3578 *class_utf8data++ = XCL_RANGE;
3579 class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);
3580 class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
3581 }
3582 #endif
3583 continue;
3584
3585 case ESC_v:
3586 SETBIT(classbits, 0x0a); /* LF */
3587 SETBIT(classbits, 0x0b); /* VT */
3588 SETBIT(classbits, 0x0c); /* FF */
3589 SETBIT(classbits, 0x0d); /* CR */
3590 SETBIT(classbits, 0x85); /* NEL */
3591 #ifdef SUPPORT_UTF8
3592 if (utf8)
3593 {
3594 class_utf8 = TRUE;
3595 *class_utf8data++ = XCL_RANGE;
3596 class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);
3597 class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
3598 }
3599 #endif
3600 continue;
3601
3602 case ESC_V:
3603 for (c = 0; c < 32; c++)
3604 {
3605 int x = 0xff;
3606 switch (c)
3607 {
3608 case 0x0a/8: x ^= 1 << (0x0a%8);
3609 x ^= 1 << (0x0b%8);
3610 x ^= 1 << (0x0c%8);
3611 x ^= 1 << (0x0d%8);
3612 break;
3613 case 0x85/8: x ^= 1 << (0x85%8); break;
3614 default: break;
3615 }
3616 classbits[c] |= x;
3617 }
3618
3619 #ifdef SUPPORT_UTF8
3620 if (utf8)
3621 {
3622 class_utf8 = TRUE;
3623 *class_utf8data++ = XCL_RANGE;
3624 class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
3625 class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);
3626 *class_utf8data++ = XCL_RANGE;
3627 class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
3628 class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
3629 }
3630 #endif
3631 continue;
3632
3633 #ifdef SUPPORT_UCP
3634 case ESC_p:
3635 case ESC_P:
3636 {
3637 BOOL negated;
3638 int pdata;
3639 int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
3640 if (ptype < 0) goto FAILED;
3641 class_utf8 = TRUE;
3642 *class_utf8data++ = ((-c == ESC_p) != negated)?
3643 XCL_PROP : XCL_NOTPROP;
3644 *class_utf8data++ = ptype;
3645 *class_utf8data++ = pdata;
3646 class_charcount -= 2; /* Not a < 256 character */
3647 continue;
3648 }
3649 #endif
3650 /* Unrecognized escapes are faulted if PCRE is running in its
3651 strict mode. By default, for compatibility with Perl, they are
3652 treated as literals. */
3653
3654 default:
3655 if ((options & PCRE_EXTRA) != 0)
3656 {
3657 *errorcodeptr = ERR7;
3658 goto FAILED;
3659 }
3660 class_charcount -= 2; /* Undo the default count from above */
3661 c = *ptr; /* Get the final character and fall through */
3662 break;
3663 }
3664 }
3665
3666 /* Fall through if we have a single character (c >= 0). This may be
3667 greater than 256 in UTF-8 mode. */
3668
3669 } /* End of backslash handling */
3670
3671 /* A single character may be followed by '-' to form a range. However,
3672 Perl does not permit ']' to be the end of the range. A '-' character
3673 at the end is treated as a literal. Perl ignores orphaned \E sequences
3674 entirely. The code for handling \Q and \E is messy. */
3675
3676 CHECK_RANGE:
3677 while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
3678 {
3679 inescq = FALSE;
3680 ptr += 2;
3681 }
3682
3683 oldptr = ptr;
3684
3685 /* Remember \r or \n */
3686
3687 if (c == CHAR_CR || c == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;
3688
3689 /* Check for range */
3690
3691 if (!inescq && ptr[1] == CHAR_MINUS)
3692 {
3693 int d;
3694 ptr += 2;
3695 while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;
3696
3697 /* If we hit \Q (not followed by \E) at this point, go into escaped
3698 mode. */
3699
3700 while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_Q)
3701 {
3702 ptr += 2;
3703 if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E)
3704 { ptr += 2; continue; }
3705 inescq = TRUE;
3706 break;
3707 }
3708
3709 if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))
3710 {
3711 ptr = oldptr;
3712 goto LONE_SINGLE_CHARACTER;
3713 }
3714
3715 #ifdef SUPPORT_UTF8
3716 if (utf8)
3717 { /* Braces are required because the */
3718 GETCHARLEN(d, ptr, ptr); /* macro generates multiple statements */
3719 }
3720 else
3721 #endif
3722 d = *ptr; /* Not UTF-8 mode */
3723
3724 /* The second part of a range can be a single-character escape, but
3725 not any of the other escapes. Perl 5.6 treats a hyphen as a literal
3726 in such circumstances. */
3727
3728 if (!inescq && d == CHAR_BACKSLASH)
3729 {
3730 d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);
3731 if (*errorcodeptr != 0) goto FAILED;
3732
3733 /* \b is backspace; any other special means the '-' was literal */
3734
3735 if (d < 0)
3736 {
3737 if (d == -ESC_b) d = CHAR_BS; else
3738 {
3739 ptr = oldptr;
3740 goto LONE_SINGLE_CHARACTER; /* A few lines below */
3741 }
3742 }
3743 }
3744
3745 /* Check that the two values are in the correct order. Optimize
3746 one-character ranges */
3747
3748 if (d < c)
3749 {
3750 *errorcodeptr = ERR8;
3751 goto FAILED;
3752 }
3753
3754 if (d == c) goto LONE_SINGLE_CHARACTER; /* A few lines below */
3755
3756 /* Remember \r or \n */
3757
3758 if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;
3759
3760 /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless
3761 matching, we have to use an XCLASS with extra data items. Caseless
3762 matching for characters > 127 is available only if UCP support is
3763 available. */
3764
3765 #ifdef SUPPORT_UTF8
3766 if (utf8 && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))
3767 {
3768 class_utf8 = TRUE;
3769
3770 /* With UCP support, we can find the other case equivalents of
3771 the relevant characters. There may be several ranges. Optimize how
3772 they fit with the basic range. */
3773
3774 #ifdef SUPPORT_UCP
3775 if ((options & PCRE_CASELESS) != 0)
3776 {
3777 unsigned int occ, ocd;
3778 unsigned int cc = c;
3779 unsigned int origd = d;
3780 while (get_othercase_range(&cc, origd, &occ, &ocd))
3781 {
3782 if (occ >= (unsigned int)c &&
3783 ocd <= (unsigned int)d)
3784 continue; /* Skip embedded ranges */
3785
3786 if (occ < (unsigned int)c &&
3787 ocd >= (unsigned int)c - 1) /* Extend the basic range */
3788 { /* if there is overlap, */
3789 c = occ; /* noting that if occ < c */
3790 continue; /* we can't have ocd > d */
3791 } /* because a subrange is */
3792 if (ocd > (unsigned int)d &&
3793 occ <= (unsigned int)d + 1) /* always shorter than */
3794 { /* the basic range. */
3795 d = ocd;
3796 continue;
3797 }
3798
3799 if (occ == ocd)
3800 {
3801 *class_utf8data++ = XCL_SINGLE;
3802 }
3803 else
3804 {
3805 *class_utf8data++ = XCL_RANGE;
3806 class_utf8data += _pcre_ord2utf8(occ, class_utf8data);
3807 }
3808 class_utf8data += _pcre_ord2utf8(ocd, class_utf8data);
3809 }
3810 }
3811 #endif /* SUPPORT_UCP */
3812
3813 /* Now record the original range, possibly modified for UCP caseless
3814 overlapping ranges. */
3815
3816 *class_utf8data++ = XCL_RANGE;
3817 class_utf8data += _pcre_ord2utf8(c, class_utf8data);
3818 class_utf8data += _pcre_ord2utf8(d, class_utf8data);
3819
3820 /* With UCP support, we are done. Without UCP support, there is no
3821 caseless matching for UTF-8 characters > 127; we can use the bit map
3822 for the smaller ones. */
3823
3824 #ifdef SUPPORT_UCP
3825 continue; /* With next character in the class */
3826 #else
3827 if ((options & PCRE_CASELESS) == 0 || c > 127) continue;
3828
3829 /* Adjust upper limit and fall through to set up the map */
3830
3831 d = 127;
3832
3833 #endif /* SUPPORT_UCP */
3834 }
3835 #endif /* SUPPORT_UTF8 */
3836
3837 /* We use the bit map for all cases when not in UTF-8 mode; else
3838 ranges that lie entirely within 0-127 when there is UCP support; else
3839 for partial ranges without UCP support. */
3840
3841 class_charcount += d - c + 1;
3842 class_lastchar = d;
3843
3844 /* We can save a bit of time by skipping this in the pre-compile. */
3845
3846 if (lengthptr == NULL) for (; c <= d; c++)
3847 {
3848 classbits[c/8] |= (1 << (c&7));
3849 if ((options & PCRE_CASELESS) != 0)
3850 {
3851 int uc = cd->fcc[c]; /* flip case */
3852 classbits[uc/8] |= (1 << (uc&7));
3853 }
3854 }
3855
3856 continue; /* Go get the next char in the class */
3857 }
3858
3859 /* Handle a lone single character - we can get here for a normal
3860 non-escape char, or after \ that introduces a single character or for an
3861 apparent range that isn't. */
3862
3863 LONE_SINGLE_CHARACTER:
3864
3865 /* Handle a character that cannot go in the bit map */
3866
3867 #ifdef SUPPORT_UTF8
3868 if (utf8 && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))
3869 {
3870 class_utf8 = TRUE;
3871 *class_utf8data++ = XCL_SINGLE;
3872 class_utf8data += _pcre_ord2utf8(c, class_utf8data);
3873
3874 #ifdef SUPPORT_UCP
3875 if ((options & PCRE_CASELESS) != 0)
3876 {
3877 unsigned int othercase;
3878 if ((othercase = UCD_OTHERCASE(c)) != c)
3879 {
3880 *class_utf8data++ = XCL_SINGLE;
3881 class_utf8data += _pcre_ord2utf8(othercase, class_utf8data);
3882 }
3883 }
3884 #endif /* SUPPORT_UCP */
3885
3886 }
3887 else
3888 #endif /* SUPPORT_UTF8 */
3889
3890 /* Handle a single-byte character */
3891 {
3892 classbits[c/8] |= (1 << (c&7));
3893 if ((options & PCRE_CASELESS) != 0)
3894 {
3895 c = cd->fcc[c]; /* flip case */
3896 classbits[c/8] |= (1 << (c&7));
3897 }
3898 class_charcount++;
3899 class_lastchar = c;
3900 }
3901 }
3902
3903 /* Loop until ']' reached. This "while" is the end of the "do" far above.
3904 If we are at the end of an internal nested string, revert to the outer
3905 string. */
3906
3907 while (((c = *(++ptr)) != 0 ||
3908 (nestptr != NULL &&
3909 (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != 0)) &&
3910 (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));
3911
3912 /* Check for missing terminating ']' */
3913
3914 if (c == 0)
3915 {
3916 *errorcodeptr = ERR6;
3917 goto FAILED;
3918 }
3919
3920 /* If class_charcount is 1, we saw precisely one character whose value is
3921 less than 256. As long as there were no characters >= 128 and there was no
3922 use of \p or \P, in other words, no use of any XCLASS features, we can
3923 optimize.
3924
3925 In UTF-8 mode, we can optimize the negative case only if there were no
3926 characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR
3927 operate on single-bytes only. This is an historical hangover. Maybe one day
3928 we can tidy these opcodes to handle multi-byte characters.
