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

Contents of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 145 - (hide annotations) (download)
Wed Apr 4 14:06:52 2007 UTC (7 years, 4 months ago) by ph10
File MIME type: text/plain
File size: 44985 byte(s)
Reworked all the WIN32 __declspec stuff in the hope of getting it right.

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 117 Copyright (c) 1997-2007 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     /* Get the definitions provided by running "configure" */
71    
72     #include "config.h"
73    
74     /* Standard C headers plus the external interface definition. The only time
75     setjmp and stdarg are used is when NO_RECURSE is set. */
76    
77     #include <ctype.h>
78     #include <limits.h>
79     #include <setjmp.h>
80     #include <stdarg.h>
81     #include <stddef.h>
82     #include <stdio.h>
83     #include <stdlib.h>
84     #include <string.h>
85    
86 ph10 145 /* When compiling a DLL for Windows, the exported symbols have to be declared
87     using some MS magic. I found some useful information on this web page:
88     http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
89     information there, using __declspec(dllesport) without "extern" we have a
90     definition; with "extern" we have a declaration. The settings here override the
91     setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
92     which is all that is needed for applications, which import the symbols. We use:
93    
94     PCRE_EXP_DECL for declarations
95     PCRE_EXP_DEFN for definitions of exported functions
96     PCRE_EXP_DATA_DEFN for definitions of exported variables
97    
98     The reason for the two DEFN macros is that in non-Windows environments, one
99     does not want to have "extern" before variable definitions because it leads to
100     compiler warnings. So we distinguish between functions and variables. In
101     Windows, the two should always be the same.
102    
103     The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
104     which is an application, but needs to import this file in order to "peek" at
105     internals, can #include pcre.h first, can get an application's-eye view.
106    
107     In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
108     special-purpose environments) might want to stick other stuff in front of
109     exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
110     PCRE_EXP_DATA_DEFN only if they are not already set. */
111    
112     #ifndef PCRE_EXP_DECL
113     # ifdef _WIN32
114     # ifdef DLL_EXPORT
115     # define PCRE_EXP_DECL extern __declspec(dllexport)
116     # define PCRE_EXP_DEFN __declspec(dllexport)
117     # define PCRE_EXP_DATA_DEFN __declspec(dllexport)
118     # else
119     # define PCRE_EXP_DECL extern
120     # define PCRE_EXP_DEFN
121     # define PCRE_EXP_DATA_DEFN
122     # endif
123     #
124     # else
125     # ifdef __cplusplus
126     # define PCRE_EXP_DECL extern "C"
127     # else
128     # define PCRE_EXP_DECL extern
129     # endif
130     # ifndef PCRE_EXP_DEFN
131     # define PCRE_EXP_DEFN PCRE_EXP_DECL
132     # endif
133     # ifndef PCRE_EXP_DATA_DEFN
134     # define PCRE_EXP_DATA_DEFN
135     # endif
136     # endif
137 nigel 77 #endif
138    
139     /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
140     cannot determine these outside the compilation (e.g. by running a program as
141     part of "configure") because PCRE is often cross-compiled for use on other
142     systems. Instead we make use of the maximum sizes that are available at
143     preprocessor time in standard C environments. */
144    
145     #if USHRT_MAX == 65535
146     typedef unsigned short pcre_uint16;
147     #elif UINT_MAX == 65535
148     typedef unsigned int pcre_uint16;
149     #else
150     #error Cannot determine a type for 16-bit unsigned integers
151     #endif
152    
153     #if UINT_MAX == 4294967295
154     typedef unsigned int pcre_uint32;
155     #elif ULONG_MAX == 4294967295
156     typedef unsigned long int pcre_uint32;
157     #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     and "all" at present). The following macros are used to package up testing for
178     newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to
179     indicate in which datablock the parameters exist, and what the start/end of
180     string field names are. */
181    
