/[pcre]/code/trunk/pcre_internal.h
ViewVC logotype

Contents of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 351 - (hide annotations) (download)
Fri Jul 4 18:27:16 2008 UTC (6 years, 1 month ago) by ph10
File MIME type: text/plain
File size: 47373 byte(s)
Final tidies for new Unicode property code; upgrade to Unicode 5.1.0.

1 nigel 77 /*************************************************
2     * Perl-Compatible Regular Expressions *
3     *************************************************/
4    
5    
6     /* PCRE is a library of functions to support regular expressions whose syntax
7     and semantics are as close as possible to those of the Perl 5 language.
8    
9     Written by Philip Hazel
10 ph10 305 Copyright (c) 1997-2008 University of Cambridge
11 nigel 77
12     -----------------------------------------------------------------------------
13     Redistribution and use in source and binary forms, with or without
14     modification, are permitted provided that the following conditions are met:
15    
16     * Redistributions of source code must retain the above copyright notice,
17     this list of conditions and the following disclaimer.
18    
19     * Redistributions in binary form must reproduce the above copyright
20     notice, this list of conditions and the following disclaimer in the
21     documentation and/or other materials provided with the distribution.
22    
23     * Neither the name of the University of Cambridge nor the names of its
24     contributors may be used to endorse or promote products derived from
25     this software without specific prior written permission.
26    
27     THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
28     AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29     IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30     ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
31     LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
32     CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
33     SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
34     INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
35     CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
36     ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
37     POSSIBILITY OF SUCH DAMAGE.
38     -----------------------------------------------------------------------------
39     */
40    
41     /* This header contains definitions that are shared between the different
42     modules, but which are not relevant to the exported API. This includes some
43     functions whose names all begin with "_pcre_". */
44    
45 nigel 85 #ifndef PCRE_INTERNAL_H
46     #define PCRE_INTERNAL_H
47 nigel 77
48     /* Define DEBUG to get debugging output on stdout. */
49    
50 nigel 85 #if 0
51 nigel 77 #define DEBUG
52 nigel 85 #endif
53 nigel 77
54     /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
55     inline, and there are *still* stupid compilers about that don't like indented
56     pre-processor statements, or at least there were when I first wrote this. After
57 nigel 93 all, it had only been about 10 years then...
58 nigel 77
59 nigel 93 It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
60     be absolutely sure we get our version. */
61    
62     #undef DPRINTF
63 nigel 77 #ifdef DEBUG
64     #define DPRINTF(p) printf p
65     #else
66 nigel 93 #define DPRINTF(p) /* Nothing */
67 nigel 77 #endif
68    
69    
70     /* Standard C headers plus the external interface definition. The only time
71     setjmp and stdarg are used is when NO_RECURSE is set. */
72    
73     #include <ctype.h>
74     #include <limits.h>
75     #include <setjmp.h>
76     #include <stdarg.h>
77     #include <stddef.h>
78     #include <stdio.h>
79     #include <stdlib.h>
80     #include <string.h>
81    
82 ph10 145 /* When compiling a DLL for Windows, the exported symbols have to be declared
83     using some MS magic. I found some useful information on this web page:
84     http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
85 ph10 154 information there, using __declspec(dllexport) without "extern" we have a
86 ph10 145 definition; with "extern" we have a declaration. The settings here override the
87     setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
88 ph10 154 which is all that is needed for applications (they just import the symbols). We
89     use:
90 ph10 145
91     PCRE_EXP_DECL for declarations
92     PCRE_EXP_DEFN for definitions of exported functions
93     PCRE_EXP_DATA_DEFN for definitions of exported variables
94    
95     The reason for the two DEFN macros is that in non-Windows environments, one
96     does not want to have "extern" before variable definitions because it leads to
97     compiler warnings. So we distinguish between functions and variables. In
98     Windows, the two should always be the same.
99    
100     The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
101     which is an application, but needs to import this file in order to "peek" at
102 ph10 154 internals, can #include pcre.h first to get an application's-eye view.
103 ph10 145
104     In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
105     special-purpose environments) might want to stick other stuff in front of
106     exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
107     PCRE_EXP_DATA_DEFN only if they are not already set. */
108    
109     #ifndef PCRE_EXP_DECL
110     # ifdef _WIN32
111 ph10 199 # ifndef PCRE_STATIC
112 ph10 145 # define PCRE_EXP_DECL extern __declspec(dllexport)
113     # define PCRE_EXP_DEFN __declspec(dllexport)
114     # define PCRE_EXP_DATA_DEFN __declspec(dllexport)
115     # else
116     # define PCRE_EXP_DECL extern
117     # define PCRE_EXP_DEFN
118     # define PCRE_EXP_DATA_DEFN
119     # endif
120     # else
121     # ifdef __cplusplus
122     # define PCRE_EXP_DECL extern "C"
123     # else
124     # define PCRE_EXP_DECL extern
125     # endif
126     # ifndef PCRE_EXP_DEFN
127     # define PCRE_EXP_DEFN PCRE_EXP_DECL
128     # endif
129     # ifndef PCRE_EXP_DATA_DEFN
130     # define PCRE_EXP_DATA_DEFN
131     # endif
132     # endif
133 nigel 77 #endif
134    
135     /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
136     cannot determine these outside the compilation (e.g. by running a program as
137     part of "configure") because PCRE is often cross-compiled for use on other
138     systems. Instead we make use of the maximum sizes that are available at
139     preprocessor time in standard C environments. */
140    
141     #if USHRT_MAX == 65535
142     typedef unsigned short pcre_uint16;
143 ph10 349 typedef short pcre_int16;
144 nigel 77 #elif UINT_MAX == 65535
145     typedef unsigned int pcre_uint16;
146 ph10 349 typedef int pcre_int16;
147 nigel 77 #else
148     #error Cannot determine a type for 16-bit unsigned integers
149     #endif
150    
151     #if UINT_MAX == 4294967295
152     typedef unsigned int pcre_uint32;
153 ph10 349 typedef int pcre_int32;
154 nigel 77 #elif ULONG_MAX == 4294967295
155     typedef unsigned long int pcre_uint32;
156 ph10 349 typedef long int pcre_int32;
157 nigel 77 #else
158     #error Cannot determine a type for 32-bit unsigned integers
159     #endif
160    
161     /* All character handling must be done as unsigned characters. Otherwise there
162     are problems with top-bit-set characters and functions such as isspace().