3929
3930 The optimization throws away the bit map. We turn the item into a
3931 1-character OP_CHAR[NC] if it's positive, or OP_NOT if it's negative. Note
3932 that OP_NOT does not support multibyte characters. In the positive case, it
3933 can cause firstbyte to be set. Otherwise, there can be no first char if
3934 this item is first, whatever repeat count may follow. In the case of
3935 reqbyte, save the previous value for reinstating. */
3936
3937 #ifdef SUPPORT_UTF8
3938 if (class_charcount == 1 && !class_utf8 &&
3939 (!utf8 || !negate_class || class_lastchar < 128))
3940 #else
3941 if (class_charcount == 1)
3942 #endif
3943 {
3944 zeroreqbyte = reqbyte;
3945
3946 /* The OP_NOT opcode works on one-byte characters only. */
3947
3948 if (negate_class)
3949 {
3950 if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
3951 zerofirstbyte = firstbyte;
3952 *code++ = OP_NOT;
3953 *code++ = class_lastchar;
3954 break;
3955 }
3956
3957 /* For a single, positive character, get the value into mcbuffer, and
3958 then we can handle this with the normal one-character code. */
3959
3960 #ifdef SUPPORT_UTF8
3961 if (utf8 && class_lastchar > 127)
3962 mclength = _pcre_ord2utf8(class_lastchar, mcbuffer);
3963 else
3964 #endif
3965 {
3966 mcbuffer[0] = class_lastchar;
3967 mclength = 1;
3968 }
3969 goto ONE_CHAR;
3970 } /* End of 1-char optimization */
3971
3972 /* The general case - not the one-char optimization. If this is the first
3973 thing in the branch, there can be no first char setting, whatever the
3974 repeat count. Any reqbyte setting must remain unchanged after any kind of
3975 repeat. */
3976
3977 if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
3978 zerofirstbyte = firstbyte;
3979 zeroreqbyte = reqbyte;
3980
3981 /* If there are characters with values > 255, we have to compile an
3982 extended class, with its own opcode, unless there was a negated special
3983 such as \S in the class, and PCRE_UCP is not set, because in that case all
3984 characters > 255 are in the class, so any that were explicitly given as
3985 well can be ignored. If (when there are explicit characters > 255 that must
3986 be listed) there are no characters < 256, we can omit the bitmap in the
3987 actual compiled code. */
3988
3989 #ifdef SUPPORT_UTF8
3990 if (class_utf8 && (!should_flip_negation || (options & PCRE_UCP) != 0))
3991 {
3992 *class_utf8data++ = XCL_END; /* Marks the end of extra data */
3993 *code++ = OP_XCLASS;
3994 code += LINK_SIZE;
3995 *code = negate_class? XCL_NOT : 0;
3996
3997 /* If the map is required, move up the extra data to make room for it;
3998 otherwise just move the code pointer to the end of the extra data. */
3999
4000 if (class_charcount > 0)
4001 {
4002 *code++ |= XCL_MAP;
4003 memmove(code + 32, code, class_utf8data - code);
4004 memcpy(code, classbits, 32);
4005 code = class_utf8data + 32;
4006 }
4007 else code = class_utf8data;
4008
4009 /* Now fill in the complete length of the item */
4010
4011 PUT(previous, 1, code - previous);
4012 break; /* End of class handling */
4013 }
4014 #endif
4015
4016 /* If there are no characters > 255, or they are all to be included or
4017 excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the
4018 whole class was negated and whether there were negative specials such as \S
4019 (non-UCP) in the class. Then copy the 32-byte map into the code vector,
4020 negating it if necessary. */
4021
4022 *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;
4023 if (negate_class)
4024 {
4025 if (lengthptr == NULL) /* Save time in the pre-compile phase */
4026 for (c = 0; c < 32; c++) code[c] = ~classbits[c];
4027 }
4028 else
4029 {
4030 memcpy(code, classbits, 32);
4031 }
4032 code += 32;
4033 break;
4034
4035
4036 /* ===================================================================*/
4037 /* Various kinds of repeat; '{' is not necessarily a quantifier, but this
4038 has been tested above. */
4039
4040 case CHAR_LEFT_CURLY_BRACKET:
4041 if (!is_quantifier) goto NORMAL_CHAR;
4042 ptr = read_repeat_counts(ptr+1, &repeat_min, &repeat_max, errorcodeptr);
4043 if (*errorcodeptr != 0) goto FAILED;
4044 goto REPEAT;
4045
4046 case CHAR_ASTERISK:
4047 repeat_min = 0;
4048 repeat_max = -1;
4049 goto REPEAT;
4050
4051 case CHAR_PLUS:
4052 repeat_min = 1;
4053 repeat_max = -1;
4054 goto REPEAT;
4055
4056 case CHAR_QUESTION_MARK:
4057 repeat_min = 0;
4058 repeat_max = 1;
4059
4060 REPEAT:
4061 if (previous == NULL)
4062 {
4063 *errorcodeptr = ERR9;
4064 goto FAILED;
4065 }
4066
4067 if (repeat_min == 0)
4068 {
4069 firstbyte = zerofirstbyte; /* Adjust for zero repeat */
4070 reqbyte = zeroreqbyte; /* Ditto */
4071 }
4072
4073 /* Remember whether this is a variable length repeat */
4074
4075 reqvary = (repeat_min == repeat_max)? 0 : REQ_VARY;
4076
4077 op_type = 0; /* Default single-char op codes */
4078 possessive_quantifier = FALSE; /* Default not possessive quantifier */
4079
4080 /* Save start of previous item, in case we have to move it up to make space
4081 for an inserted OP_ONCE for the additional '+' extension. */
4082
4083 tempcode = previous;
4084
4085 /* If the next character is '+', we have a possessive quantifier. This
4086 implies greediness, whatever the setting of the PCRE_UNGREEDY option.
4087 If the next character is '?' this is a minimizing repeat, by default,
4088 but if PCRE_UNGREEDY is set, it works the other way round. We change the
4089 repeat type to the non-default. */
4090
4091 if (ptr[1] == CHAR_PLUS)
4092 {
4093 repeat_type = 0; /* Force greedy */
4094 possessive_quantifier = TRUE;
4095 ptr++;
4096 }
4097 else if (ptr[1] == CHAR_QUESTION_MARK)
4098 {
4099 repeat_type = greedy_non_default;
4100 ptr++;
4101 }
4102 else repeat_type = greedy_default;
4103
4104 /* If previous was a character match, abolish the item and generate a
4105 repeat item instead. If a char item has a minumum of more than one, ensure
4106 that it is set in reqbyte - it might not be if a sequence such as x{3} is
4107 the first thing in a branch because the x will have gone into firstbyte
4108 instead. */
4109
4110 if (*previous == OP_CHAR || *previous == OP_CHARNC)
4111 {
4112 /* Deal with UTF-8 characters that take up more than one byte. It's
4113 easier to write this out separately than try to macrify it. Use c to
4114 hold the length of the character in bytes, plus 0x80 to flag that it's a
4115 length rather than a small character. */
4116
4117 #ifdef SUPPORT_UTF8
4118 if (utf8 && (code[-1] & 0x80) != 0)
4119 {
4120 uschar *lastchar = code - 1;
4121 while((*lastchar & 0xc0) == 0x80) lastchar--;
4122 c = code - lastchar; /* Length of UTF-8 character */
4123 memcpy(utf8_char, lastchar, c); /* Save the char */
4124 c |= 0x80; /* Flag c as a length */
4125 }
4126 else
4127 #endif
4128
4129 /* Handle the case of a single byte - either with no UTF8 support, or
4130 with UTF-8 disabled, or for a UTF-8 character < 128. */
4131
4132 {
4133 c = code[-1];
4134 if (repeat_min > 1) reqbyte = c | req_caseopt | cd->req_varyopt;
4135 }
4136
4137 /* If the repetition is unlimited, it pays to see if the next thing on
4138 the line is something that cannot possibly match this character. If so,
4139 automatically possessifying this item gains some performance in the case
4140 where the match fails. */
4141
4142 if (!possessive_quantifier &&
4143 repeat_max < 0 &&
4144 check_auto_possessive(previous, utf8, ptr + 1, options, cd))
4145 {
4146 repeat_type = 0; /* Force greedy */
4147 possessive_quantifier = TRUE;
4148 }
4149
4150 goto OUTPUT_SINGLE_REPEAT; /* Code shared with single character types */
4151 }
4152
4153 /* If previous was a single negated character ([^a] or similar), we use
4154 one of the special opcodes, replacing it. The code is shared with single-
4155 character repeats by setting opt_type to add a suitable offset into
4156 repeat_type. We can also test for auto-possessification. OP_NOT is
4157 currently used only for single-byte chars. */
4158
4159 else if (*previous == OP_NOT)
4160 {
4161 op_type = OP_NOTSTAR - OP_STAR; /* Use "not" opcodes */
4162 c = previous[1];
4163 if (!possessive_quantifier &&
4164 repeat_max < 0 &&
4165 check_auto_possessive(previous, utf8, ptr + 1, options, cd))
4166 {
4167 repeat_type = 0; /* Force greedy */
4168 possessive_quantifier = TRUE;
4169 }
4170 goto OUTPUT_SINGLE_REPEAT;
4171 }
4172
4173 /* If previous was a character type match (\d or similar), abolish it and
4174 create a suitable repeat item. The code is shared with single-character
4175 repeats by setting op_type to add a suitable offset into repeat_type. Note
4176 the the Unicode property types will be present only when SUPPORT_UCP is
4177 defined, but we don't wrap the little bits of code here because it just
4178 makes it horribly messy. */
4179
4180 else if (*previous < OP_EODN)
4181 {
4182 uschar *oldcode;
4183 int prop_type, prop_value;
4184 op_type = OP_TYPESTAR - OP_STAR; /* Use type opcodes */
4185 c = *previous;
4186
4187 if (!possessive_quantifier &&
4188 repeat_max < 0 &&
4189 check_auto_possessive(previous, utf8, ptr + 1, options, cd))
4190 {
4191 repeat_type = 0; /* Force greedy */
4192 possessive_quantifier = TRUE;
4193 }
4194
4195 OUTPUT_SINGLE_REPEAT:
4196 if (*previous == OP_PROP || *previous == OP_NOTPROP)
4197 {
4198 prop_type = previous[1];
4199 prop_value = previous[2];
4200 }
4201 else prop_type = prop_value = -1;
4202
4203 oldcode = code;
4204 code = previous; /* Usually overwrite previous item */
4205
4206 /* If the maximum is zero then the minimum must also be zero; Perl allows
4207 this case, so we do too - by simply omitting the item altogether. */
4208
4209 if (repeat_max == 0) goto END_REPEAT;
4210
4211 /*--------------------------------------------------------------------*/
4212 /* This code is obsolete from release 8.00; the restriction was finally
4213 removed: */
4214
4215 /* All real repeats make it impossible to handle partial matching (maybe
4216 one day we will be able to remove this restriction). */
4217
4218 /* if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; */
4219 /*--------------------------------------------------------------------*/
4220
4221 /* Combine the op_type with the repeat_type */
4222
4223 repeat_type += op_type;
4224
4225 /* A minimum of zero is handled either as the special case * or ?, or as
4226 an UPTO, with the maximum given. */
4227
4228 if (repeat_min == 0)
4229 {
4230 if (repeat_max == -1) *code++ = OP_STAR + repeat_type;
4231 else if (repeat_max == 1) *code++ = OP_QUERY + repeat_type;
4232 else
4233 {
4234 *code++ = OP_UPTO + repeat_type;
4235 PUT2INC(code, 0, repeat_max);
4236 }
4237 }
4238
4239 /* A repeat minimum of 1 is optimized into some special cases. If the
4240 maximum is unlimited, we use OP_PLUS. Otherwise, the original item is
4241 left in place and, if the maximum is greater than 1, we use OP_UPTO with
4242 one less than the maximum. */
4243
4244 else if (repeat_min == 1)
4245 {
4246 if (repeat_max == -1)
4247 *code++ = OP_PLUS + repeat_type;
4248 else
4249 {
4250 code = oldcode; /* leave previous item in place */
4251 if (repeat_max == 1) goto END_REPEAT;
4252 *code++ = OP_UPTO + repeat_type;
4253 PUT2INC(code, 0, repeat_max - 1);
4254 }
4255 }
4256
4257 /* The case {n,n} is just an EXACT, while the general case {n,m} is
4258 handled as an EXACT followed by an UPTO. */
4259
4260 else
4261 {
4262 *code++ = OP_EXACT + op_type; /* NB EXACT doesn't have repeat_type */
4263 PUT2INC(code, 0, repeat_min);
4264
4265 /* If the maximum is unlimited, insert an OP_STAR. Before doing so,
4266 we have to insert the character for the previous code. For a repeated
4267 Unicode property match, there are two extra bytes that define the
4268 required property. In UTF-8 mode, long characters have their length in
4269 c, with the 0x80 bit as a flag. */
4270
4271 if (repeat_max < 0)
4272 {
4273 #ifdef SUPPORT_UTF8
4274 if (utf8 && c >= 128)
4275 {
4276 memcpy(code, utf8_char, c & 7);
4277 code += c & 7;
4278 }
4279 else
4280 #endif
4281 {
4282 *code++ = c;
4283 if (prop_type >= 0)
4284 {
4285 *code++ = prop_type;
4286 *code++ = prop_value;
4287 }
4288 }
4289 *code++ = OP_STAR + repeat_type;
4290 }
4291
4292 /* Else insert an UPTO if the max is greater than the min, again
4293 preceded by the character, for the previously inserted code. If the
4294 UPTO is just for 1 instance, we can use QUERY instead. */
4295
4296 else if (repeat_max != repeat_min)
4297 {
4298 #ifdef SUPPORT_UTF8
4299 if (utf8 && c >= 128)
4300 {
4301 memcpy(code, utf8_char, c & 7);
4302 code += c & 7;
4303 }
4304 else
4305 #endif
4306 *code++ = c;
4307 if (prop_type >= 0)
4308 {
4309 *code++ = prop_type;
4310 *code++ = prop_value;
4311 }
4312 repeat_max -= repeat_min;
4313
4314 if (repeat_max == 1)
4315 {
4316 *code++ = OP_QUERY + repeat_type;
4317 }
4318 else
4319 {
4320 *code++ = OP_UPTO + repeat_type;
4321 PUT2INC(code, 0, repeat_max);
4322 }
4323 }
4324 }
4325
4326 /* The character or character type itself comes last in all cases. */
4327
4328 #ifdef SUPPORT_UTF8
4329 if (utf8 && c >= 128)
4330 {
4331 memcpy(code, utf8_char, c & 7);
4332 code += c & 7;
4333 }
4334 else
4335 #endif
4336 *code++ = c;
4337
4338 /* For a repeated Unicode property match, there are two extra bytes that
4339 define the required property. */
4340
4341 #ifdef SUPPORT_UCP
4342 if (prop_type >= 0)
4343 {
4344 *code++ = prop_type;
4345 *code++ = prop_value;
4346 }
4347 #endif
4348 }
4349
4350 /* If previous was a character class or a back reference, we put the repeat
4351 stuff after it, but just skip the item if the repeat was {0,0}. */
4352
4353 else if (*previous == OP_CLASS ||
4354 *previous == OP_NCLASS ||
4355 #ifdef SUPPORT_UTF8
4356 *previous == OP_XCLASS ||
4357 #endif
4358 *previous == OP_REF)
4359 {
4360 if (repeat_max == 0)
4361 {
4362 code = previous;
4363 goto END_REPEAT;
4364 }
4365
4366 /*--------------------------------------------------------------------*/
4367 /* This code is obsolete from release 8.00; the restriction was finally
4368 removed: */
4369
4370 /* All real repeats make it impossible to handle partial matching (maybe
4371 one day we will be able to remove this restriction). */
4372
4373 /* if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; */
4374 /*--------------------------------------------------------------------*/
4375
4376 if (repeat_min == 0 && repeat_max == -1)
4377 *code++ = OP_CRSTAR + repeat_type;
4378 else if (repeat_min == 1 && repeat_max == -1)
4379 *code++ = OP_CRPLUS + repeat_type;
4380 else if (repeat_min == 0 && repeat_max == 1)
4381 *code++ = OP_CRQUERY + repeat_type;
4382 else
4383 {
4384 *code++ = OP_CRRANGE + repeat_type;
4385 PUT2INC(code, 0, repeat_min);
4386 if (repeat_max == -1) repeat_max = 0; /* 2-byte encoding for max */
4387 PUT2INC(code, 0, repeat_max);
4388 }
4389 }
4390
4391 /* If previous was a bracket group, we may have to replicate it in certain
4392 cases. */
4393
4394 else if (*previous == OP_BRA || *previous == OP_CBRA ||
4395 *previous == OP_ONCE || *previous == OP_COND)
4396 {
4397 register int i;
4398 int ketoffset = 0;
4399 int len = (int)(code - previous);
4400 uschar *bralink = NULL;
4401
4402 /* Repeating a DEFINE group is pointless */
4403
4404 if (*previous == OP_COND && previous[LINK_SIZE+1] == OP_DEF)
4405 {
4406 *errorcodeptr = ERR55;
4407 goto FAILED;
4408 }
4409
4410 /* If the maximum repeat count is unlimited, find the end of the bracket
4411 by scanning through from the start, and compute the offset back to it
4412 from the current code pointer. There may be an OP_OPT setting following
4413 the final KET, so we can't find the end just by going back from the code
4414 pointer. */
4415
4416 if (repeat_max == -1)
4417 {
4418 register uschar *ket = previous;
4419 do ket += GET(ket, 1); while (*ket != OP_KET);
4420 ketoffset = (int)(code - ket);
4421 }
4422
4423 /* The case of a zero minimum is special because of the need to stick
4424 OP_BRAZERO in front of it, and because the group appears once in the
4425 data, whereas in other cases it appears the minimum number of times. For
4426 this reason, it is simplest to treat this case separately, as otherwise
4427 the code gets far too messy. There are several special subcases when the
4428 minimum is zero. */
4429
4430 if (repeat_min == 0)
4431 {
4432 /* If the maximum is also zero, we used to just omit the group from the
4433 output altogether, like this:
4434
4435 ** if (repeat_max == 0)
4436 ** {
4437 ** code = previous;
4438 ** goto END_REPEAT;
4439 ** }
4440
4441 However, that fails when a group is referenced as a subroutine from
4442 elsewhere in the pattern, so now we stick in OP_SKIPZERO in front of it
4443 so that it is skipped on execution. As we don't have a list of which
4444 groups are referenced, we cannot do this selectively.