182     #define NLTYPE_FIXED 0 /* Newline is a fixed length string */
183     #define NLTYPE_ANY 1 /* Newline is any Unicode line ending */
184    
185     /* This macro checks for a newline at the given position */
186    
187 nigel 91 #define IS_NEWLINE(p) \
188 nigel 93 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
189     ((p) < NLBLOCK->PSEND && \
190     _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \
191     ) \
192     : \
193     ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
194     (p)[0] == NLBLOCK->nl[0] && \
195     (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
196     ) \
197     )
198 nigel 91
199 nigel 93 /* This macro checks for a newline immediately preceding the given position */
200    
201     #define WAS_NEWLINE(p) \
202     ((NLBLOCK->nltype != NLTYPE_FIXED)? \
203     ((p) > NLBLOCK->PSSTART && \
204     _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \
205     ) \
206     : \
207     ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
208     (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
209     (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
210     ) \
211     )
212    
213 nigel 87 /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
214     with a custom type. This makes it possible, for example, to allow pcre_exec()
215     to process subject strings that are discontinuous by using a smart pointer
216     class. It must always be possible to inspect all of the subject string in
217     pcre_exec() because of the way it backtracks. Two macros are required in the
218     normal case, for sign-unspecified and unsigned char pointers. The former is
219     used for the external interface and appears in pcre.h, which is why its name
220     must begin with PCRE_. */
221    
222     #ifdef CUSTOM_SUBJECT_PTR
223     #define PCRE_SPTR CUSTOM_SUBJECT_PTR
224     #define USPTR CUSTOM_SUBJECT_PTR
225     #else
226     #define PCRE_SPTR const char *
227     #define USPTR const unsigned char *
228     #endif
229    
230 ph10 145
231    
232 nigel 85 /* Include the public PCRE header and the definitions of UCP character property
233     values. */
234 nigel 77
235 ph10 137 #include <pcre.h>
236 nigel 77 #include "ucp.h"
237    
238     /* When compiling for use with the Virtual Pascal compiler, these functions
239     need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
240     option on the command line. */
241    
242     #ifdef VPCOMPAT
243 ph10 101 #define strlen(s) _strlen(s)
244 nigel 77 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
245 ph10 101 #define memcmp(s,c,n) _memcmp(s,c,n)
246 nigel 77 #define memcpy(d,s,n) _memcpy(d,s,n)
247     #define memmove(d,s,n) _memmove(d,s,n)
248     #define memset(s,c,n) _memset(s,c,n)
249     #else /* VPCOMPAT */
250    
251     /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
252     define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
253     is set. Otherwise, include an emulating function for those systems that have
254 ph10 118 neither (there some non-Unix environments where this is the case). */
255 nigel 77
256 ph10 137 #ifndef HAVE_MEMMOVE
257 nigel 77 #undef memmove /* some systems may have a macro */
258 ph10 137 #ifdef HAVE_BCOPY
259 nigel 77 #define memmove(a, b, c) bcopy(b, a, c)
260     #else /* HAVE_BCOPY */
261 nigel 91 static void *
262 ph10 108 pcre_memmove(void *d, const void *s, size_t n)
263 nigel 77 {
264 nigel 87 size_t i;
265 ph10 108 unsigned char *dest = (unsigned char *)d;
266     const unsigned char *src = (const unsigned char *)s;
267 ph10 118 if (dest > src)
268     {
269     dest += n;
270     src += n;
271     for (i = 0; i < n; ++i) *(--dest) = *(--src);
272     return (void *)dest;
273     }
274     else
275     {
276     for (i = 0; i < n; ++i) *dest++ = *src++;
277 ph10 123 return (void *)(dest - n);
278     }
279 nigel 77 }
280     #define memmove(a, b, c) pcre_memmove(a, b, c)
281     #endif /* not HAVE_BCOPY */
282     #endif /* not HAVE_MEMMOVE */
283     #endif /* not VPCOMPAT */
284    
285    
286     /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
287     in big-endian order) by default. These are used, for example, to link from the
288     start of a subpattern to its alternatives and its end. The use of 2 bytes per
289     offset limits the size of the compiled regex to around 64K, which is big enough
290     for almost everybody. However, I received a request for an even bigger limit.