163     However, we leave the interface to the outside world as char *, because that
164     should make things easier for callers. We define a short type for unsigned char
165     to save lots of typing. I tried "uchar", but it causes problems on Digital
166     Unix, where it is defined in sys/types, so use "uschar" instead. */
167    
168     typedef unsigned char uschar;
169    
170 nigel 93 /* This is an unsigned int value that no character can ever have. UTF-8
171     characters only go up to 0x7fffffff (though Unicode doesn't go beyond
172     0x0010ffff). */
173 nigel 91
174 nigel 93 #define NOTACHAR 0xffffffff
175    
176     /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
177 ph10 149 "any" and "anycrlf" at present). The following macros are used to package up
178     testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
179     modules to indicate in which datablock the parameters exist, and what the
180     start/end of string field names are. */
181 nigel 93
182 ph10 149 #define NLTYPE_FIXED 0 /* Newline is a fixed length string */
183     #define NLTYPE_ANY 1 /* Newline is any Unicode line ending */
184     #define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */
185 nigel 93
186     /* This macro checks for a newline at the given position */
187    
188 nigel 91 #define IS_NEWLINE(p) \
189 nigel 93 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
190     ((p) < NLBLOCK->PSEND && \
191 ph10 149 _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
192     utf8)) \
193 nigel 93 : \
194     ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
195     (p)[0] == NLBLOCK->nl[0] && \
196     (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
197     ) \
198     )
199 nigel 91
200 nigel 93 /* This macro checks for a newline immediately preceding the given position */
201    
202     #define WAS_NEWLINE(p) \
203     ((NLBLOCK->nltype != NLTYPE_FIXED)? \
204     ((p) > NLBLOCK->PSSTART && \
205 ph10 149 _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
206     &(NLBLOCK->nllen), utf8)) \
207 nigel 93 : \
208     ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
209     (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
210     (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
211     ) \
212     )
213    
214 nigel 87 /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
215     with a custom type. This makes it possible, for example, to allow pcre_exec()
216     to process subject strings that are discontinuous by using a smart pointer
217     class. It must always be possible to inspect all of the subject string in
218     pcre_exec() because of the way it backtracks. Two macros are required in the
219     normal case, for sign-unspecified and unsigned char pointers. The former is
220     used for the external interface and appears in pcre.h, which is why its name
221     must begin with PCRE_. */
222    
223     #ifdef CUSTOM_SUBJECT_PTR
224     #define PCRE_SPTR CUSTOM_SUBJECT_PTR
225     #define USPTR CUSTOM_SUBJECT_PTR
226     #else
227     #define PCRE_SPTR const char *
228     #define USPTR const unsigned char *
229     #endif
230    
231 ph10 145
232    
233 nigel 85 /* Include the public PCRE header and the definitions of UCP character property
234     values. */
235 nigel 77
236 ph10 189 #include "pcre.h"
237 nigel 77 #include "ucp.h"
238    
239     /* When compiling for use with the Virtual Pascal compiler, these functions
240     need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
241     option on the command line. */
242    
243     #ifdef VPCOMPAT
244 ph10 101 #define strlen(s) _strlen(s)
245 nigel 77 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
246 ph10 101 #define memcmp(s,c,n) _memcmp(s,c,n)
247 nigel 77 #define memcpy(d,s,n) _memcpy(d,s,n)
248     #define memmove(d,s,n) _memmove(d,s,n)
249     #define memset(s,c,n) _memset(s,c,n)
250     #else /* VPCOMPAT */
251    
252     /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
253     define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
254     is set. Otherwise, include an emulating function for those systems that have
255 ph10 118 neither (there some non-Unix environments where this is the case). */
256 nigel 77
257 ph10 137 #ifndef HAVE_MEMMOVE
258 nigel 77 #undef memmove /* some systems may have a macro */
259 ph10 137 #ifdef HAVE_BCOPY
260 nigel 77 #define memmove(a, b, c) bcopy(b, a, c)
261     #else /* HAVE_BCOPY */
262 nigel 91 static void *
263 ph10 108 pcre_memmove(void *d, const void *s, size_t n)
264 nigel 77 {
265 nigel 87 size_t i;
266 ph10 108 unsigned char *dest = (unsigned char *)d;
267     const unsigned char *src = (const unsigned char *)s;
268 ph10 118 if (dest > src)
269     {
270     dest += n;
271     src += n;
272     for (i = 0; i < n; ++i) *(--dest) = *(--src);
273     return (void *)dest;
274     }
275     else
276     {
277     for (i = 0; i < n; ++i) *dest++ = *src++;
278 ph10 123 return (void *)(dest - n);
279     }
280 nigel 77 }
281     #define memmove(a, b, c) pcre_memmove(a, b, c)
282     #endif /* not HAVE_BCOPY */
283     #endif /* not HAVE_MEMMOVE */
284     #endif /* not VPCOMPAT */
285    
286    
287     /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
288     in big-endian order) by default. These are used, for example, to link from the
289     start of a subpattern to its alternatives and its end. The use of 2 bytes per
290     offset limits the size of the compiled regex to around 64K, which is big enough
291     for almost everybody. However, I received a request for an even bigger limit.
292     For this reason, and also to make the code easier to maintain, the storing and
293     loading of offsets from the byte string is now handled by the macros that are
294     defined here.