4445
4446 If the maximum is 1 or unlimited, we just have to stick in the BRAZERO
4447 and do no more at this point. However, we do need to adjust any
4448 OP_RECURSE calls inside the group that refer to the group itself or any
4449 internal or forward referenced group, because the offset is from the
4450 start of the whole regex. Temporarily terminate the pattern while doing
4451 this. */
4452
4453 if (repeat_max <= 1) /* Covers 0, 1, and unlimited */
4454 {
4455 *code = OP_END;
4456 adjust_recurse(previous, 1, utf8, cd, save_hwm);
4457 memmove(previous+1, previous, len);
4458 code++;
4459 if (repeat_max == 0)
4460 {
4461 *previous++ = OP_SKIPZERO;
4462 goto END_REPEAT;
4463 }
4464 *previous++ = OP_BRAZERO + repeat_type;
4465 }
4466
4467 /* If the maximum is greater than 1 and limited, we have to replicate
4468 in a nested fashion, sticking OP_BRAZERO before each set of brackets.
4469 The first one has to be handled carefully because it's the original
4470 copy, which has to be moved up. The remainder can be handled by code
4471 that is common with the non-zero minimum case below. We have to
4472 adjust the value or repeat_max, since one less copy is required. Once
4473 again, we may have to adjust any OP_RECURSE calls inside the group. */
4474
4475 else
4476 {
4477 int offset;
4478 *code = OP_END;
4479 adjust_recurse(previous, 2 + LINK_SIZE, utf8, cd, save_hwm);
4480 memmove(previous + 2 + LINK_SIZE, previous, len);
4481 code += 2 + LINK_SIZE;
4482 *previous++ = OP_BRAZERO + repeat_type;
4483 *previous++ = OP_BRA;
4484
4485 /* We chain together the bracket offset fields that have to be
4486 filled in later when the ends of the brackets are reached. */
4487
4488 offset = (bralink == NULL)? 0 : (int)(previous - bralink);
4489 bralink = previous;
4490 PUTINC(previous, 0, offset);
4491 }
4492
4493 repeat_max--;
4494 }
4495
4496 /* If the minimum is greater than zero, replicate the group as many
4497 times as necessary, and adjust the maximum to the number of subsequent
4498 copies that we need. If we set a first char from the group, and didn't
4499 set a required char, copy the latter from the former. If there are any
4500 forward reference subroutine calls in the group, there will be entries on
4501 the workspace list; replicate these with an appropriate increment. */
4502
4503 else
4504 {
4505 if (repeat_min > 1)
4506 {
4507 /* In the pre-compile phase, we don't actually do the replication. We
4508 just adjust the length as if we had. Do some paranoid checks for
4509 potential integer overflow. The INT64_OR_DOUBLE type is a 64-bit
4510 integer type when available, otherwise double. */
4511
4512 if (lengthptr != NULL)
4513 {
4514 int delta = (repeat_min - 1)*length_prevgroup;
4515 if ((INT64_OR_DOUBLE)(repeat_min - 1)*
4516 (INT64_OR_DOUBLE)length_prevgroup >
4517 (INT64_OR_DOUBLE)INT_MAX ||
4518 OFLOW_MAX - *lengthptr < delta)
4519 {
4520 *errorcodeptr = ERR20;
4521 goto FAILED;
4522 }
4523 *lengthptr += delta;
4524 }
4525
4526 /* This is compiling for real */
4527
4528 else
4529 {
4530 if (groupsetfirstbyte && reqbyte < 0) reqbyte = firstbyte;
4531 for (i = 1; i < repeat_min; i++)
4532 {
4533 uschar *hc;
4534 uschar *this_hwm = cd->hwm;
4535 memcpy(code, previous, len);
4536 for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)
4537 {
4538 PUT(cd->hwm, 0, GET(hc, 0) + len);
4539 cd->hwm += LINK_SIZE;
4540 }
4541 save_hwm = this_hwm;
4542 code += len;
4543 }
4544 }
4545 }
4546
4547 if (repeat_max > 0) repeat_max -= repeat_min;
4548 }
4549
4550 /* This code is common to both the zero and non-zero minimum cases. If
4551 the maximum is limited, it replicates the group in a nested fashion,
4552 remembering the bracket starts on a stack. In the case of a zero minimum,
4553 the first one was set up above. In all cases the repeat_max now specifies
4554 the number of additional copies needed. Again, we must remember to
4555 replicate entries on the forward reference list. */
4556
4557 if (repeat_max >= 0)
4558 {
4559 /* In the pre-compile phase, we don't actually do the replication. We
4560 just adjust the length as if we had. For each repetition we must add 1
4561 to the length for BRAZERO and for all but the last repetition we must
4562 add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some
4563 paranoid checks to avoid integer overflow. The INT64_OR_DOUBLE type is
4564 a 64-bit integer type when available, otherwise double. */
4565
4566 if (lengthptr != NULL && repeat_max > 0)
4567 {
4568 int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -
4569 2 - 2*LINK_SIZE; /* Last one doesn't nest */
4570 if ((INT64_OR_DOUBLE)repeat_max *
4571 (INT64_OR_DOUBLE)(length_prevgroup + 1 + 2 + 2*LINK_SIZE)
4572 > (INT64_OR_DOUBLE)INT_MAX ||
4573 OFLOW_MAX - *lengthptr < delta)
4574 {
4575 *errorcodeptr = ERR20;
4576 goto FAILED;
4577 }
4578 *lengthptr += delta;
4579 }
4580
4581 /* This is compiling for real */
4582
4583 else for (i = repeat_max - 1; i >= 0; i--)
4584 {
4585 uschar *hc;
4586 uschar *this_hwm = cd->hwm;
4587
4588 *code++ = OP_BRAZERO + repeat_type;
4589
4590 /* All but the final copy start a new nesting, maintaining the
4591 chain of brackets outstanding. */
4592
4593 if (i != 0)
4594 {
4595 int offset;
4596 *code++ = OP_BRA;
4597 offset = (bralink == NULL)? 0 : (int)(code - bralink);
4598 bralink = code;
4599 PUTINC(code, 0, offset);
4600 }
4601
4602 memcpy(code, previous, len);
4603 for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)
4604 {
4605 PUT(cd->hwm, 0, GET(hc, 0) + len + ((i != 0)? 2+LINK_SIZE : 1));
4606 cd->hwm += LINK_SIZE;
4607 }
4608 save_hwm = this_hwm;
4609 code += len;
4610 }
4611
4612 /* Now chain through the pending brackets, and fill in their length
4613 fields (which are holding the chain links pro tem). */
4614
4615 while (bralink != NULL)
4616 {
4617 int oldlinkoffset;
4618 int offset = (int)(code - bralink + 1);
4619 uschar *bra = code - offset;
4620 oldlinkoffset = GET(bra, 1);
4621 bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset;
4622 *code++ = OP_KET;
4623 PUTINC(code, 0, offset);
4624 PUT(bra, 1, offset);
4625 }
4626 }
4627
4628 /* If the maximum is unlimited, set a repeater in the final copy. We
4629 can't just offset backwards from the current code point, because we
4630 don't know if there's been an options resetting after the ket. The
4631 correct offset was computed above.
4632
4633 Then, when we are doing the actual compile phase, check to see whether
4634 this group is a non-atomic one that could match an empty string. If so,
4635 convert the initial operator to the S form (e.g. OP_BRA -> OP_SBRA) so
4636 that runtime checking can be done. [This check is also applied to
4637 atomic groups at runtime, but in a different way.] */
4638
4639 else
4640 {
4641 uschar *ketcode = code - ketoffset;
4642 uschar *bracode = ketcode - GET(ketcode, 1);
4643 *ketcode = OP_KETRMAX + repeat_type;
4644 if (lengthptr == NULL && *bracode != OP_ONCE)
4645 {
4646 uschar *scode = bracode;
4647 do
4648 {
4649 if (could_be_empty_branch(scode, ketcode, utf8, cd))
4650 {
4651 *bracode += OP_SBRA - OP_BRA;
4652 break;
4653 }
4654 scode += GET(scode, 1);
4655 }
4656 while (*scode == OP_ALT);
4657 }
4658 }
4659 }
4660
4661 /* If previous is OP_FAIL, it was generated by an empty class [] in
4662 JavaScript mode. The other ways in which OP_FAIL can be generated, that is
4663 by (*FAIL) or (?!) set previous to NULL, which gives a "nothing to repeat"
4664 error above. We can just ignore the repeat in JS case. */
4665
4666 else if (*previous == OP_FAIL) goto END_REPEAT;
4667
4668 /* Else there's some kind of shambles */
4669
4670 else
4671 {
4672 *errorcodeptr = ERR11;
4673 goto FAILED;
4674 }
4675
4676 /* If the character following a repeat is '+', or if certain optimization
4677 tests above succeeded, possessive_quantifier is TRUE. For some of the
4678 simpler opcodes, there is an special alternative opcode for this. For
4679 anything else, we wrap the entire repeated item inside OP_ONCE brackets.
4680 The '+' notation is just syntactic sugar, taken from Sun's Java package,
4681 but the special opcodes can optimize it a bit. The repeated item starts at
4682 tempcode, not at previous, which might be the first part of a string whose
4683 (former) last char we repeated.