291     For this reason, and also to make the code easier to maintain, the storing and
292     loading of offsets from the byte string is now handled by the macros that are
293     defined here.
294    
295     The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
296     the config.h file, but can be overridden by using -D on the command line. This
297     is automated on Unix systems via the "configure" command. */
298    
299     #if LINK_SIZE == 2
300    
301     #define PUT(a,n,d) \
302     (a[n] = (d) >> 8), \
303     (a[(n)+1] = (d) & 255)
304    
305     #define GET(a,n) \
306     (((a)[n] << 8) | (a)[(n)+1])
307    
308     #define MAX_PATTERN_SIZE (1 << 16)
309    
310    
311     #elif LINK_SIZE == 3
312    
313     #define PUT(a,n,d) \
314     (a[n] = (d) >> 16), \
315     (a[(n)+1] = (d) >> 8), \
316     (a[(n)+2] = (d) & 255)
317    
318     #define GET(a,n) \
319     (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
320    
321     #define MAX_PATTERN_SIZE (1 << 24)
322    
323    
324     #elif LINK_SIZE == 4
325    
326     #define PUT(a,n,d) \
327     (a[n] = (d) >> 24), \
328     (a[(n)+1] = (d) >> 16), \
329     (a[(n)+2] = (d) >> 8), \
330     (a[(n)+3] = (d) & 255)
331    
332     #define GET(a,n) \
333     (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
334    
335     #define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */
336    
337    
338     #else
339     #error LINK_SIZE must be either 2, 3, or 4
340     #endif
341    
342    
343     /* Convenience macro defined in terms of the others */
344    
345     #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
346    
347    
348     /* PCRE uses some other 2-byte quantities that do not change when the size of
349     offsets changes. There are used for repeat counts and for other things such as
350     capturing parenthesis numbers in back references. */
351    
352     #define PUT2(a,n,d) \
353     a[n] = (d) >> 8; \
354     a[(n)+1] = (d) & 255
355    
356     #define GET2(a,n) \
357     (((a)[n] << 8) | (a)[(n)+1])
358    
359     #define PUT2INC(a,n,d) PUT2(a,n,d), a += 2
360    
361    
362     /* When UTF-8 encoding is being used, a character is no longer just a single
363     byte. The macros for character handling generate simple sequences when used in
364     byte-mode, and more complicated ones for UTF-8 characters. */
365    
366     #ifndef SUPPORT_UTF8
367     #define GETCHAR(c, eptr) c = *eptr;
368     #define GETCHARTEST(c, eptr) c = *eptr;
369     #define GETCHARINC(c, eptr) c = *eptr++;
370     #define GETCHARINCTEST(c, eptr) c = *eptr++;
371     #define GETCHARLEN(c, eptr, len) c = *eptr;
372     #define BACKCHAR(eptr)
373    
374     #else /* SUPPORT_UTF8 */
375    
376     /* Get the next UTF-8 character, not advancing the pointer. This is called when
377     we know we are in UTF-8 mode. */
378    
379     #define GETCHAR(c, eptr) \
380     c = *eptr; \
381 nigel 93 if (c >= 0xc0) \
382 nigel 77 { \
383     int gcii; \
384     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
385     int gcss = 6*gcaa; \
386     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
387     for (gcii = 1; gcii <= gcaa; gcii++) \
388     { \
389     gcss -= 6; \
390     c |= (eptr[gcii] & 0x3f) << gcss; \
391     } \
392     }
393    
394     /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
395     pointer. */
396    
397     #define GETCHARTEST(c, eptr) \
398     c = *eptr; \
399 nigel 93 if (utf8 && c >= 0xc0) \
400 nigel 77 { \
401     int gcii; \
402     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
403     int gcss = 6*gcaa; \
404     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
405     for (gcii = 1; gcii <= gcaa; gcii++) \
406     { \
407     gcss -= 6; \
408     c |= (eptr[gcii] & 0x3f) << gcss; \
409     } \
410     }
411    
412     /* Get the next UTF-8 character, advancing the pointer. This is called when we
413     know we are in UTF-8 mode. */
414    
415     #define GETCHARINC(c, eptr) \