295    
296     The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
297     the config.h file, but can be overridden by using -D on the command line. This
298     is automated on Unix systems via the "configure" command. */
299    
300     #if LINK_SIZE == 2
301    
302     #define PUT(a,n,d) \
303     (a[n] = (d) >> 8), \
304     (a[(n)+1] = (d) & 255)
305    
306     #define GET(a,n) \
307     (((a)[n] << 8) | (a)[(n)+1])
308    
309     #define MAX_PATTERN_SIZE (1 << 16)
310    
311    
312     #elif LINK_SIZE == 3
313    
314     #define PUT(a,n,d) \
315     (a[n] = (d) >> 16), \
316     (a[(n)+1] = (d) >> 8), \
317     (a[(n)+2] = (d) & 255)
318    
319     #define GET(a,n) \
320     (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
321    
322     #define MAX_PATTERN_SIZE (1 << 24)
323    
324    
325     #elif LINK_SIZE == 4
326    
327     #define PUT(a,n,d) \
328     (a[n] = (d) >> 24), \
329     (a[(n)+1] = (d) >> 16), \
330     (a[(n)+2] = (d) >> 8), \
331     (a[(n)+3] = (d) & 255)
332    
333     #define GET(a,n) \
334     (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
335    
336     #define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */
337    
338    
339     #else
340     #error LINK_SIZE must be either 2, 3, or 4
341     #endif
342    
343    
344     /* Convenience macro defined in terms of the others */
345    
346     #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
347    
348    
349     /* PCRE uses some other 2-byte quantities that do not change when the size of
350     offsets changes. There are used for repeat counts and for other things such as
351     capturing parenthesis numbers in back references. */
352    
353     #define PUT2(a,n,d) \
354     a[n] = (d) >> 8; \
355     a[(n)+1] = (d) & 255
356    
357     #define GET2(a,n) \
358     (((a)[n] << 8) | (a)[(n)+1])
359    
360     #define PUT2INC(a,n,d) PUT2(a,n,d), a += 2
361    
362    
363     /* When UTF-8 encoding is being used, a character is no longer just a single
364     byte. The macros for character handling generate simple sequences when used in
365 ph10 208 byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
366     never be called in byte mode. To make sure it can never even appear when UTF-8
367 ph10 207 support is omitted, we don't even define it. */
368 nigel 77
369     #ifndef SUPPORT_UTF8
370 ph10 294 #define NEXTCHAR(p) p++;
371 nigel 77 #define GETCHAR(c, eptr) c = *eptr;
372     #define GETCHARTEST(c, eptr) c = *eptr;
373     #define GETCHARINC(c, eptr) c = *eptr++;
374     #define GETCHARINCTEST(c, eptr) c = *eptr++;
375     #define GETCHARLEN(c, eptr, len) c = *eptr;
376 ph10 207 /* #define BACKCHAR(eptr) */
377 nigel 77
378     #else /* SUPPORT_UTF8 */
379    
380 ph10 294 /* Advance a character pointer one byte in non-UTF-8 mode and by one character
381     in UTF-8 mode. */
382    
383     #define NEXTCHAR(p) \
384     p++; \
385 ph10 298 if (utf8) { while((*p & 0xc0) == 0x80) p++; }
386 ph10 294
387 nigel 77 /* Get the next UTF-8 character, not advancing the pointer. This is called when
388     we know we are in UTF-8 mode. */
389    
390     #define GETCHAR(c, eptr) \
391     c = *eptr; \
392 nigel 93 if (c >= 0xc0) \
393 nigel 77 { \
394     int gcii; \
395     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
396     int gcss = 6*gcaa; \
397     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
398     for (gcii = 1; gcii <= gcaa; gcii++) \
399     { \
400     gcss -= 6; \
401     c |= (eptr[gcii] & 0x3f) << gcss; \
402     } \
403     }
404    
405     /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
406     pointer. */
407    
408     #define GETCHARTEST(c, eptr) \
409     c = *eptr; \
410 nigel 93 if (utf8 && c >= 0xc0) \
411 nigel 77 { \
412     int gcii; \
413     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
414     int gcss = 6*gcaa; \
415     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
416     for (gcii = 1; gcii <= gcaa; gcii++) \
417     { \
418     gcss -= 6; \
419     c |= (eptr[gcii] & 0x3f) << gcss; \
420     } \
421     }
422    
423     /* Get the next UTF-8 character, advancing the pointer. This is called when we
424     know we are in UTF-8 mode. */
425    
426     #define GETCHARINC(c, eptr) \
427     c = *eptr++; \
428 nigel 93 if (c >= 0xc0) \
429 nigel 77 { \
430     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
431     int gcss = 6*gcaa; \
432     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
433     while (gcaa-- > 0) \
434     { \
435     gcss -= 6; \
436     c |= (*eptr++ & 0x3f) << gcss; \
437     } \
438     }
439    
440     /* Get the next character, testing for UTF-8 mode, and advancing the pointer */
441    
442     #define GETCHARINCTEST(c, eptr) \
443     c = *eptr++; \
444 nigel 93 if (utf8 && c >= 0xc0) \
445 nigel 77 { \
446     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
447     int gcss = 6*gcaa; \
448     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
449     while (gcaa-- > 0) \
450     { \
451     gcss -= 6; \
452     c |= (*eptr++ & 0x3f) << gcss; \
453     } \
454     }
455    
456     /* Get the next UTF-8 character, not advancing the pointer, incrementing length
457     if there are extra bytes. This is called when we know we are in UTF-8 mode. */
458    
459     #define GETCHARLEN(c, eptr, len) \
460     c = *eptr; \
461 nigel 93 if (c >= 0xc0) \
462 nigel 77 { \
463     int gcii; \
464     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
465     int gcss = 6*gcaa; \
466     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
467     for (gcii = 1; gcii <= gcaa; gcii++) \
468     { \
469     gcss -= 6; \
470     c |= (eptr[gcii] & 0x3f) << gcss; \
471     } \
472     len += gcaa; \
473     }
474    
475     /* If the pointer is not at the start of a character, move it back until
476 ph10 208 it is. This is called only in UTF-8 mode - we don't put a test within the macro
477 ph10 207 because almost all calls are already within a block of UTF-8 only code. */
478 nigel 77
479 ph10 207 #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
480 nigel 77
481     #endif
482    
483    
484     /* In case there is no definition of offsetof() provided - though any proper