4684
4685 Possessifying an 'exact' quantifier has no effect, so we can ignore it. But
4686 an 'upto' may follow. We skip over an 'exact' item, and then test the
4687 length of what remains before proceeding. */
4688
4689 if (possessive_quantifier)
4690 {
4691 int len;
4692
4693 if (*tempcode == OP_TYPEEXACT)
4694 tempcode += _pcre_OP_lengths[*tempcode] +
4695 ((tempcode[3] == OP_PROP || tempcode[3] == OP_NOTPROP)? 2 : 0);
4696
4697 else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)
4698 {
4699 tempcode += _pcre_OP_lengths[*tempcode];
4700 #ifdef SUPPORT_UTF8
4701 if (utf8 && tempcode[-1] >= 0xc0)
4702 tempcode += _pcre_utf8_table4[tempcode[-1] & 0x3f];
4703 #endif
4704 }
4705
4706 len = (int)(code - tempcode);
4707 if (len > 0) switch (*tempcode)
4708 {
4709 case OP_STAR: *tempcode = OP_POSSTAR; break;
4710 case OP_PLUS: *tempcode = OP_POSPLUS; break;
4711 case OP_QUERY: *tempcode = OP_POSQUERY; break;
4712 case OP_UPTO: *tempcode = OP_POSUPTO; break;
4713
4714 case OP_TYPESTAR: *tempcode = OP_TYPEPOSSTAR; break;
4715 case OP_TYPEPLUS: *tempcode = OP_TYPEPOSPLUS; break;
4716 case OP_TYPEQUERY: *tempcode = OP_TYPEPOSQUERY; break;
4717 case OP_TYPEUPTO: *tempcode = OP_TYPEPOSUPTO; break;
4718
4719 case OP_NOTSTAR: *tempcode = OP_NOTPOSSTAR; break;
4720 case OP_NOTPLUS: *tempcode = OP_NOTPOSPLUS; break;
4721 case OP_NOTQUERY: *tempcode = OP_NOTPOSQUERY; break;
4722 case OP_NOTUPTO: *tempcode = OP_NOTPOSUPTO; break;
4723
4724 /* Because we are moving code along, we must ensure that any
4725 pending recursive references are updated. */
4726
4727 default:
4728 *code = OP_END;
4729 adjust_recurse(tempcode, 1 + LINK_SIZE, utf8, cd, save_hwm);
4730 memmove(tempcode + 1+LINK_SIZE, tempcode, len);
4731 code += 1 + LINK_SIZE;
4732 len += 1 + LINK_SIZE;
4733 tempcode[0] = OP_ONCE;
4734 *code++ = OP_KET;
4735 PUTINC(code, 0, len);
4736 PUT(tempcode, 1, len);
4737 break;
4738 }
4739 }
4740
4741 /* In all case we no longer have a previous item. We also set the
4742 "follows varying string" flag for subsequently encountered reqbytes if
4743 it isn't already set and we have just passed a varying length item. */
4744
4745 END_REPEAT:
4746 previous = NULL;
4747 cd->req_varyopt |= reqvary;
4748 break;
4749
4750
4751 /* ===================================================================*/
4752 /* Start of nested parenthesized sub-expression, or comment or lookahead or
4753 lookbehind or option setting or condition or all the other extended
4754 parenthesis forms. */
4755
4756 case CHAR_LEFT_PARENTHESIS:
4757 newoptions = options;
4758 skipbytes = 0;
4759 bravalue = OP_CBRA;
4760 save_hwm = cd->hwm;
4761 reset_bracount = FALSE;
4762
4763 /* First deal with various "verbs" that can be introduced by '*'. */
4764
4765 if (*(++ptr) == CHAR_ASTERISK &&
4766 ((cd->ctypes[ptr[1]] & ctype_letter) != 0 || ptr[1] == ':'))
4767 {
4768 int i, namelen;
4769 int arglen = 0;
4770 const char *vn = verbnames;
4771 const uschar *name = ptr + 1;
4772 const uschar *arg = NULL;
4773 previous = NULL;
4774 while ((cd->ctypes[*++ptr] & ctype_letter) != 0) {};
4775 namelen = (int)(ptr - name);
4776
4777 if (*ptr == CHAR_COLON)
4778 {
4779 arg = ++ptr;
4780 while ((cd->ctypes[*ptr] & (ctype_letter|ctype_digit)) != 0
4781 || *ptr == '_') ptr++;
4782 arglen = (int)(ptr - arg);
4783 }
4784
4785 if (*ptr != CHAR_RIGHT_PARENTHESIS)
4786 {
4787 *errorcodeptr = ERR60;
4788 goto FAILED;
4789 }
4790
4791 /* Scan the table of verb names */
4792
4793 for (i = 0; i < verbcount; i++)
4794 {
4795 if (namelen == verbs[i].len &&
4796 strncmp((char *)name, vn, namelen) == 0)
4797 {
4798 /* Check for open captures before ACCEPT */
4799
4800 if (verbs[i].op == OP_ACCEPT)
4801 {
4802 open_capitem *oc;
4803 cd->had_accept = TRUE;
4804 for (oc = cd->open_caps; oc != NULL; oc = oc->next)
4805 {
4806 *code++ = OP_CLOSE;
4807 PUT2INC(code, 0, oc->number);
4808 }
4809 }
4810
4811 /* Handle the cases with/without an argument */
4812
4813 if (arglen == 0)
4814 {
4815 if (verbs[i].op < 0) /* Argument is mandatory */
4816 {
4817 *errorcodeptr = ERR66;
4818 goto FAILED;
4819 }
4820 *code = verbs[i].op;
4821 if (*code++ == OP_THEN)
4822 {
4823 PUT(code, 0, code - bcptr->current_branch - 1);
4824 code += LINK_SIZE;
4825 }
4826 }
4827
4828 else
4829 {
4830 if (verbs[i].op_arg < 0) /* Argument is forbidden */
4831 {
4832 *errorcodeptr = ERR59;
4833 goto FAILED;
4834 }
4835 *code = verbs[i].op_arg;
4836 if (*code++ == OP_THEN_ARG)
4837 {
4838 PUT(code, 0, code - bcptr->current_branch - 1);
4839 code += LINK_SIZE;
4840 }
4841 *code++ = arglen;
4842 memcpy(code, arg, arglen);
4843 code += arglen;
4844 *code++ = 0;
4845 }
4846
4847 break; /* Found verb, exit loop */
4848 }
4849
4850 vn += verbs[i].len + 1;
4851 }
4852
4853 if (i < verbcount) continue; /* Successfully handled a verb */
4854 *errorcodeptr = ERR60; /* Verb not recognized */
4855 goto FAILED;
4856 }
4857
4858 /* Deal with the extended parentheses; all are introduced by '?', and the
4859 appearance of any of them means that this is not a capturing group. */
4860
4861 else if (*ptr == CHAR_QUESTION_MARK)
4862 {
4863 int i, set, unset, namelen;
4864 int *optset;
4865 const uschar *name;
4866 uschar *slot;
4867
4868 switch (*(++ptr))
4869 {
4870 case CHAR_NUMBER_SIGN: /* Comment; skip to ket */
4871 ptr++;
4872 while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
4873 if (*ptr == 0)
4874 {
4875 *errorcodeptr = ERR18;
4876 goto FAILED;
4877 }
4878 continue;
4879
4880
4881 /* ------------------------------------------------------------ */
4882 case CHAR_VERTICAL_LINE: /* Reset capture count for each branch */
4883 reset_bracount = TRUE;
4884 /* Fall through */
4885
4886 /* ------------------------------------------------------------ */
4887 case CHAR_COLON: /* Non-capturing bracket */
4888 bravalue = OP_BRA;
4889 ptr++;
4890 break;
4891
4892
4893 /* ------------------------------------------------------------ */
4894 case CHAR_LEFT_PARENTHESIS:
4895 bravalue = OP_COND; /* Conditional group */
4896
4897 /* A condition can be an assertion, a number (referring to a numbered
4898 group), a name (referring to a named group), or 'R', referring to
4899 recursion. R<digits> and R&name are also permitted for recursion tests.
4900
4901 There are several syntaxes for testing a named group: (?(name)) is used
4902 by Python; Perl 5.10 onwards uses (?(<name>) or (?('name')).
4903
4904 There are two unfortunate ambiguities, caused by history. (a) 'R' can
4905 be the recursive thing or the name 'R' (and similarly for 'R' followed
4906 by digits), and (b) a number could be a name that consists of digits.
4907 In both cases, we look for a name first; if not found, we try the other
4908 cases. */
4909
4910 /* For conditions that are assertions, check the syntax, and then exit
4911 the switch. This will take control down to where bracketed groups,
4912 including assertions, are processed. */
4913
4914 if (ptr[1] == CHAR_QUESTION_MARK && (ptr[2] == CHAR_EQUALS_SIGN ||
4915 ptr[2] == CHAR_EXCLAMATION_MARK || ptr[2] == CHAR_LESS_THAN_SIGN))
4916 break;
4917
4918 /* Most other conditions use OP_CREF (a couple change to OP_RREF
4919 below), and all need to skip 3 bytes at the start of the group. */
4920
4921 code[1+LINK_SIZE] = OP_CREF;
4922 skipbytes = 3;
4923 refsign = -1;
4924
4925 /* Check for a test for recursion in a named group. */
4926
4927 if (ptr[1] == CHAR_R && ptr[2] == CHAR_AMPERSAND)
4928 {
4929 terminator = -1;
4930 ptr += 2;
4931 code[1+LINK_SIZE] = OP_RREF; /* Change the type of test */
4932 }
4933
4934 /* Check for a test for a named group's having been set, using the Perl
4935 syntax (?(<name>) or (?('name') */
4936
4937 else if (ptr[1] == CHAR_LESS_THAN_SIGN)
4938 {
4939 terminator = CHAR_GREATER_THAN_SIGN;
4940 ptr++;
4941 }
4942 else if (ptr[1] == CHAR_APOSTROPHE)
4943 {
4944 terminator = CHAR_APOSTROPHE;
4945 ptr++;
4946 }
4947 else
4948 {
4949 terminator = 0;
4950 if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);
4951 }
4952
4953 /* We now expect to read a name; any thing else is an error */
4954
4955 if ((cd->ctypes[ptr[1]] & ctype_word) == 0)
4956 {
4957 ptr += 1; /* To get the right offset */
4958 *errorcodeptr = ERR28;
4959 goto FAILED;
4960 }
4961
4962 /* Read the name, but also get it as a number if it's all digits */
4963
4964 recno = 0;
4965 name = ++ptr;
4966 while ((cd->ctypes[*ptr] & ctype_word) != 0)
4967 {
4968 if (recno >= 0)
4969 recno = ((digitab[*ptr] & ctype_digit) != 0)?
4970 recno * 10 + *ptr - CHAR_0 : -1;
4971 ptr++;
4972 }
4973 namelen = (int)(ptr - name);
4974
4975 if ((terminator > 0 && *ptr++ != terminator) ||
4976 *ptr++ != CHAR_RIGHT_PARENTHESIS)
4977 {
4978 ptr--; /* Error offset */
4979 *errorcodeptr = ERR26;
4980 goto FAILED;
4981 }
4982
4983 /* Do no further checking in the pre-compile phase. */
4984
4985 if (lengthptr != NULL) break;
4986
4987 /* In the real compile we do the work of looking for the actual
4988 reference. If the string started with "+" or "-" we require the rest to
4989 be digits, in which case recno will be set. */
4990
4991 if (refsign > 0)
4992 {
4993 if (recno <= 0)
4994 {
4995 *errorcodeptr = ERR58;
4996 goto FAILED;
4997 }
4998 recno = (refsign == CHAR_MINUS)?
4999 cd->bracount - recno + 1 : recno +cd->bracount;
5000 if (recno <= 0 || recno > cd->final_bracount)
5001 {
5002 *errorcodeptr = ERR15;
5003 goto FAILED;
5004 }
5005 PUT2(code, 2+LINK_SIZE, recno);
5006 break;
5007 }
5008
5009 /* Otherwise (did not start with "+" or "-"), start by looking for the
5010 name. If we find a name, add one to the opcode to change OP_CREF or
5011 OP_RREF into OP_NCREF or OP_NRREF. These behave exactly the same,
5012 except they record that the reference was originally to a name. The
5013 information is used to check duplicate names. */
5014
5015 slot = cd->name_table;
5016 for (i = 0; i < cd->names_found; i++)
5017 {
5018 if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break;
5019 slot += cd->name_entry_size;
5020 }
5021
5022 /* Found a previous named subpattern */
5023
5024 if (i < cd->names_found)
5025 {
5026 recno = GET2(slot, 0);
5027 PUT2(code, 2+LINK_SIZE, recno);
5028 code[1+LINK_SIZE]++;
5029 }
5030
5031 /* Search the pattern for a forward reference */
5032
5033 else if ((i = find_parens(cd, name, namelen,
5034 (options & PCRE_EXTENDED) != 0)) > 0)
5035 {
5036 PUT2(code, 2+LINK_SIZE, i);
5037 code[1+LINK_SIZE]++;
5038 }
5039
5040 /* If terminator == 0 it means that the name followed directly after
5041 the opening parenthesis [e.g. (?(abc)...] and in this case there are
5042 some further alternatives to try. For the cases where terminator != 0
5043 [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have
5044 now checked all the possibilities, so give an error. */
5045
5046 else if (terminator != 0)
5047 {
5048 *errorcodeptr = ERR15;
5049 goto FAILED;
5050 }
5051
5052 /* Check for (?(R) for recursion. Allow digits after R to specify a
5053 specific group number. */
5054
5055 else if (*name == CHAR_R)
5056 {
5057 recno = 0;
5058 for (i = 1; i < namelen; i++)
5059 {
5060 if ((digitab[name[i]] & ctype_digit) == 0)
5061 {
5062 *errorcodeptr = ERR15;
5063 goto FAILED;
5064 }
5065 recno = recno * 10 + name[i] - CHAR_0;
5066 }
5067 if (recno == 0) recno = RREF_ANY;
5068 code[1+LINK_SIZE] = OP_RREF; /* Change test type */
5069 PUT2(code, 2+LINK_SIZE, recno);
5070 }
5071
5072 /* Similarly, check for the (?(DEFINE) "condition", which is always
5073 false. */
5074
5075 else if (namelen == 6 && strncmp((char *)name, STRING_DEFINE, 6) == 0)
5076 {
5077 code[1+LINK_SIZE] = OP_DEF;
5078 skipbytes = 1;
5079 }
5080
5081 /* Check for the "name" actually being a subpattern number. We are
5082 in the second pass here, so final_bracount is set. */
5083
5084 else if (recno > 0 && recno <= cd->final_bracount)
5085 {
5086 PUT2(code, 2+LINK_SIZE, recno);
5087 }
5088
5089 /* Either an unidentified subpattern, or a reference to (?(0) */
5090
5091 else
5092 {
5093 *errorcodeptr = (recno == 0)? ERR35: ERR15;
5094 goto FAILED;
5095 }
5096 break;
5097
5098
5099 /* ------------------------------------------------------------ */
5100 case CHAR_EQUALS_SIGN: /* Positive lookahead */
5101 bravalue = OP_ASSERT;
5102 ptr++;
5103 break;
5104
5105
5106 /* ------------------------------------------------------------ */
5107 case CHAR_EXCLAMATION_MARK: /* Negative lookahead */
5108 ptr++;
5109 if (*ptr == CHAR_RIGHT_PARENTHESIS) /* Optimize (?!) */
5110 {
5111 *code++ = OP_FAIL;
5112 previous = NULL;
5113 continue;
5114 }
5115 bravalue = OP_ASSERT_NOT;
5116 break;
5117
5118
5119 /* ------------------------------------------------------------ */
5120 case CHAR_LESS_THAN_SIGN: /* Lookbehind or named define */
5121 switch (ptr[1])
5122 {
5123 case CHAR_EQUALS_SIGN: /* Positive lookbehind */
5124 bravalue = OP_ASSERTBACK;
5125 ptr += 2;
5126 break;
5127
5128 case CHAR_EXCLAMATION_MARK: /* Negative lookbehind */
5129 bravalue = OP_ASSERTBACK_NOT;
5130 ptr += 2;
5131 break;
5132
5133 default: /* Could be name define, else bad */
5134 if ((cd->ctypes[ptr[1]] & ctype_word) != 0) goto DEFINE_NAME;
5135 ptr++; /* Correct offset for error */
5136 *errorcodeptr = ERR24;
5137 goto FAILED;
5138 }
5139 break;
5140
5141
5142 /* ------------------------------------------------------------ */
5143 case CHAR_GREATER_THAN_SIGN: /* One-time brackets */
5144 bravalue = OP_ONCE;
5145 ptr++;
5146 break;
5147
5148
5149 /* ------------------------------------------------------------ */
5150 case CHAR_C: /* Callout - may be followed by digits; */
5151 previous_callout = code; /* Save for later completion */
5152 after_manual_callout = 1; /* Skip one item before completing */
5153 *code++ = OP_CALLOUT;
5154 {
5155 int n = 0;
5156 while ((digitab[*(++ptr)] & ctype_digit) != 0)
5157 n = n * 10 + *ptr - CHAR_0;
5158 if (*ptr != CHAR_RIGHT_PARENTHESIS)
5159 {
5160 *errorcodeptr = ERR39;
5161 goto FAILED;
5162 }
5163 if (n > 255)
5164 {
5165 *errorcodeptr = ERR38;
5166 goto FAILED;
5167 }
5168 *code++ = n;
5169 PUT(code, 0, (int)(ptr - cd->start_pattern + 1)); /* Pattern offset */
5170 PUT(code, LINK_SIZE, 0); /* Default length */
5171 code += 2 * LINK_SIZE;
5172 }
5173 previous = NULL;
5174 continue;
5175
5176
5177 /* ------------------------------------------------------------ */
5178 case CHAR_P: /* Python-style named subpattern handling */
5179 if (*(++ptr) == CHAR_EQUALS_SIGN ||
5180 *ptr == CHAR_GREATER_THAN_SIGN) /* Reference or recursion */
5181 {
5182 is_recurse = *ptr == CHAR_GREATER_THAN_SIGN;
5183 terminator = CHAR_RIGHT_PARENTHESIS;
5184 goto NAMED_REF_OR_RECURSE;
5185 }
5186 else if (*ptr != CHAR_LESS_THAN_SIGN) /* Test for Python-style defn */
5187 {
5188 *errorcodeptr = ERR41;
5189 goto FAILED;
5190 }
5191 /* Fall through to handle (?P< as (?< is handled */
5192
5193
5194 /* ------------------------------------------------------------ */
5195 DEFINE_NAME: /* Come here from (?< handling */
5196 case CHAR_APOSTROPHE:
5197 {
5198 terminator = (*ptr == CHAR_LESS_THAN_SIGN)?