416     c = *eptr++; \
417 nigel 93 if (c >= 0xc0) \
418 nigel 77 { \
419     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
420     int gcss = 6*gcaa; \
421     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
422     while (gcaa-- > 0) \
423     { \
424     gcss -= 6; \
425     c |= (*eptr++ & 0x3f) << gcss; \
426     } \
427     }
428    
429     /* Get the next character, testing for UTF-8 mode, and advancing the pointer */
430    
431     #define GETCHARINCTEST(c, eptr) \
432     c = *eptr++; \
433 nigel 93 if (utf8 && c >= 0xc0) \
434 nigel 77 { \
435     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
436     int gcss = 6*gcaa; \
437     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
438     while (gcaa-- > 0) \
439     { \
440     gcss -= 6; \
441     c |= (*eptr++ & 0x3f) << gcss; \
442     } \
443     }
444    
445     /* Get the next UTF-8 character, not advancing the pointer, incrementing length
446     if there are extra bytes. This is called when we know we are in UTF-8 mode. */
447    
448     #define GETCHARLEN(c, eptr, len) \
449     c = *eptr; \
450 nigel 93 if (c >= 0xc0) \
451 nigel 77 { \
452     int gcii; \
453     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
454     int gcss = 6*gcaa; \
455     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
456     for (gcii = 1; gcii <= gcaa; gcii++) \
457     { \
458     gcss -= 6; \
459     c |= (eptr[gcii] & 0x3f) << gcss; \
460     } \
461     len += gcaa; \
462     }
463    
464     /* If the pointer is not at the start of a character, move it back until
465     it is. Called only in UTF-8 mode. */
466    
467     #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;
468    
469     #endif
470    
471    
472     /* In case there is no definition of offsetof() provided - though any proper
473     Standard C system should have one. */
474    
475     #ifndef offsetof
476     #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
477     #endif
478    
479    
480     /* These are the public options that can change during matching. */
481    
482     #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
483    
484 nigel 91 /* Private options flags start at the most significant end of the four bytes.
485     The public options defined in pcre.h start at the least significant end. Make
486     sure they don't overlap! The bits are getting a bit scarce now -- when we run
487     out, there is a dummy word in the structure that could be used for the private
488     bits. */
489 nigel 77
490 nigel 91 #define PCRE_NOPARTIAL 0x80000000 /* can't use partial with this regex */
491 nigel 77 #define PCRE_FIRSTSET 0x40000000 /* first_byte is set */
492     #define PCRE_REQCHSET 0x20000000 /* req_byte is set */
493     #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
494 nigel 91 #define PCRE_JCHANGED 0x08000000 /* j option changes within regex */
495 nigel 77
496     /* Options for the "extra" block produced by pcre_study(). */
497    
498     #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
499    
500     /* Masks for identifying the public options that are permitted at compile
501     time, run time, or study time, respectively. */
502    
503 nigel 93 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)
504    
505 nigel 77 #define PUBLIC_OPTIONS \
506     (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
507     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
508 nigel 91 PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
509 nigel 93 PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
510 nigel 77
511     #define PUBLIC_EXEC_OPTIONS \
512     (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
513 nigel 93 PCRE_PARTIAL|PCRE_NEWLINE_BITS)
514 nigel 77
515     #define PUBLIC_DFA_EXEC_OPTIONS \
516     (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
517 nigel 93 PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
518 nigel 77
519     #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
520    