485     Standard C system should have one. */
486    
487     #ifndef offsetof
488     #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
489     #endif
490    
491    
492     /* These are the public options that can change during matching. */
493    
494     #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
495    
496 ph10 231 /* Private flags containing information about the compiled regex. They used to
497     live at the top end of the options word, but that got almost full, so now they
498 ph10 230 are in a 16-bit flags word. */
499 nigel 77
500 ph10 230 #define PCRE_NOPARTIAL 0x0001 /* can't use partial with this regex */
501     #define PCRE_FIRSTSET 0x0002 /* first_byte is set */
502     #define PCRE_REQCHSET 0x0004 /* req_byte is set */
503     #define PCRE_STARTLINE 0x0008 /* start after \n for multiline */
504     #define PCRE_JCHANGED 0x0010 /* j option used in regex */
505     #define PCRE_HASCRORLF 0x0020 /* explicit \r or \n in pattern */
506 nigel 77
507     /* Options for the "extra" block produced by pcre_study(). */
508    
509     #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
510    
511     /* Masks for identifying the public options that are permitted at compile
512     time, run time, or study time, respectively. */
513    
514 ph10 149 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
515     PCRE_NEWLINE_ANYCRLF)
516 nigel 93
517 nigel 77 #define PUBLIC_OPTIONS \
518     (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
519     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
520 nigel 91 PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
521 ph10 336 PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
522     PCRE_JAVASCRIPT_COMPAT)
523 nigel 77
524     #define PUBLIC_EXEC_OPTIONS \
525     (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
526 ph10 231 PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
527 nigel 77
528     #define PUBLIC_DFA_EXEC_OPTIONS \
529     (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
530 ph10 231 PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
531     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
532 nigel 77
533     #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
534    
535     /* Magic number to provide a small check against being handed junk. Also used
536     to detect whether a pattern was compiled on a host of different endianness. */
537    
538     #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
539    
540     /* Negative values for the firstchar and reqchar variables */
541    
542     #define REQ_UNSET (-2)
543     #define REQ_NONE (-1)
544    
545     /* The maximum remaining length of subject we are prepared to search for a
546     req_byte match. */
547    
548     #define REQ_BYTE_MAX 1000
549    
550     /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a
551     variable-length repeat, or a anything other than literal characters. */
552    
553     #define REQ_CASELESS 0x0100 /* indicates caselessness */
554     #define REQ_VARY 0x0200 /* reqbyte followed non-literal item */
555    
556     /* Miscellaneous definitions */
557    
558     typedef int BOOL;
559    
560     #define FALSE 0
561     #define TRUE 1
562    
563 nigel 93 /* Escape items that are just an encoding of a particular data value. */
564 nigel 77
565     #ifndef ESC_e
566     #define ESC_e 27
567     #endif
568    
569     #ifndef ESC_f
570     #define ESC_f '\f'
571     #endif
572    
573     #ifndef ESC_n
574 nigel 93 #define ESC_n '\n'
575 nigel 77 #endif
576    
577     #ifndef ESC_r
578     #define ESC_r '\r'
579     #endif
580    
581     /* We can't officially use ESC_t because it is a POSIX reserved identifier
582     (presumably because of all the others like size_t). */
583    
584     #ifndef ESC_tee
585     #define ESC_tee '\t'
586     #endif
587    
588 nigel 87 /* Codes for different types of Unicode property */
589    
590     #define PT_ANY 0 /* Any property - matches all chars */
591     #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
592     #define PT_GC 2 /* General characteristic (e.g. L) */
593     #define PT_PC 3 /* Particular characteristic (e.g. Lu) */
594     #define PT_SC 4 /* Script (e.g. Han) */
595    
596     /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
597     contain UTF-8 characters with values greater than 255. */
598    
599     #define XCL_NOT 0x01 /* Flag: this is a negative class */
600     #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
601    
602     #define XCL_END 0 /* Marks end of individual items */
603     #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
604     #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
605     #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
606     #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
607    
608 nigel 77 /* These are escaped items that aren't just an encoding of a particular data
609     value such as \n. They must have non-zero values, as check_escape() returns
610     their negation. Also, they must appear in the same order as in the opcode
611     definitions below, up to ESC_z. There's a dummy for OP_ANY because it
612 ph10 345 corresponds to "." rather than an escape sequence, and another for OP_ALLANY
613 ph10 341 (which is used for [^] in JavaScript compatibility mode).
614 nigel 77
615 ph10 341 The final escape must be ESC_REF as subsequent values are used for
616     backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
617     greater than ESC_b and less than ESC_Z to detect the types that may be
618     repeated. These are the types that consume characters. If any new escapes are
619     put in between that don't consume a character, that code will have to change.
620     */
621    
622 ph10 172 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
623 ph10 345 ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
624     ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
625 ph10 341 ESC_REF };
626 nigel 77
627 nigel 93
628 ph10 210 /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
629 nigel 77 OP_EOD must correspond in order to the list of escapes immediately above.