5199 CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE;
5200 name = ++ptr;
5201
5202 while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
5203 namelen = (int)(ptr - name);
5204
5205 /* In the pre-compile phase, just do a syntax check. */
5206
5207 if (lengthptr != NULL)
5208 {
5209 if (*ptr != terminator)
5210 {
5211 *errorcodeptr = ERR42;
5212 goto FAILED;
5213 }
5214 if (cd->names_found >= MAX_NAME_COUNT)
5215 {
5216 *errorcodeptr = ERR49;
5217 goto FAILED;
5218 }
5219 if (namelen + 3 > cd->name_entry_size)
5220 {
5221 cd->name_entry_size = namelen + 3;
5222 if (namelen > MAX_NAME_SIZE)
5223 {
5224 *errorcodeptr = ERR48;
5225 goto FAILED;
5226 }
5227 }
5228 }
5229
5230 /* In the real compile, create the entry in the table, maintaining
5231 alphabetical order. Duplicate names for different numbers are
5232 permitted only if PCRE_DUPNAMES is set. Duplicate names for the same
5233 number are always OK. (An existing number can be re-used if (?|
5234 appears in the pattern.) In either event, a duplicate name results in
5235 a duplicate entry in the table, even if the number is the same. This
5236 is because the number of names, and hence the table size, is computed
5237 in the pre-compile, and it affects various numbers and pointers which
5238 would all have to be modified, and the compiled code moved down, if
5239 duplicates with the same number were omitted from the table. This
5240 doesn't seem worth the hassle. However, *different* names for the
5241 same number are not permitted. */
5242
5243 else
5244 {
5245 BOOL dupname = FALSE;
5246 slot = cd->name_table;
5247
5248 for (i = 0; i < cd->names_found; i++)
5249 {
5250 int crc = memcmp(name, slot+2, namelen);
5251 if (crc == 0)
5252 {
5253 if (slot[2+namelen] == 0)
5254 {
5255 if (GET2(slot, 0) != cd->bracount + 1 &&
5256 (options & PCRE_DUPNAMES) == 0)
5257 {
5258 *errorcodeptr = ERR43;
5259 goto FAILED;
5260 }
5261 else dupname = TRUE;
5262 }
5263 else crc = -1; /* Current name is a substring */
5264 }
5265
5266 /* Make space in the table and break the loop for an earlier
5267 name. For a duplicate or later name, carry on. We do this for
5268 duplicates so that in the simple case (when ?(| is not used) they
5269 are in order of their numbers. */
5270
5271 if (crc < 0)
5272 {
5273 memmove(slot + cd->name_entry_size, slot,
5274 (cd->names_found - i) * cd->name_entry_size);
5275 break;
5276 }
5277
5278 /* Continue the loop for a later or duplicate name */
5279
5280 slot += cd->name_entry_size;
5281 }
5282
5283 /* For non-duplicate names, check for a duplicate number before
5284 adding the new name. */
5285
5286 if (!dupname)
5287 {
5288 uschar *cslot = cd->name_table;
5289 for (i = 0; i < cd->names_found; i++)
5290 {
5291 if (cslot != slot)
5292 {
5293 if (GET2(cslot, 0) == cd->bracount + 1)
5294 {
5295 *errorcodeptr = ERR65;
5296 goto FAILED;
5297 }
5298 }
5299 else i--;
5300 cslot += cd->name_entry_size;
5301 }
5302 }
5303
5304 PUT2(slot, 0, cd->bracount + 1);
5305 memcpy(slot + 2, name, namelen);
5306 slot[2+namelen] = 0;
5307 }
5308 }
5309
5310 /* In both pre-compile and compile, count the number of names we've
5311 encountered. */
5312
5313 cd->names_found++;
5314 ptr++; /* Move past > or ' */
5315 goto NUMBERED_GROUP;
5316
5317
5318 /* ------------------------------------------------------------ */
5319 case CHAR_AMPERSAND: /* Perl recursion/subroutine syntax */
5320 terminator = CHAR_RIGHT_PARENTHESIS;
5321 is_recurse = TRUE;
5322 /* Fall through */
5323
5324 /* We come here from the Python syntax above that handles both
5325 references (?P=name) and recursion (?P>name), as well as falling
5326 through from the Perl recursion syntax (?&name). We also come here from
5327 the Perl \k<name> or \k'name' back reference syntax and the \k{name}
5328 .NET syntax, and the Oniguruma \g<...> and \g'...' subroutine syntax. */
5329
5330 NAMED_REF_OR_RECURSE:
5331 name = ++ptr;
5332 while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
5333 namelen = (int)(ptr - name);
5334
5335 /* In the pre-compile phase, do a syntax check and set a dummy
5336 reference number. */
5337
5338 if (lengthptr != NULL)
5339 {
5340 if (namelen == 0)
5341 {
5342 *errorcodeptr = ERR62;
5343 goto FAILED;
5344 }
5345 if (*ptr != terminator)
5346 {
5347 *errorcodeptr = ERR42;
5348 goto FAILED;
5349 }
5350 if (namelen > MAX_NAME_SIZE)
5351 {
5352 *errorcodeptr = ERR48;
5353 goto FAILED;
5354 }
5355 recno = 0;
5356 }
5357
5358 /* In the real compile, seek the name in the table. We check the name
5359 first, and then check that we have reached the end of the name in the
5360 table. That way, if the name that is longer than any in the table,
5361 the comparison will fail without reading beyond the table entry. */
5362
5363 else
5364 {
5365 slot = cd->name_table;
5366 for (i = 0; i < cd->names_found; i++)
5367 {
5368 if (strncmp((char *)name, (char *)slot+2, namelen) == 0 &&
5369 slot[2+namelen] == 0)
5370 break;
5371 slot += cd->name_entry_size;
5372 }
5373
5374 if (i < cd->names_found) /* Back reference */
5375 {
5376 recno = GET2(slot, 0);
5377 }
5378 else if ((recno = /* Forward back reference */
5379 find_parens(cd, name, namelen,
5380 (options & PCRE_EXTENDED) != 0)) <= 0)
5381 {
5382 *errorcodeptr = ERR15;
5383 goto FAILED;
5384 }
5385 }
5386
5387 /* In both phases, we can now go to the code than handles numerical
5388 recursion or backreferences. */
5389
5390 if (is_recurse) goto HANDLE_RECURSION;
5391 else goto HANDLE_REFERENCE;
5392
5393
5394 /* ------------------------------------------------------------ */
5395 case CHAR_R: /* Recursion */
5396 ptr++; /* Same as (?0) */
5397 /* Fall through */
5398
5399
5400 /* ------------------------------------------------------------ */
5401 case CHAR_MINUS: case CHAR_PLUS: /* Recursion or subroutine */
5402 case CHAR_0: case CHAR_1: case CHAR_2: case CHAR_3: case CHAR_4:
5403 case CHAR_5: case CHAR_6: case CHAR_7: case CHAR_8: case CHAR_9:
5404 {
5405 const uschar *called;
5406 terminator = CHAR_RIGHT_PARENTHESIS;
5407
5408 /* Come here from the \g<...> and \g'...' code (Oniguruma
5409 compatibility). However, the syntax has been checked to ensure that
5410 the ... are a (signed) number, so that neither ERR63 nor ERR29 will
5411 be called on this path, nor with the jump to OTHER_CHAR_AFTER_QUERY
5412 ever be taken. */
5413
5414 HANDLE_NUMERICAL_RECURSION:
5415
5416 if ((refsign = *ptr) == CHAR_PLUS)
5417 {
5418 ptr++;
5419 if ((digitab[*ptr] & ctype_digit) == 0)
5420 {
5421 *errorcodeptr = ERR63;
5422 goto FAILED;
5423 }
5424 }
5425 else if (refsign == CHAR_MINUS)
5426 {
5427 if ((digitab[ptr[1]] & ctype_digit) == 0)
5428 goto OTHER_CHAR_AFTER_QUERY;
5429 ptr++;
5430 }
5431
5432 recno = 0;
5433 while((digitab[*ptr] & ctype_digit) != 0)
5434 recno = recno * 10 + *ptr++ - CHAR_0;
5435
5436 if (*ptr != terminator)
5437 {
5438 *errorcodeptr = ERR29;
5439 goto FAILED;
5440 }
5441
5442 if (refsign == CHAR_MINUS)
5443 {
5444 if (recno == 0)
5445 {
5446 *errorcodeptr = ERR58;
5447 goto FAILED;
5448 }
5449 recno = cd->bracount - recno + 1;
5450 if (recno <= 0)
5451 {
5452 *errorcodeptr = ERR15;
5453 goto FAILED;
5454 }
5455 }
5456 else if (refsign == CHAR_PLUS)
5457 {
5458 if (recno == 0)
5459 {
5460 *errorcodeptr = ERR58;
5461 goto FAILED;
5462 }
5463 recno += cd->bracount;
5464 }
5465
5466 /* Come here from code above that handles a named recursion */
5467
5468 HANDLE_RECURSION:
5469
5470 previous = code;
5471 called = cd->start_code;
5472
5473 /* When we are actually compiling, find the bracket that is being
5474 referenced. Temporarily end the regex in case it doesn't exist before
5475 this point. If we end up with a forward reference, first check that
5476 the bracket does occur later so we can give the error (and position)
5477 now. Then remember this forward reference in the workspace so it can
5478 be filled in at the end. */
5479
5480 if (lengthptr == NULL)
5481 {
5482 *code = OP_END;
5483 if (recno != 0)
5484 called = _pcre_find_bracket(cd->start_code, utf8, recno);
5485
5486 /* Forward reference */
5487
5488 if (called == NULL)
5489 {
5490 if (find_parens(cd, NULL, recno,
5491 (options & PCRE_EXTENDED) != 0) < 0)
5492 {
5493 *errorcodeptr = ERR15;
5494 goto FAILED;
5495 }
5496
5497 /* Fudge the value of "called" so that when it is inserted as an
5498 offset below, what it actually inserted is the reference number
5499 of the group. */
5500
5501 called = cd->start_code + recno;
5502 PUTINC(cd->hwm, 0, (int)(code + 2 + LINK_SIZE - cd->start_code));
5503 }
5504
5505 /* If not a forward reference, and the subpattern is still open,
5506 this is a recursive call. We check to see if this is a left
5507 recursion that could loop for ever, and diagnose that case. */
5508
5509 else if (GET(called, 1) == 0 &&
5510 could_be_empty(called, code, bcptr, utf8, cd))
5511 {
5512 *errorcodeptr = ERR40;
5513 goto FAILED;
5514 }
5515 }
5516
5517 /* Insert the recursion/subroutine item, automatically wrapped inside
5518 "once" brackets. Set up a "previous group" length so that a
5519 subsequent quantifier will work. */
5520
5521 *code = OP_ONCE;
5522 PUT(code, 1, 2 + 2*LINK_SIZE);
5523 code += 1 + LINK_SIZE;
5524
5525 *code = OP_RECURSE;
5526 PUT(code, 1, (int)(called - cd->start_code));
5527 code += 1 + LINK_SIZE;
5528
5529 *code = OP_KET;
5530 PUT(code, 1, 2 + 2*LINK_SIZE);
5531 code += 1 + LINK_SIZE;
5532
5533 length_prevgroup = 3 + 3*LINK_SIZE;
5534 }
5535
5536 /* Can't determine a first byte now */
5537
5538 if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
5539 continue;
5540
5541
5542 /* ------------------------------------------------------------ */
5543 default: /* Other characters: check option setting */
5544 OTHER_CHAR_AFTER_QUERY:
5545 set = unset = 0;
5546 optset = &set;
5547
5548 while (*ptr != CHAR_RIGHT_PARENTHESIS && *ptr != CHAR_COLON)
5549 {
5550 switch (*ptr++)
5551 {
5552 case CHAR_MINUS: optset = &unset; break;
5553
5554 case CHAR_J: /* Record that it changed in the external options */
5555 *optset |= PCRE_DUPNAMES;
5556 cd->external_flags |= PCRE_JCHANGED;
5557 break;
5558
5559 case CHAR_i: *optset |= PCRE_CASELESS; break;
5560 case CHAR_m: *optset |= PCRE_MULTILINE; break;
5561 case CHAR_s: *optset |= PCRE_DOTALL; break;
5562 case CHAR_x: *optset |= PCRE_EXTENDED; break;
5563 case CHAR_U: *optset |= PCRE_UNGREEDY; break;
5564 case CHAR_X: *optset |= PCRE_EXTRA; break;
5565
5566 default: *errorcodeptr = ERR12;
5567 ptr--; /* Correct the offset */
5568 goto FAILED;
5569 }
5570 }
5571
5572 /* Set up the changed option bits, but don't change anything yet. */
5573
5574 newoptions = (options | set) & (~unset);
5575
5576 /* If the options ended with ')' this is not the start of a nested
5577 group with option changes, so the options change at this level. If this
5578 item is right at the start of the pattern, the options can be
5579 abstracted and made external in the pre-compile phase, and ignored in
5580 the compile phase. This can be helpful when matching -- for instance in
5581 caseless checking of required bytes.