521     /* Magic number to provide a small check against being handed junk. Also used
522     to detect whether a pattern was compiled on a host of different endianness. */
523    
524     #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
525    
526     /* Negative values for the firstchar and reqchar variables */
527    
528     #define REQ_UNSET (-2)
529     #define REQ_NONE (-1)
530    
531     /* The maximum remaining length of subject we are prepared to search for a
532     req_byte match. */
533    
534     #define REQ_BYTE_MAX 1000
535    
536     /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a
537     variable-length repeat, or a anything other than literal characters. */
538    
539     #define REQ_CASELESS 0x0100 /* indicates caselessness */
540     #define REQ_VARY 0x0200 /* reqbyte followed non-literal item */
541    
542     /* Miscellaneous definitions */
543    
544     typedef int BOOL;
545    
546     #define FALSE 0
547     #define TRUE 1
548    
549 nigel 93 /* Escape items that are just an encoding of a particular data value. */
550 nigel 77
551     #ifndef ESC_e
552     #define ESC_e 27
553     #endif
554    
555     #ifndef ESC_f
556     #define ESC_f '\f'
557     #endif
558    
559     #ifndef ESC_n
560 nigel 93 #define ESC_n '\n'
561 nigel 77 #endif
562    
563     #ifndef ESC_r
564     #define ESC_r '\r'
565     #endif
566    
567     /* We can't officially use ESC_t because it is a POSIX reserved identifier
568     (presumably because of all the others like size_t). */
569    
570     #ifndef ESC_tee
571     #define ESC_tee '\t'
572     #endif
573    
574 nigel 87 /* Codes for different types of Unicode property */
575    
576     #define PT_ANY 0 /* Any property - matches all chars */
577     #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
578     #define PT_GC 2 /* General characteristic (e.g. L) */
579     #define PT_PC 3 /* Particular characteristic (e.g. Lu) */
580     #define PT_SC 4 /* Script (e.g. Han) */
581    
582     /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
583     contain UTF-8 characters with values greater than 255. */
584    
585     #define XCL_NOT 0x01 /* Flag: this is a negative class */
586     #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
587    
588     #define XCL_END 0 /* Marks end of individual items */
589     #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
590     #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
591     #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
592     #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
593    
594 nigel 77 /* These are escaped items that aren't just an encoding of a particular data
595     value such as \n. They must have non-zero values, as check_escape() returns
596     their negation. Also, they must appear in the same order as in the opcode
597     definitions below, up to ESC_z. There's a dummy for OP_ANY because it
598     corresponds to "." rather than an escape sequence. The final one must be
599 nigel 93 ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
600     There are two tests in the code for an escape greater than ESC_b and less than
601     ESC_Z to detect the types that may be repeated. These are the types that
602     consume characters. If any new escapes are put in between that don't consume a
603 nigel 77 character, that code will have to change. */
604    
605     enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,
606 nigel 93 ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,
607     ESC_E, ESC_Q, ESC_k, ESC_REF };
608 nigel 77
609 nigel 93
610 nigel 77 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
611     that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
612     OP_EOD must correspond in order to the list of escapes immediately above.
613    
614 nigel 93 To keep stored, compiled patterns compatible, new opcodes should be added
615     immediately before OP_BRA, where (since release 7.0) a gap is left for this
616     purpose.