630    
631 nigel 93 *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
632     that follow must also be updated to match. There is also a table called
633     "coptable" in pcre_dfa_exec.c that must be updated. */
634    
635 nigel 77 enum {
636     OP_END, /* 0 End of pattern */
637    
638     /* Values corresponding to backslashed metacharacters */
639    
640     OP_SOD, /* 1 Start of data: \A */
641     OP_SOM, /* 2 Start of match (subject + offset): \G */
642 ph10 172 OP_SET_SOM, /* 3 Set start of match (\K) */
643 ph10 168 OP_NOT_WORD_BOUNDARY, /* 4 \B */
644     OP_WORD_BOUNDARY, /* 5 \b */
645     OP_NOT_DIGIT, /* 6 \D */
646     OP_DIGIT, /* 7 \d */
647     OP_NOT_WHITESPACE, /* 8 \S */
648     OP_WHITESPACE, /* 9 \s */
649     OP_NOT_WORDCHAR, /* 10 \W */
650     OP_WORDCHAR, /* 11 \w */
651 ph10 341 OP_ANY, /* 12 Match any character (subject to DOTALL) */
652     OP_ALLANY, /* 13 Match any character (not subject to DOTALL) */
653     OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
654     OP_NOTPROP, /* 15 \P (not Unicode property) */
655     OP_PROP, /* 16 \p (Unicode property) */
656     OP_ANYNL, /* 17 \R (any newline sequence) */
657     OP_NOT_HSPACE, /* 18 \H (not horizontal whitespace) */
658     OP_HSPACE, /* 19 \h (horizontal whitespace) */
659     OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */
660     OP_VSPACE, /* 21 \v (vertical whitespace) */
661     OP_EXTUNI, /* 22 \X (extended Unicode sequence */
662     OP_EODN, /* 23 End of data or \n at end of data: \Z. */
663     OP_EOD, /* 24 End of data: \z */
664 nigel 77
665 ph10 341 OP_OPT, /* 25 Set runtime options */
666     OP_CIRC, /* 26 Start of line - varies with multiline switch */
667     OP_DOLL, /* 27 End of line - varies with multiline switch */
668     OP_CHAR, /* 28 Match one character, casefully */
669     OP_CHARNC, /* 29 Match one character, caselessly */
670     OP_NOT, /* 30 Match one character, not the following one */
671 nigel 77
672 ph10 341 OP_STAR, /* 31 The maximizing and minimizing versions of */
673     OP_MINSTAR, /* 32 these six opcodes must come in pairs, with */
674     OP_PLUS, /* 33 the minimizing one second. */
675     OP_MINPLUS, /* 34 This first set applies to single characters.*/
676     OP_QUERY, /* 35 */
677     OP_MINQUERY, /* 36 */
678 nigel 77
679 ph10 341 OP_UPTO, /* 37 From 0 to n matches */
680     OP_MINUPTO, /* 38 */
681     OP_EXACT, /* 39 Exactly n matches */
682 nigel 77
683 ph10 341 OP_POSSTAR, /* 40 Possessified star */
684     OP_POSPLUS, /* 41 Possessified plus */
685     OP_POSQUERY, /* 42 Posesssified query */
686     OP_POSUPTO, /* 43 Possessified upto */
687 nigel 77
688 ph10 341 OP_NOTSTAR, /* 44 The maximizing and minimizing versions of */
689     OP_NOTMINSTAR, /* 45 these six opcodes must come in pairs, with */
690     OP_NOTPLUS, /* 46 the minimizing one second. They must be in */
691     OP_NOTMINPLUS, /* 47 exactly the same order as those above. */
692     OP_NOTQUERY, /* 48 This set applies to "not" single characters. */
693     OP_NOTMINQUERY, /* 49 */
694 nigel 77
695 ph10 341 OP_NOTUPTO, /* 50 From 0 to n matches */
696     OP_NOTMINUPTO, /* 51 */
697     OP_NOTEXACT, /* 52 Exactly n matches */
698 nigel 93
699 ph10 341 OP_NOTPOSSTAR, /* 53 Possessified versions */
700     OP_NOTPOSPLUS, /* 54 */
701     OP_NOTPOSQUERY, /* 55 */
702     OP_NOTPOSUPTO, /* 56 */
703 nigel 93
704 ph10 341 OP_TYPESTAR, /* 57 The maximizing and minimizing versions of */
705     OP_TYPEMINSTAR, /* 58 these six opcodes must come in pairs, with */
706     OP_TYPEPLUS, /* 59 the minimizing one second. These codes must */
707     OP_TYPEMINPLUS, /* 60 be in exactly the same order as those above. */
708     OP_TYPEQUERY, /* 61 This set applies to character types such as \d */
709     OP_TYPEMINQUERY, /* 62 */
710 nigel 93
711 ph10 341 OP_TYPEUPTO, /* 63 From 0 to n matches */
712     OP_TYPEMINUPTO, /* 64 */
713     OP_TYPEEXACT, /* 65 Exactly n matches */
714 nigel 93
715 ph10 341 OP_TYPEPOSSTAR, /* 66 Possessified versions */
716     OP_TYPEPOSPLUS, /* 67 */
717     OP_TYPEPOSQUERY, /* 68 */
718     OP_TYPEPOSUPTO, /* 69 */
719 nigel 93
720 ph10 341 OP_CRSTAR, /* 70 The maximizing and minimizing versions of */
721     OP_CRMINSTAR, /* 71 all these opcodes must come in pairs, with */
722     OP_CRPLUS, /* 72 the minimizing one second. These codes must */
723     OP_CRMINPLUS, /* 73 be in exactly the same order as those above. */
724     OP_CRQUERY, /* 74 These are for character classes and back refs */
725     OP_CRMINQUERY, /* 75 */
726     OP_CRRANGE, /* 76 These are different to the three sets above. */
727     OP_CRMINRANGE, /* 77 */
728 nigel 93
729 ph10 341 OP_CLASS, /* 78 Match a character class, chars < 256 only */
730     OP_NCLASS, /* 79 Same, but the bitmap was created from a negative
731 nigel 77 class - the difference is relevant only when a UTF-8
732     character > 255 is encountered. */
733    
734 ph10 341 OP_XCLASS, /* 80 Extended class for handling UTF-8 chars within the
735 nigel 77 class. This does both positive and negative. */
736    
737 ph10 341 OP_REF, /* 81 Match a back reference */
738     OP_RECURSE, /* 82 Match a numbered subpattern (possibly recursive) */
739     OP_CALLOUT, /* 83 Call out to external function if provided */
740 nigel 77
741 ph10 341 OP_ALT, /* 84 Start of alternation */
742     OP_KET, /* 85 End of group that doesn't have an unbounded repeat */
743     OP_KETRMAX, /* 86 These two must remain together and in this */
744     OP_KETRMIN, /* 87 order. They are for groups the repeat for ever. */
745 nigel 77
746 nigel 93 /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
747 nigel 77
748 ph10 341 OP_ASSERT, /* 88 Positive lookahead */
749     OP_ASSERT_NOT, /* 89 Negative lookahead */
750     OP_ASSERTBACK, /* 90 Positive lookbehind */
751     OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
752     OP_REVERSE, /* 92 Move pointer back - used in lookbehind assertions */
753 nigel 77
754 nigel 93 /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
755     as there's a test for >= ONCE for a subpattern that isn't an assertion. */
756 nigel 77
757 ph10 341 OP_ONCE, /* 93 Atomic group */
758     OP_BRA, /* 94 Start of non-capturing bracket */
759     OP_CBRA, /* 95 Start of capturing bracket */
760     OP_COND, /* 96 Conditional group */
761 nigel 77
762 nigel 93 /* These three must follow the previous three, in the same order. There's a
763     check for >= SBRA to distinguish the two sets. */
764 nigel 77
765 ph10 341 OP_SBRA, /* 97 Start of non-capturing bracket, check empty */