5582
5583 If the code pointer is not (cd->start_code + 1 + LINK_SIZE), we are
5584 definitely *not* at the start of the pattern because something has been
5585 compiled. In the pre-compile phase, however, the code pointer can have
5586 that value after the start, because it gets reset as code is discarded
5587 during the pre-compile. However, this can happen only at top level - if
5588 we are within parentheses, the starting BRA will still be present. At
5589 any parenthesis level, the length value can be used to test if anything
5590 has been compiled at that level. Thus, a test for both these conditions
5591 is necessary to ensure we correctly detect the start of the pattern in
5592 both phases.
5593
5594 If we are not at the pattern start, compile code to change the ims
5595 options if this setting actually changes any of them, and reset the
5596 greedy defaults and the case value for firstbyte and reqbyte. */
5597
5598 if (*ptr == CHAR_RIGHT_PARENTHESIS)
5599 {
5600 if (code == cd->start_code + 1 + LINK_SIZE &&
5601 (lengthptr == NULL || *lengthptr == 2 + 2*LINK_SIZE))
5602 {
5603 cd->external_options = newoptions;
5604 }
5605 else
5606 {
5607 if ((options & PCRE_IMS) != (newoptions & PCRE_IMS))
5608 {
5609 *code++ = OP_OPT;
5610 *code++ = newoptions & PCRE_IMS;
5611 }
5612 greedy_default = ((newoptions & PCRE_UNGREEDY) != 0);
5613 greedy_non_default = greedy_default ^ 1;
5614 req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;
5615 }
5616
5617 /* Change options at this level, and pass them back for use
5618 in subsequent branches. When not at the start of the pattern, this
5619 information is also necessary so that a resetting item can be
5620 compiled at the end of a group (if we are in a group). */
5621
5622 *optionsptr = options = newoptions;
5623 previous = NULL; /* This item can't be repeated */
5624 continue; /* It is complete */
5625 }
5626
5627 /* If the options ended with ':' we are heading into a nested group
5628 with possible change of options. Such groups are non-capturing and are
5629 not assertions of any kind. All we need to do is skip over the ':';
5630 the newoptions value is handled below. */
5631
5632 bravalue = OP_BRA;
5633 ptr++;
5634 } /* End of switch for character following (? */
5635 } /* End of (? handling */
5636
5637 /* Opening parenthesis not followed by '*' or '?'. If PCRE_NO_AUTO_CAPTURE
5638 is set, all unadorned brackets become non-capturing and behave like (?:...)
5639 brackets. */
5640
5641 else if ((options & PCRE_NO_AUTO_CAPTURE) != 0)
5642 {
5643 bravalue = OP_BRA;
5644 }
5645
5646 /* Else we have a capturing group. */
5647
5648 else
5649 {
5650 NUMBERED_GROUP:
5651 cd->bracount += 1;
5652 PUT2(code, 1+LINK_SIZE, cd->bracount);
5653 skipbytes = 2;
5654 }
5655
5656 /* Process nested bracketed regex. Assertions may not be repeated, but
5657 other kinds can be. All their opcodes are >= OP_ONCE. We copy code into a
5658 non-register variable in order to be able to pass its address because some
5659 compilers complain otherwise. Pass in a new setting for the ims options if
5660 they have changed. */
5661
5662 previous = (bravalue >= OP_ONCE)? code : NULL;
5663 *code = bravalue;
5664 tempcode = code;
5665 tempreqvary = cd->req_varyopt; /* Save value before bracket */
5666 length_prevgroup = 0; /* Initialize for pre-compile phase */
5667
5668 if (!compile_regex(
5669 newoptions, /* The complete new option state */
5670 options & PCRE_IMS, /* The previous ims option state */
5671 &tempcode, /* Where to put code (updated) */
5672 &ptr, /* Input pointer (updated) */
5673 errorcodeptr, /* Where to put an error message */
5674 (bravalue == OP_ASSERTBACK ||
5675 bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
5676 reset_bracount, /* True if (?| group */
5677 skipbytes, /* Skip over bracket number */
5678 &subfirstbyte, /* For possible first char */
5679 &subreqbyte, /* For possible last char */
5680 bcptr, /* Current branch chain */
5681 cd, /* Tables block */
5682 (lengthptr == NULL)? NULL : /* Actual compile phase */
5683 &length_prevgroup /* Pre-compile phase */
5684 ))
5685 goto FAILED;
5686
5687 /* At the end of compiling, code is still pointing to the start of the
5688 group, while tempcode has been updated to point past the end of the group
5689 and any option resetting that may follow it. The pattern pointer (ptr)
5690 is on the bracket. */
5691
5692 /* If this is a conditional bracket, check that there are no more than
5693 two branches in the group, or just one if it's a DEFINE group. We do this
5694 in the real compile phase, not in the pre-pass, where the whole group may
5695 not be available. */
5696
5697 if (bravalue == OP_COND && lengthptr == NULL)
5698 {
5699 uschar *tc = code;
5700 int condcount = 0;
5701
5702 do {
5703 condcount++;
5704 tc += GET(tc,1);
5705 }
5706 while (*tc != OP_KET);
5707
5708 /* A DEFINE group is never obeyed inline (the "condition" is always
5709 false). It must have only one branch. */
5710
5711 if (code[LINK_SIZE+1] == OP_DEF)
5712 {
5713 if (condcount > 1)
5714 {
5715 *errorcodeptr = ERR54;
5716 goto FAILED;
5717 }
5718 bravalue = OP_DEF; /* Just a flag to suppress char handling below */
5719 }
5720
5721 /* A "normal" conditional group. If there is just one branch, we must not
5722 make use of its firstbyte or reqbyte, because this is equivalent to an
5723 empty second branch. */
5724
5725 else
5726 {
5727 if (condcount > 2)
5728 {
5729 *errorcodeptr = ERR27;
5730 goto FAILED;
5731 }
5732 if (condcount == 1) subfirstbyte = subreqbyte = REQ_NONE;
5733 }
5734 }
5735
5736 /* Error if hit end of pattern */
5737
5738 if (*ptr != CHAR_RIGHT_PARENTHESIS)
5739 {
5740 *errorcodeptr = ERR14;
5741 goto FAILED;
5742 }
5743
5744 /* In the pre-compile phase, update the length by the length of the group,
5745 less the brackets at either end. Then reduce the compiled code to just a
5746 set of non-capturing brackets so that it doesn't use much memory if it is
5747 duplicated by a quantifier.*/
5748
5749 if (lengthptr != NULL)
5750 {
5751 if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE)
5752 {
5753 *errorcodeptr = ERR20;
5754 goto FAILED;
5755 }
5756 *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;
5757 *code++ = OP_BRA;
5758 PUTINC(code, 0, 1 + LINK_SIZE);
5759 *code++ = OP_KET;
5760 PUTINC(code, 0, 1 + LINK_SIZE);
5761 break; /* No need to waste time with special character handling */
5762 }
5763
5764 /* Otherwise update the main code pointer to the end of the group. */
5765
5766 code = tempcode;
5767
5768 /* For a DEFINE group, required and first character settings are not
5769 relevant. */
5770
5771 if (bravalue == OP_DEF) break;
5772
5773 /* Handle updating of the required and first characters for other types of
5774 group. Update for normal brackets of all kinds, and conditions with two
5775 branches (see code above). If the bracket is followed by a quantifier with
5776 zero repeat, we have to back off. Hence the definition of zeroreqbyte and
5777 zerofirstbyte outside the main loop so that they can be accessed for the
5778 back off. */
5779
5780 zeroreqbyte = reqbyte;
5781 zerofirstbyte = firstbyte;
5782 groupsetfirstbyte = FALSE;
5783
5784 if (bravalue >= OP_ONCE)
5785 {
5786 /* If we have not yet set a firstbyte in this branch, take it from the
5787 subpattern, remembering that it was set here so that a repeat of more
5788 than one can replicate it as reqbyte if necessary. If the subpattern has
5789 no firstbyte, set "none" for the whole branch. In both cases, a zero
5790 repeat forces firstbyte to "none". */
5791
5792 if (firstbyte == REQ_UNSET)
5793 {
5794 if (subfirstbyte >= 0)
5795 {
5796 firstbyte = subfirstbyte;
5797 groupsetfirstbyte = TRUE;
5798 }
5799 else firstbyte = REQ_NONE;
5800 zerofirstbyte = REQ_NONE;
5801 }
5802
5803 /* If firstbyte was previously set, convert the subpattern's firstbyte
5804 into reqbyte if there wasn't one, using the vary flag that was in
5805 existence beforehand. */
5806
5807 else if (subfirstbyte >= 0 && subreqbyte < 0)
5808 subreqbyte = subfirstbyte | tempreqvary;
5809
5810 /* If the subpattern set a required byte (or set a first byte that isn't
5811 really the first byte - see above), set it. */
5812
5813 if (subreqbyte >= 0) reqbyte = subreqbyte;
5814 }
5815
5816 /* For a forward assertion, we take the reqbyte, if set. This can be
5817 helpful if the pattern that follows the assertion doesn't set a different
5818 char. For example, it's useful for /(?=abcde).+/. We can't set firstbyte
5819 for an assertion, however because it leads to incorrect effect for patterns
5820 such as /(?=a)a.+/ when the "real" "a" would then become a reqbyte instead
5821 of a firstbyte. This is overcome by a scan at the end if there's no
5822 firstbyte, looking for an asserted first char. */
5823
5824 else if (bravalue == OP_ASSERT && subreqbyte >= 0) reqbyte = subreqbyte;
5825 break; /* End of processing '(' */
5826
5827
5828 /* ===================================================================*/
5829 /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
5830 are arranged to be the negation of the corresponding OP_values in the
5831 default case when PCRE_UCP is not set. For the back references, the values
5832 are ESC_REF plus the reference number. Only back references and those types
5833 that consume a character may be repeated. We can test for values between
5834 ESC_b and ESC_Z for the latter; this may have to change if any new ones are
5835 ever created. */
5836
5837 case CHAR_BACKSLASH:
5838 tempptr = ptr;
5839 c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE);
5840 if (*errorcodeptr != 0) goto FAILED;
5841
5842 if (c < 0)
5843 {
5844 if (-c == ESC_Q) /* Handle start of quoted string */
5845 {
5846 if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
5847 ptr += 2; /* avoid empty string */
5848 else inescq = TRUE;
5849 continue;
5850 }
5851
5852 if (-c == ESC_E) continue; /* Perl ignores an orphan \E */
5853
5854 /* For metasequences that actually match a character, we disable the
5855 setting of a first character if it hasn't already been set. */
5856
5857 if (firstbyte == REQ_UNSET && -c > ESC_b && -c < ESC_Z)
5858 firstbyte = REQ_NONE;
5859
5860 /* Set values to reset to if this is followed by a zero repeat. */
5861
5862 zerofirstbyte = firstbyte;
5863 zeroreqbyte = reqbyte;
5864
5865 /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'
5866 is a subroutine call by number (Oniguruma syntax). In fact, the value
5867 -ESC_g is returned only for these cases. So we don't need to check for <
5868 or ' if the value is -ESC_g. For the Perl syntax \g{n} the value is
5869 -ESC_REF+n, and for the Perl syntax \g{name} the result is -ESC_k (as
5870 that is a synonym for a named back reference). */
5871
5872 if (-c == ESC_g)
5873 {
5874 const uschar *p;
5875 save_hwm = cd->hwm; /* Normally this is set when '(' is read */
5876 terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)?