766     OP_SCBRA, /* 98 Start of capturing bracket, check empty */
767     OP_SCOND, /* 99 Conditional group, check empty */
768 nigel 77
769 ph10 341 OP_CREF, /* 100 Used to hold a capture number as condition */
770     OP_RREF, /* 101 Used to hold a recursion number as condition */
771     OP_DEF, /* 102 The DEFINE condition */
772 nigel 93
773 ph10 341 OP_BRAZERO, /* 103 These two must remain together and in this */
774     OP_BRAMINZERO, /* 104 order. */
775 ph10 210
776     /* These are backtracking control verbs */
777    
778 ph10 341 OP_PRUNE, /* 105 */
779     OP_SKIP, /* 106 */
780     OP_THEN, /* 107 */
781     OP_COMMIT, /* 108 */
782 ph10 210
783     /* These are forced failure and success verbs */
784    
785 ph10 341 OP_FAIL, /* 109 */
786     OP_ACCEPT, /* 110 */
787 ph10 345
788 ph10 335 /* This is used to skip a subpattern with a {0} quantifier */
789 ph10 345
790 ph10 341 OP_SKIPZERO /* 111 */
791 nigel 77 };
792    
793    
794     /* This macro defines textual names for all the opcodes. These are used only
795     for debugging. The macro is referenced only in pcre_printint.c. */
796    
797     #define OP_NAME_LIST \
798 ph10 168 "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \
799 ph10 341 "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte", \
800 ph10 178 "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \
801     "extuni", "\\Z", "\\z", \
802 nigel 77 "Opt", "^", "$", "char", "charnc", "not", \
803     "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
804 nigel 93 "*+","++", "?+", "{", \
805 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
806 nigel 93 "*+","++", "?+", "{", \
807 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
808 nigel 93 "*+","++", "?+", "{", \
809 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", \
810     "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \
811     "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \
812 nigel 93 "AssertB", "AssertB not", "Reverse", \
813 ph10 192 "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond", \
814 ph10 210 "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero", \
815 ph10 335 "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT", \
816 ph10 345 "Skip zero"
817 nigel 77
818    
819     /* This macro defines the length of fixed length operations in the compiled
820     regex. The lengths are used when searching for specific things, and also in the
821     debugging printing of a compiled regex. We use a macro so that it can be
822     defined close to the definitions of the opcodes themselves.
823    
824     As things have been extended, some of these are no longer fixed lenths, but are
825     minima instead. For example, the length of a single-character repeat may vary
826     in UTF-8 mode. The code that uses this table must know about such things. */
827    
828     #define OP_LENGTHS \
829     1, /* End */ \
830 ph10 168 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \
831     1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \
832 ph10 341 1, 1, 1, /* Any, AllAny, Anybyte */ \
833 ph10 178 3, 3, 1, /* NOTPROP, PROP, EXTUNI */ \
834     1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \
835 nigel 77 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \
836     2, /* Char - the minimum length */ \
837     2, /* Charnc - the minimum length */ \
838     2, /* not */ \
839     /* Positive single-char repeats ** These are */ \
840     2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
841     4, 4, 4, /* upto, minupto, exact ** UTF-8 mode */ \
842 nigel 93 2, 2, 2, 4, /* *+, ++, ?+, upto+ */ \
843 nigel 77 /* Negative single-char repeats - only for chars < 256 */ \
844     2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
845     4, 4, 4, /* NOT upto, minupto, exact */ \
846 nigel 93 2, 2, 2, 4, /* Possessive *, +, ?, upto */ \
847 nigel 77 /* Positive type repeats */ \
848     2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
849     4, 4, 4, /* Type upto, minupto, exact */ \
850 nigel 93 2, 2, 2, 4, /* Possessive *+, ++, ?+, upto+ */ \
851 nigel 77 /* Character class & ref repeats */ \
852     1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
853     5, 5, /* CRRANGE, CRMINRANGE */ \
854     33, /* CLASS */ \
855     33, /* NCLASS */ \
856     0, /* XCLASS - variable length */ \
857     3, /* REF */ \
858     1+LINK_SIZE, /* RECURSE */ \
859     2+2*LINK_SIZE, /* CALLOUT */ \
860     1+LINK_SIZE, /* Alt */ \
861     1+LINK_SIZE, /* Ket */ \
862     1+LINK_SIZE, /* KetRmax */ \
863     1+LINK_SIZE, /* KetRmin */ \
864     1+LINK_SIZE, /* Assert */ \
865     1+LINK_SIZE, /* Assert not */ \
866     1+LINK_SIZE, /* Assert behind */ \
867     1+LINK_SIZE, /* Assert behind not */ \
868     1+LINK_SIZE, /* Reverse */ \
869 nigel 93 1+LINK_SIZE, /* ONCE */ \
870     1+LINK_SIZE, /* BRA */ \
871     3+LINK_SIZE, /* CBRA */ \
872 nigel 77 1+LINK_SIZE, /* COND */ \
873 nigel 93 1+LINK_SIZE, /* SBRA */ \
874     3+LINK_SIZE, /* SCBRA */ \
875     1+LINK_SIZE, /* SCOND */ \
876 nigel 77 3, /* CREF */ \
877 nigel 93 3, /* RREF */ \
878     1, /* DEF */ \
879 nigel 77 1, 1, /* BRAZERO, BRAMINZERO */ \
880 ph10 210 1, 1, 1, 1, /* PRUNE, SKIP, THEN, COMMIT, */ \
881 ph10 335 1, 1, 1 /* FAIL, ACCEPT, SKIPZERO */
882 nigel 77
883    
884 nigel 93 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
885 nigel 77
886 nigel 93 #define RREF_ANY 0xffff
887 nigel 77
888     /* Error code numbers. They are given names so that they can more easily be
889     tracked. */
890    
891     enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
892     ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
893     ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
894     ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
895 nigel 91 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
896 ph10 210 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
897 ph10 336 ERR60, ERR61, ERR62, ERR63, ERR64 };
898 nigel 77
899     /* The real format of the start of the pcre block; the index of names and the
900     code vector run on as long as necessary after the end. We store an explicit
901     offset to the name table so that if a regex is compiled on one host, saved, and
902     then run on another where the size of pointers is different, all might still
903     be well. For the case of compiled-on-4 and run-on-8, we include an extra
904     pointer that is always NULL. For future-proofing, a few dummy fields were
905     originally included - even though you can never get this planning right - but
906     there is only one left now.
907    
908     NOTE NOTE NOTE:
909     Because people can now save and re-use compiled patterns, any additions to this
910     structure should be made at the end, and something earlier (e.g. a new
911     flag in the options or one of the dummy fields) should indicate that the new
912     fields are present. Currently PCRE always sets the dummy fields to zero.
913     NOTE NOTE NOTE:
914     */
915    
916     typedef struct real_pcre {
917     pcre_uint32 magic_number;
918     pcre_uint32 size; /* Total that was malloced */
919 ph10 230 pcre_uint32 options; /* Public options */
920     pcre_uint16 flags; /* Private flags */
921     pcre_uint16 dummy1; /* For future use */
922 nigel 77 pcre_uint16 top_bracket;
923     pcre_uint16 top_backref;
924     pcre_uint16 first_byte;
925     pcre_uint16 req_byte;
926     pcre_uint16 name_table_offset; /* Offset to name table that follows */
927     pcre_uint16 name_entry_size; /* Size of any name items */
928     pcre_uint16 name_count; /* Number of name items */
929     pcre_uint16 ref_count; /* Reference count */
930    
931     const unsigned char *tables; /* Pointer to tables or NULL for std */
932     const unsigned char *nullpad; /* NULL padding */
933     } real_pcre;
934    
935     /* The format of the block used to store data from pcre_study(). The same
936     remark (see NOTE above) about extending this structure applies. */
937    
938     typedef struct pcre_study_data {
939     pcre_uint32 size; /* Total that was malloced */
940     pcre_uint32 options;
941     uschar start_bits[32];
942     } pcre_study_data;
943    
944     /* Structure for passing "static" information around between the functions
945     doing the compiling, so that they are thread-safe. */
946    
947     typedef struct compile_data {
948     const uschar *lcc; /* Points to lower casing table */
949     const uschar *fcc; /* Points to case-flipping table */
950     const uschar *cbits; /* Points to character type table */
951     const uschar *ctypes; /* Points to table of type maps */
952 nigel 93 const uschar *start_workspace;/* The start of working space */
953 nigel 77 const uschar *start_code; /* The start of the compiled code */
954     const uschar *start_pattern; /* The start of the pattern */
955 nigel 93 const uschar *end_pattern; /* The end of the pattern */
956     uschar *hwm; /* High watermark of workspace */
957 nigel 77 uschar *name_table; /* The name/number table */
958     int names_found; /* Number of entries so far */
959     int name_entry_size; /* Size of each entry */
960 ph10 270 int bracount; /* Count of capturing parens as we compile */
961 ph10 286 int final_bracount; /* Saved value after first pass */
962 nigel 77 int top_backref; /* Maximum back reference */
963     unsigned int backref_map; /* Bitmap of low back refs */
964 nigel 93 int external_options; /* External (initial) options */
965 ph10 231 int external_flags; /* External flag bits to be set */
966 nigel 77 int req_varyopt; /* "After variable item" flag for reqbyte */
967 ph10 211 BOOL had_accept; /* (*ACCEPT) encountered */
968 nigel 93 int nltype; /* Newline type */
969     int nllen; /* Newline string length */
970     uschar nl[4]; /* Newline string when fixed length */
971 nigel 77 } compile_data;
972    
973     /* Structure for maintaining a chain of pointers to the currently incomplete
974     branches, for testing for left recursion. */
975    
976     typedef struct branch_chain {
977     struct branch_chain *outer;
978     uschar *current;
979     } branch_chain;
980    
981     /* Structure for items in a linked list that represents an explicit recursive
982     call within the pattern. */
983    
984     typedef struct recursion_info {
985     struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
986     int group_num; /* Number of group that was called */
987     const uschar *after_call; /* "Return value": points after the call in the expr */
988 ph10 168 USPTR save_start; /* Old value of mstart */
989 nigel 77 int *offset_save; /* Pointer to start of saved offsets */
990     int saved_max; /* Number of saved offsets */
991     } recursion_info;
992    
993 nigel 93 /* Structure for building a chain of data for holding the values of the subject
994     pointer at the start of each subpattern, so as to detect when an empty string
995     has been matched by a subpattern - to break infinite loops. */
996    
997     typedef struct eptrblock {
998     struct eptrblock *epb_prev;
999     USPTR epb_saved_eptr;
1000     } eptrblock;
1001    
1002    
1003 nigel 77 /* Structure for passing "static" information around between the functions
1004     doing traditional NFA matching, so that they are thread-safe. */
1005    
1006     typedef struct match_data {
1007 nigel 87 unsigned long int match_call_count; /* As it says */
1008     unsigned long int match_limit; /* As it says */
1009     unsigned long int match_limit_recursion; /* As it says */
1010 nigel 77 int *offset_vector; /* Offset vector */
1011     int offset_end; /* One past the end */
1012     int offset_max; /* The maximum usable for return data */
1013 nigel 93 int nltype; /* Newline type */
1014     int nllen; /* Newline string length */
1015     uschar nl[4]; /* Newline string when fixed */
1016 nigel 77 const uschar *lcc; /* Points to lower casing table */
1017     const uschar *ctypes; /* Points to table of type maps */
1018     BOOL offset_overflow; /* Set if too many extractions */
1019     BOOL notbol; /* NOTBOL flag */
1020     BOOL noteol; /* NOTEOL flag */
1021     BOOL utf8; /* UTF8 flag */
1022 ph10 345 BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */
1023 nigel 77 BOOL endonly; /* Dollar not before final \n */
1024     BOOL notempty; /* Empty string match not wanted */
1025     BOOL partial; /* PARTIAL flag */
1026     BOOL hitend; /* Hit the end of the subject at some point */
1027 ph10 231 BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */
1028 nigel 77 const uschar *start_code; /* For use when recursing */
1029 nigel 87 USPTR start_subject; /* Start of the subject string */