5877 CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE;
5878
5879 /* These two statements stop the compiler for warning about possibly
5880 unset variables caused by the jump to HANDLE_NUMERICAL_RECURSION. In
5881 fact, because we actually check for a number below, the paths that
5882 would actually be in error are never taken. */
5883
5884 skipbytes = 0;
5885 reset_bracount = FALSE;
5886
5887 /* Test for a name */
5888
5889 if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
5890 {
5891 BOOL isnumber = TRUE;
5892 for (p = ptr + 1; *p != 0 && *p != terminator; p++)
5893 {
5894 if ((cd->ctypes[*p] & ctype_digit) == 0) isnumber = FALSE;
5895 if ((cd->ctypes[*p] & ctype_word) == 0) break;
5896 }
5897 if (*p != terminator)
5898 {
5899 *errorcodeptr = ERR57;
5900 break;
5901 }
5902 if (isnumber)
5903 {
5904 ptr++;
5905 goto HANDLE_NUMERICAL_RECURSION;
5906 }
5907 is_recurse = TRUE;
5908 goto NAMED_REF_OR_RECURSE;
5909 }
5910
5911 /* Test a signed number in angle brackets or quotes. */
5912
5913 p = ptr + 2;
5914 while ((digitab[*p] & ctype_digit) != 0) p++;
5915 if (*p != terminator)
5916 {
5917 *errorcodeptr = ERR57;
5918 break;
5919 }
5920 ptr++;
5921 goto HANDLE_NUMERICAL_RECURSION;
5922 }
5923
5924 /* \k<name> or \k'name' is a back reference by name (Perl syntax).
5925 We also support \k{name} (.NET syntax) */
5926
5927 if (-c == ESC_k && (ptr[1] == CHAR_LESS_THAN_SIGN ||
5928 ptr[1] == CHAR_APOSTROPHE || ptr[1] == CHAR_LEFT_CURLY_BRACKET))
5929 {
5930 is_recurse = FALSE;
5931 terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)?
5932 CHAR_GREATER_THAN_SIGN : (*ptr == CHAR_APOSTROPHE)?
5933 CHAR_APOSTROPHE : CHAR_RIGHT_CURLY_BRACKET;
5934 goto NAMED_REF_OR_RECURSE;
5935 }
5936
5937 /* Back references are handled specially; must disable firstbyte if
5938 not set to cope with cases like (?=(\w+))\1: which would otherwise set
5939 ':' later. */
5940
5941 if (-c >= ESC_REF)
5942 {
5943 open_capitem *oc;
5944 recno = -c - ESC_REF;
5945
5946 HANDLE_REFERENCE: /* Come here from named backref handling */
5947 if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
5948 previous = code;
5949 *code++ = OP_REF;
5950 PUT2INC(code, 0, recno);
5951 cd->backref_map |= (recno < 32)? (1 << recno) : 1;
5952 if (recno > cd->top_backref) cd->top_backref = recno;
5953
5954 /* Check to see if this back reference is recursive, that it, it
5955 is inside the group that it references. A flag is set so that the
5956 group can be made atomic. */
5957
5958 for (oc = cd->open_caps; oc != NULL; oc = oc->next)
5959 {
5960 if (oc->number == recno)
5961 {
5962 oc->flag = TRUE;
5963 break;
5964 }
5965 }
5966 }
5967
5968 /* So are Unicode property matches, if supported. */
5969
5970 #ifdef SUPPORT_UCP
5971 else if (-c == ESC_P || -c == ESC_p)
5972 {
5973 BOOL negated;
5974 int pdata;
5975 int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
5976 if (ptype < 0) goto FAILED;
5977 previous = code;
5978 *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
5979 *code++ = ptype;
5980 *code++ = pdata;
5981 }
5982 #else
5983
5984 /* If Unicode properties are not supported, \X, \P, and \p are not
5985 allowed. */
5986
5987 else if (-c == ESC_X || -c == ESC_P || -c == ESC_p)
5988 {
5989 *errorcodeptr = ERR45;
5990 goto FAILED;
5991 }
5992 #endif
5993
5994 /* For the rest (including \X when Unicode properties are supported), we
5995 can obtain the OP value by negating the escape value in the default
5996 situation when PCRE_UCP is not set. When it *is* set, we substitute
5997 Unicode property tests. */
5998
5999 else
6000 {
6001 #ifdef SUPPORT_UCP
6002 if (-c >= ESC_DU && -c <= ESC_wu)
6003 {
6004 nestptr = ptr + 1; /* Where to resume */
6005 ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */
6006 }
6007 else
6008 #endif
6009 {
6010 previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
6011 *code++ = -c;
6012 }
6013 }
6014 continue;
6015 }
6016
6017 /* We have a data character whose value is in c. In UTF-8 mode it may have
6018 a value > 127. We set its representation in the length/buffer, and then
6019 handle it as a data character. */
6020
6021 #ifdef SUPPORT_UTF8
6022 if (utf8 && c > 127)
6023 mclength = _pcre_ord2utf8(c, mcbuffer);
6024 else
6025 #endif
6026
6027 {
6028 mcbuffer[0] = c;
6029 mclength = 1;
6030 }
6031 goto ONE_CHAR;
6032
6033
6034 /* ===================================================================*/
6035 /* Handle a literal character. It is guaranteed not to be whitespace or #
6036 when the extended flag is set. If we are in UTF-8 mode, it may be a
6037 multi-byte literal character. */
6038
6039 default:
6040 NORMAL_CHAR:
6041 mclength = 1;
6042 mcbuffer[0] = c;
6043
6044 #ifdef SUPPORT_UTF8
6045 if (utf8 && c >= 0xc0)
6046 {
6047 while ((ptr[1] & 0xc0) == 0x80)
6048 mcbuffer[mclength++] = *(++ptr);
6049 }
6050 #endif
6051
6052 /* At this point we have the character's bytes in mcbuffer, and the length
6053 in mclength. When not in UTF-8 mode, the length is always 1. */
6054
6055 ONE_CHAR:
6056 previous = code;
6057 *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;
6058 for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
6059
6060 /* Remember if \r or \n were seen */
6061
6062 if (mcbuffer[0] == CHAR_CR || mcbuffer[0] == CHAR_NL)
6063 cd->external_flags |= PCRE_HASCRORLF;
6064
6065 /* Set the first and required bytes appropriately. If no previous first
6066 byte, set it from this character, but revert to none on a zero repeat.
6067 Otherwise, leave the firstbyte value alone, and don't change it on a zero
6068 repeat. */
6069
6070 if (firstbyte == REQ_UNSET)
6071 {
6072 zerofirstbyte = REQ_NONE;
6073 zeroreqbyte = reqbyte;
6074
6075 /* If the character is more than one byte long, we can set firstbyte
6076 only if it is not to be matched caselessly. */
6077
6078 if (mclength == 1 || req_caseopt == 0)
6079 {
6080 firstbyte = mcbuffer[0] | req_caseopt;
6081 if (mclength != 1) reqbyte = code[-1] | cd->req_varyopt;
6082 }
6083 else firstbyte = reqbyte = REQ_NONE;
6084 }
6085
6086 /* firstbyte was previously set; we can set reqbyte only the length is
6087 1 or the matching is caseful. */
6088
6089 else
6090 {
6091 zerofirstbyte = firstbyte;
6092 zeroreqbyte = reqbyte;
6093 if (mclength == 1 || req_caseopt == 0)
6094 reqbyte = code[-1] | req_caseopt | cd->req_varyopt;
6095 }
6096
6097 break; /* End of literal character handling */
6098 }
6099 } /* end of big loop */
6100
6101
6102 /* Control never reaches here by falling through, only by a goto for all the
6103 error states. Pass back the position in the pattern so that it can be displayed
6104 to the user for diagnosing the error. */
6105
6106 FAILED:
6107 *ptrptr = ptr;
6108 return FALSE;
6109 }
6110
6111
6112
6113
6114 /*************************************************
6115 * Compile sequence of alternatives *
6116 *************************************************/
6117
6118 /* On entry, ptr is pointing past the bracket character, but on return it
6119 points to the closing bracket, or vertical bar, or end of string. The code
6120 variable is pointing at the byte into which the BRA operator has been stored.
6121 If the ims options are changed at the start (for a (?ims: group) or during any
6122 branch, we need to insert an OP_OPT item at the start of every following branch
6123 to ensure they get set correctly at run time, and also pass the new options
6124 into every subsequent branch compile.
6125
6126 This function is used during the pre-compile phase when we are trying to find
6127 out the amount of memory needed, as well as during the real compile phase. The
6128 value of lengthptr distinguishes the two phases.
6129
6130 Arguments:
6131 options option bits, including any changes for this subpattern
6132 oldims previous settings of ims option bits
6133 codeptr -> the address of the current code pointer
6134 ptrptr -> the address of the current pattern pointer
6135 errorcodeptr -> pointer to error code variable
6136 lookbehind TRUE if this is a lookbehind assertion
6137 reset_bracount TRUE to reset the count for each branch
6138 skipbytes skip this many bytes at start (for brackets and OP_COND)
6139 firstbyteptr place to put the first required character, or a negative number
6140 reqbyteptr place to put the last required character, or a negative number
6141 bcptr pointer to the chain of currently open branches
6142 cd points to the data block with tables pointers etc.
6143 lengthptr NULL during the real compile phase
6144 points to length accumulator during pre-compile phase
6145
6146 Returns: TRUE on success
6147 */
6148
6149 static BOOL
6150 compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,
6151 int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
6152 int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,
6153 int *lengthptr)
6154 {
6155 const uschar *ptr = *ptrptr;
6156 uschar *code = *codeptr;
6157 uschar *last_branch = code;
6158 uschar *start_bracket = code;
6159 uschar *reverse_count = NULL;
6160 open_capitem capitem;
6161 int capnumber = 0;
6162 int firstbyte, reqbyte;
6163 int branchfirstbyte, branchreqbyte;
6164 int length;
6165 int orig_bracount;
6166 int max_bracount;
6167 int old_external_options = cd->external_options;
6168 branch_chain bc;
6169
6170 bc.outer = bcptr;
6171 bc.current_branch = code;
6172
6173 firstbyte = reqbyte = REQ_UNSET;
6174
6175 /* Accumulate the length for use in the pre-compile phase. Start with the
6176 length of the BRA and KET and any extra bytes that are required at the
6177 beginning. We accumulate in a local variable to save frequent testing of
6178 lenthptr for NULL. We cannot do this by looking at the value of code at the
6179 start and end of each alternative, because compiled items are discarded during
6180 the pre-compile phase so that the work space is not exceeded. */
6181
6182 length = 2 + 2*LINK_SIZE + skipbytes;
6183
6184 /* WARNING: If the above line is changed for any reason, you must also change
6185 the code that abstracts option settings at the start of the pattern and makes
6186 them global. It tests the value of length for (2 + 2*LINK_SIZE) in the
6187 pre-compile phase to find out whether anything has yet been compiled or not. */
6188
6189 /* If this is a capturing subpattern, add to the chain of open capturing items
6190 so that we can detect them if (*ACCEPT) is encountered. This is also used to
6191 detect groups that contain recursive back references to themselves. */
6192
6193 if (*code == OP_CBRA)
6194 {
6195 capnumber = GET2(code, 1 + LINK_SIZE);
6196 capitem.number = capnumber;
6197 capitem.next = cd->open_caps;
6198 capitem.flag = FALSE;
6199 cd->open_caps = &capitem;
6200 }
6201
6202 /* Offset is set zero to mark that this bracket is still open */
6203
6204 PUT(code, 1, 0);
6205 code += 1 + LINK_SIZE + skipbytes;
6206
6207 /* Loop for each alternative branch */
6208
6209 orig_bracount = max_bracount = cd->bracount;
6210 for (;;)
6211 {
6212 /* For a (?| group, reset the capturing bracket count so that each branch
6213 uses the same numbers. */
6214
6215 if (reset_bracount) cd->bracount = orig_bracount;
6216
6217 /* Handle a change of ims options at the start of the branch */
6218
6219 if ((options & PCRE_IMS) != oldims)
6220 {
6221 *code++ = OP_OPT;
6222 *code++ = options & PCRE_IMS;
6223 length += 2;
6224 }
6225
6226 /* Set up dummy OP_REVERSE if lookbehind assertion */
6227
6228 if (lookbehind)
6229 {
6230 *code++ = OP_REVERSE;
6231 reverse_count = code;
6232 PUTINC(code, 0, 0);
6233 length += 1 + LINK_SIZE;
6234 }
6235
6236 /* Now compile the branch; in the pre-compile phase its length gets added
6237 into the length. */
6238
6239 if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstbyte,
6240 &branchreqbyte, &bc, cd, (lengthptr == NULL)? NULL : &length))
6241 {
6242 *ptrptr = ptr;
6243 return FALSE;
6244 }
6245
6246 /* If the external options have changed during this branch, it means that we
6247 are at the top level, and a leading option setting has been encountered. We
6248 need to re-set the original option values to take account of this so that,
6249 during the pre-compile phase, we know to allow for a re-set at the start of
6250 subsequent branches. */
6251
6252 if (old_external_options != cd->external_options)
6253 oldims = cd->external_options & PCRE_IMS;
6254
6255 /* Keep the highest bracket count in case (?| was used and some branch
6256 has fewer than the rest. */
6257
6258 if (cd->bracount > max_bracount) max_bracount = cd->bracount;
6259
6260 /* In the real compile phase, there is some post-processing to be done. */
6261
6262 if (lengthptr == NULL)
6263 {
6264 /* If this is the first branch, the firstbyte and reqbyte values for the
6265 branch become the values for the regex. */
6266
6267 if (*last_branch != OP_ALT)
6268 {
6269 firstbyte = branchfirstbyte;
6270 reqbyte = branchreqbyte;
6271 }
6272
6273 /* If this is not the first branch, the first char and reqbyte have to
6274 match the values from all the previous branches, except that if the
6275 previous value for reqbyte didn't have REQ_VARY set, it can still match,
6276 and we set REQ_VARY for the regex. */
6277
6278 else
6279 {
6280 /* If we previously had a firstbyte, but it doesn't match the new branch,
6281 we have to abandon the firstbyte for the regex, but if there was
6282 previously no reqbyte, it takes on the value of the old firstbyte. */
6283
6284 if (firstbyte >= 0 && firstbyte != branchfirstbyte)
6285 {
6286 if (reqbyte < 0) reqbyte = firstbyte;
6287 firstbyte = REQ_NONE;
6288 }
6289
6290 /* If we (now or from before) have no firstbyte, a firstbyte from the
6291 branch becomes a reqbyte if there isn't a branch reqbyte. */
6292
6293 if (firstbyte < 0 && branchfirstbyte >= 0 && branchreqbyte < 0)
6294 branchreqbyte = branchfirstbyte;
6295
6296 /* Now ensure that the reqbytes match */
6297
6298 if ((reqbyte & ~REQ_VARY) != (branchreqbyte & ~REQ_VARY))
6299 reqbyte = REQ_NONE;
6300 else reqbyte |= branchreqbyte; /* To "or" REQ_VARY */
6301 }
6302
6303 /* If lookbehind, check that this branch matches a fixed-length string, and
6304 put the length into the OP_REVERSE item. Temporarily mark the end of the
6305 branch with OP_END. If the branch contains OP_RECURSE, the result is -3
6306 because there may be forward references that we can't check here. Set a
6307 flag to cause another lookbehind check at the end. Why not do it all at the
6308 end? Because common, erroneous checks are picked up here and the offset of
6309 the problem can be shown. */
6310
6311 if (lookbehind)
6312 {
6313 int fixed_length;
6314 *code = OP_END;
6315 fixed_length = find_fixedlength(last_branch, options, FALSE, cd);
6316 DPRINTF(("fixed length = %d\n", fixed_length));
6317 if (fixed_length == -3)
6318 {
6319 cd->check_lookbehind = TRUE;
6320 }
6321 else if (fixed_length < 0)
6322 {
6323 *errorcodeptr = (fixed_length == -2)? ERR36 : ERR25;
6324 *ptrptr = ptr;
6325 return FALSE;
6326 }
6327 else { PUT(reverse_count, 0, fixed_length); }
6328 }
6329 }
6330
6331 /* Reached end of expression, either ')' or end of pattern. In the real
6332 compile phase, go back through the alternative branches and reverse the chain
6333 of offsets, with the field in the BRA item now becoming an offset to the
6334 first alternative. If there are no alternatives, it points to the end of the
6335 group. The length in the terminating ket is always the length of the whole
6336 bracketed item. If any of the ims options were changed inside the group,
6337 compile a resetting op-code following, except at the very end of the pattern.