1030     USPTR end_subject; /* End of the subject string */
1031 ph10 168 USPTR start_match_ptr; /* Start of matched string */
1032 nigel 87 USPTR end_match_ptr; /* Subject position at end match */
1033 nigel 77 int end_offset_top; /* Highwater mark at end of match */
1034     int capture_last; /* Most recent capture number */
1035     int start_offset; /* The start offset value */
1036 nigel 93 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
1037     int eptrn; /* Next free eptrblock */
1038 nigel 77 recursion_info *recursive; /* Linked list of recursion data */
1039     void *callout_data; /* To pass back to callouts */
1040     } match_data;
1041    
1042     /* A similar structure is used for the same purpose by the DFA matching
1043     functions. */
1044    
1045     typedef struct dfa_match_data {
1046     const uschar *start_code; /* Start of the compiled pattern */
1047     const uschar *start_subject; /* Start of the subject string */
1048     const uschar *end_subject; /* End of subject string */
1049     const uschar *tables; /* Character tables */
1050     int moptions; /* Match options */
1051     int poptions; /* Pattern options */
1052 nigel 93 int nltype; /* Newline type */
1053     int nllen; /* Newline string length */
1054     uschar nl[4]; /* Newline string when fixed */
1055 nigel 77 void *callout_data; /* To pass back to callouts */
1056     } dfa_match_data;
1057    
1058     /* Bit definitions for entries in the pcre_ctypes table. */
1059    
1060     #define ctype_space 0x01
1061     #define ctype_letter 0x02
1062     #define ctype_digit 0x04
1063     #define ctype_xdigit 0x08
1064 ph10 274 #define ctype_word 0x10 /* alphanumeric or '_' */
1065 nigel 77 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
1066    
1067     /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
1068     of bits for a class map. Some classes are built by combining these tables. */
1069    
1070     #define cbit_space 0 /* [:space:] or \s */
1071     #define cbit_xdigit 32 /* [:xdigit:] */
1072     #define cbit_digit 64 /* [:digit:] or \d */
1073     #define cbit_upper 96 /* [:upper:] */
1074     #define cbit_lower 128 /* [:lower:] */
1075     #define cbit_word 160 /* [:word:] or \w */
1076     #define cbit_graph 192 /* [:graph:] */
1077     #define cbit_print 224 /* [:print:] */
1078     #define cbit_punct 256 /* [:punct:] */
1079     #define cbit_cntrl 288 /* [:cntrl:] */
1080     #define cbit_length 320 /* Length of the cbits table */
1081    
1082     /* Offsets of the various tables from the base tables pointer, and
1083     total length. */
1084    
1085     #define lcc_offset 0
1086     #define fcc_offset 256
1087     #define cbits_offset 512
1088     #define ctypes_offset (cbits_offset + cbit_length)
1089     #define tables_length (ctypes_offset + 256)
1090    
1091 nigel 87 /* Layout of the UCP type table that translates property names into types and
1092 ph10 240 codes. Each entry used to point directly to a name, but to reduce the number of
1093 ph10 243 relocations in shared libraries, it now has an offset into a single string
1094 ph10 240 instead. */
1095 nigel 77
1096     typedef struct {
1097 ph10 243 pcre_uint16 name_offset;
1098 nigel 87 pcre_uint16 type;
1099     pcre_uint16 value;
1100 nigel 77 } ucp_type_table;
1101    
1102    
1103     /* Internal shared data tables. These are tables that are used by more than one
1104     of the exported public functions. They have to be "external" in the C sense,
1105     but are not part of the PCRE public API. The data for these tables is in the
1106     pcre_tables.c module. */
1107    
1108     extern const int _pcre_utf8_table1[];
1109     extern const int _pcre_utf8_table2[];
1110     extern const int _pcre_utf8_table3[];
1111     extern const uschar _pcre_utf8_table4[];
1112    
1113     extern const int _pcre_utf8_table1_size;
1114    
1115 ph10 240 extern const char _pcre_utt_names[];
1116 nigel 77 extern const ucp_type_table _pcre_utt[];
1117     extern const int _pcre_utt_size;
1118    
1119     extern const uschar _pcre_default_tables[];
1120    
1121     extern const uschar _pcre_OP_lengths[];
1122    
1123    
1124     /* Internal shared functions. These are functions that are used by more than
1125     one of the exported public functions. They have to be "external" in the C
1126     sense, but are not part of the PCRE public API. */
1127    
1128 ph10 150 extern BOOL _pcre_is_newline(const uschar *, int, const uschar *,
1129 ph10 149 int *, BOOL);
1130 nigel 93 extern int _pcre_ord2utf8(int, uschar *);
1131     extern real_pcre *_pcre_try_flipped(const real_pcre *, real_pcre *,
1132     const pcre_study_data *, pcre_study_data *);
1133     extern int _pcre_valid_utf8(const uschar *, int);
1134 ph10 150 extern BOOL _pcre_was_newline(const uschar *, int, const uschar *,
1135 ph10 149 int *, BOOL);
1136 nigel 93 extern BOOL _pcre_xclass(int, const uschar *);
1137 nigel 77
1138 ph10 349
1139     /* Unicode character database (UCD) */
1140    
1141     typedef struct {
1142     uschar script;
1143     uschar chartype;
1144 ph10 351 pcre_int32 other_case;
1145 ph10 349 } ucd_record;
1146    
1147 ph10 351 extern const ucd_record _pcre_ucd_records[];
1148     extern const uschar _pcre_ucd_stage1[];
1149     extern const pcre_uint16 _pcre_ucd_stage2[];
1150     extern const int _pcre_ucp_gentype[];
1151 ph10 349
1152 ph10 351
1153 ph10 349 /* UCD access macros */
1154    
1155     #define UCD_BLOCK_SIZE 128
1156 ph10 351 #define GET_UCD(ch) (_pcre_ucd_records + \
1157     _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1158 ph10 349 UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1159    
1160     #define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype
1161     #define UCD_SCRIPT(ch) GET_UCD(ch)->script
1162 ph10 351 #define UCD_CATEGORY(ch) _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1163     #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1164 ph10 349
1165 nigel 85 #endif
1166    
1167 nigel 77 /* End of pcre_internal.h */

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12