6338 Return leaving the pointer at the terminating char. */
6339
6340 if (*ptr != CHAR_VERTICAL_LINE)
6341 {
6342 if (lengthptr == NULL)
6343 {
6344 int branch_length = (int)(code - last_branch);
6345 do
6346 {
6347 int prev_length = GET(last_branch, 1);
6348 PUT(last_branch, 1, branch_length);
6349 branch_length = prev_length;
6350 last_branch -= branch_length;
6351 }
6352 while (branch_length > 0);
6353 }
6354
6355 /* Fill in the ket */
6356
6357 *code = OP_KET;
6358 PUT(code, 1, (int)(code - start_bracket));
6359 code += 1 + LINK_SIZE;
6360
6361 /* If it was a capturing subpattern, check to see if it contained any
6362 recursive back references. If so, we must wrap it in atomic brackets.
6363 In any event, remove the block from the chain. */
6364
6365 if (capnumber > 0)
6366 {
6367 if (cd->open_caps->flag)
6368 {
6369 memmove(start_bracket + 1 + LINK_SIZE, start_bracket,
6370 code - start_bracket);
6371 *start_bracket = OP_ONCE;
6372 code += 1 + LINK_SIZE;
6373 PUT(start_bracket, 1, (int)(code - start_bracket));
6374 *code = OP_KET;
6375 PUT(code, 1, (int)(code - start_bracket));
6376 code += 1 + LINK_SIZE;
6377 length += 2 + 2*LINK_SIZE;
6378 }
6379 cd->open_caps = cd->open_caps->next;
6380 }
6381
6382 /* Reset options if needed. */
6383
6384 if ((options & PCRE_IMS) != oldims && *ptr == CHAR_RIGHT_PARENTHESIS)
6385 {
6386 *code++ = OP_OPT;
6387 *code++ = oldims;
6388 length += 2;
6389 }
6390
6391 /* Retain the highest bracket number, in case resetting was used. */
6392
6393 cd->bracount = max_bracount;
6394
6395 /* Set values to pass back */
6396
6397 *codeptr = code;
6398 *ptrptr = ptr;
6399 *firstbyteptr = firstbyte;
6400 *reqbyteptr = reqbyte;
6401 if (lengthptr != NULL)
6402 {
6403 if (OFLOW_MAX - *lengthptr < length)
6404 {
6405 *errorcodeptr = ERR20;
6406 return FALSE;
6407 }
6408 *lengthptr += length;
6409 }
6410 return TRUE;
6411 }
6412
6413 /* Another branch follows. In the pre-compile phase, we can move the code
6414 pointer back to where it was for the start of the first branch. (That is,
6415 pretend that each branch is the only one.)
6416
6417 In the real compile phase, insert an ALT node. Its length field points back
6418 to the previous branch while the bracket remains open. At the end the chain
6419 is reversed. It's done like this so that the start of the bracket has a
6420 zero offset until it is closed, making it possible to detect recursion. */
6421
6422 if (lengthptr != NULL)
6423 {
6424 code = *codeptr + 1 + LINK_SIZE + skipbytes;
6425 length += 1 + LINK_SIZE;
6426 }
6427 else
6428 {
6429 *code = OP_ALT;
6430 PUT(code, 1, (int)(code - last_branch));
6431 bc.current_branch = last_branch = code;
6432 code += 1 + LINK_SIZE;
6433 }
6434
6435 ptr++;
6436 }
6437 /* Control never reaches here */
6438 }
6439
6440
6441
6442
6443 /*************************************************
6444 * Check for anchored expression *
6445 *************************************************/
6446
6447 /* Try to find out if this is an anchored regular expression. Consider each
6448 alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket
6449 all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then
6450 it's anchored. However, if this is a multiline pattern, then only OP_SOD
6451 counts, since OP_CIRC can match in the middle.
6452
6453 We can also consider a regex to be anchored if OP_SOM starts all its branches.
6454 This is the code for \G, which means "match at start of match position, taking
6455 into account the match offset".
6456
6457 A branch is also implicitly anchored if it starts with .* and DOTALL is set,
6458 because that will try the rest of the pattern at all possible matching points,
6459 so there is no point trying again.... er ....
6460
6461 .... except when the .* appears inside capturing parentheses, and there is a
6462 subsequent back reference to those parentheses. We haven't enough information
6463 to catch that case precisely.
6464
6465 At first, the best we could do was to detect when .* was in capturing brackets
6466 and the highest back reference was greater than or equal to that level.
6467 However, by keeping a bitmap of the first 31 back references, we can catch some
6468 of the more common cases more precisely.
6469
6470 Arguments:
6471 code points to start of expression (the bracket)
6472 options points to the options setting
6473 bracket_map a bitmap of which brackets we are inside while testing; this
6474 handles up to substring 31; after that we just have to take
6475 the less precise approach
6476 backref_map the back reference bitmap
6477
6478 Returns: TRUE or FALSE
6479 */
6480
6481 static BOOL
6482 is_anchored(register const uschar *code, int *options, unsigned int bracket_map,
6483 unsigned int backref_map)
6484 {
6485 do {
6486 const uschar *scode = first_significant_code(code + _pcre_OP_lengths[*code],
6487 options, PCRE_MULTILINE, FALSE);
6488 register int op = *scode;
6489
6490 /* Non-capturing brackets */
6491
6492 if (op == OP_BRA)
6493 {
6494 if (!is_anchored(scode, options, bracket_map, backref_map)) return FALSE;
6495 }
6496
6497 /* Capturing brackets */
6498
6499 else if (op == OP_CBRA)
6500 {
6501 int n = GET2(scode, 1+LINK_SIZE);
6502 int new_map = bracket_map | ((n < 32)? (1 << n) : 1);
6503 if (!is_anchored(scode, options, new_map, backref_map)) return FALSE;
6504 }
6505
6506 /* Other brackets */
6507
6508 else if (op == OP_ASSERT || op == OP_ONCE || op == OP_COND)
6509 {
6510 if (!is_anchored(scode, options, bracket_map, backref_map)) return FALSE;
6511 }
6512
6513 /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and
6514 it isn't in brackets that are or may be referenced. */
6515
6516 else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR ||
6517 op == OP_TYPEPOSSTAR))
6518 {
6519 if (scode[1] != OP_ALLANY || (bracket_map & backref_map) != 0)
6520 return FALSE;
6521 }
6522
6523 /* Check for explicit anchoring */
6524
6525 else if (op != OP_SOD && op != OP_SOM &&
6526 ((*options & PCRE_MULTILINE) != 0 || op != OP_CIRC))
6527 return FALSE;
6528 code += GET(code, 1);
6529 }
6530 while (*code == OP_ALT); /* Loop for each alternative */
6531 return TRUE;
6532 }
6533
6534
6535
6536 /*************************************************
6537 * Check for starting with ^ or .* *
6538 *************************************************/
6539
6540 /* This is called to find out if every branch starts with ^ or .* so that
6541 "first char" processing can be done to speed things up in multiline
6542 matching and for non-DOTALL patterns that start with .* (which must start at
6543 the beginning or after \n). As in the case of is_anchored() (see above), we
6544 have to take account of back references to capturing brackets that contain .*
6545 because in that case we can't make the assumption.
6546
6547 Arguments:
6548 code points to start of expression (the bracket)
6549 bracket_map a bitmap of which brackets we are inside while testing; this
6550 handles up to substring 31; after that we just have to take
6551 the less precise approach
6552 backref_map the back reference bitmap
6553
6554 Returns: TRUE or FALSE
6555 */
6556
6557 static BOOL
6558 is_startline(const uschar *code, unsigned int bracket_map,
6559 unsigned int backref_map)
6560 {
6561 do {
6562 const uschar *scode = first_significant_code(code + _pcre_OP_lengths[*code],
6563 NULL, 0, FALSE);
6564 register int op = *scode;
6565
6566 /* If we are at the start of a conditional assertion group, *both* the
6567 conditional assertion *and* what follows the condition must satisfy the test
6568 for start of line. Other kinds of condition fail. Note that there may be an
6569 auto-callout at the start of a condition. */
6570
6571 if (op == OP_COND)
6572 {
6573 scode += 1 + LINK_SIZE;
6574 if (*scode == OP_CALLOUT) scode += _pcre_OP_lengths[OP_CALLOUT];
6575 switch (*scode)
6576 {
6577 case OP_CREF:
6578 case OP_NCREF:
6579 case OP_RREF:
6580 case OP_NRREF:
6581 case OP_DEF:
6582 return FALSE;
6583
6584 default: /* Assertion */
6585 if (!is_startline(scode, bracket_map, backref_map)) return FALSE;
6586 do scode += GET(scode, 1); while (*scode == OP_ALT);
6587 scode += 1 + LINK_SIZE;
6588 break;
6589 }
6590 scode = first_significant_code(scode, NULL, 0, FALSE);
6591 op = *scode;
6592 }
6593
6594 /* Non-capturing brackets */
6595
6596 if (op == OP_BRA)
6597 {
6598 if (!is_startline(scode, bracket_map, backref_map)) return FALSE;
6599 }
6600
6601 /* Capturing brackets */
6602
6603 else if (op == OP_CBRA)
6604 {
6605 int n = GET2(scode, 1+LINK_SIZE);
6606 int new_map = bracket_map | ((n < 32)? (1 << n) : 1);
6607 if (!is_startline(scode, new_map, backref_map)) return FALSE;
6608 }
6609
6610 /* Other brackets */
6611
6612 else if (op == OP_ASSERT || op == OP_ONCE)
6613 {
6614 if (!is_startline(scode, bracket_map, backref_map)) return FALSE;
6615 }
6616
6617 /* .* means "start at start or after \n" if it isn't in brackets that
6618 may be referenced. */
6619
6620 else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR || op == OP_TYPEPOSSTAR)
6621 {
6622 if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE;
6623 }
6624
6625 /* Check for explicit circumflex */
6626
6627 else if (op != OP_CIRC) return FALSE;
6628
6629 /* Move on to the next alternative */
6630
6631 code += GET(code, 1);
6632 }
6633 while (*code == OP_ALT); /* Loop for each alternative */
6634 return TRUE;
6635 }
6636
6637
6638
6639 /*************************************************
6640 * Check for asserted fixed first char *
6641 *************************************************/
6642
6643 /* During compilation, the "first char" settings from forward assertions are
6644 discarded, because they can cause conflicts with actual literals that follow.
6645 However, if we end up without a first char setting for an unanchored pattern,
6646 it is worth scanning the regex to see if there is an initial asserted first
6647 char. If all branches start with the same asserted char, or with a bracket all
6648 of whose alternatives start with the same ass