617    
618     *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
619     that follow must also be updated to match. There is also a table called
620     "coptable" in pcre_dfa_exec.c that must be updated. */
621    
622 nigel 77 enum {
623     OP_END, /* 0 End of pattern */
624    
625     /* Values corresponding to backslashed metacharacters */
626    
627     OP_SOD, /* 1 Start of data: \A */
628     OP_SOM, /* 2 Start of match (subject + offset): \G */
629     OP_NOT_WORD_BOUNDARY, /* 3 \B */
630     OP_WORD_BOUNDARY, /* 4 \b */
631     OP_NOT_DIGIT, /* 5 \D */
632     OP_DIGIT, /* 6 \d */
633     OP_NOT_WHITESPACE, /* 7 \S */
634     OP_WHITESPACE, /* 8 \s */
635     OP_NOT_WORDCHAR, /* 9 \W */
636     OP_WORDCHAR, /* 10 \w */
637     OP_ANY, /* 11 Match any character */
638     OP_ANYBYTE, /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */
639     OP_NOTPROP, /* 13 \P (not Unicode property) */
640     OP_PROP, /* 14 \p (Unicode property) */
641 nigel 93 OP_ANYNL, /* 15 \R (any newline sequence) */
642     OP_EXTUNI, /* 16 \X (extended Unicode sequence */
643     OP_EODN, /* 17 End of data or \n at end of data: \Z. */
644     OP_EOD, /* 18 End of data: \z */
645 nigel 77
646 nigel 93 OP_OPT, /* 19 Set runtime options */
647     OP_CIRC, /* 20 Start of line - varies with multiline switch */
648     OP_DOLL, /* 21 End of line - varies with multiline switch */
649     OP_CHAR, /* 22 Match one character, casefully */
650     OP_CHARNC, /* 23 Match one character, caselessly */
651     OP_NOT, /* 24 Match one character, not the following one */
652 nigel 77
653 nigel 93 OP_STAR, /* 25 The maximizing and minimizing versions of */
654     OP_MINSTAR, /* 26 these six opcodes must come in pairs, with */
655     OP_PLUS, /* 27 the minimizing one second. */
656     OP_MINPLUS, /* 28 This first set applies to single characters.*/
657     OP_QUERY, /* 29 */
658     OP_MINQUERY, /* 30 */
659 nigel 77
660 nigel 93 OP_UPTO, /* 31 From 0 to n matches */
661     OP_MINUPTO, /* 32 */
662     OP_EXACT, /* 33 Exactly n matches */
663 nigel 77
664 nigel 93 OP_POSSTAR, /* 34 Possessified star */
665     OP_POSPLUS, /* 35 Possessified plus */
666     OP_POSQUERY, /* 36 Posesssified query */
667     OP_POSUPTO, /* 37 Possessified upto */
668 nigel 77
669 nigel 93 OP_NOTSTAR, /* 38 The maximizing and minimizing versions of */
670     OP_NOTMINSTAR, /* 39 these six opcodes must come in pairs, with */
671     OP_NOTPLUS, /* 40 the minimizing one second. They must be in */
672     OP_NOTMINPLUS, /* 41 exactly the same order as those above. */
673     OP_NOTQUERY, /* 42 This set applies to "not" single characters. */
674     OP_NOTMINQUERY, /* 43 */
675 nigel 77
676 nigel 93 OP_NOTUPTO, /* 44 From 0 to n matches */
677     OP_NOTMINUPTO, /* 45 */
678     OP_NOTEXACT, /* 46 Exactly n matches */
679    
680     OP_NOTPOSSTAR, /* 47 Possessified versions */
681     OP_NOTPOSPLUS, /* 48 */
682     OP_NOTPOSQUERY, /* 49 */
683     OP_NOTPOSUPTO, /* 50 */
684    
685     OP_TYPESTAR, /* 51 The maximizing and minimizing versions of */
686     OP_TYPEMINSTAR, /* 52 these six opcodes must come in pairs, with */
687     OP_TYPEPLUS, /* 53 the minimizing one second. These codes must */
688     OP_TYPEMINPLUS, /* 54 be in exactly the same order as those above. */
689     OP_TYPEQUERY, /* 55 This set applies to character types such as \d */
690     OP_TYPEMINQUERY, /* 56 */
691    
692     OP_TYPEUPTO, /* 57 From 0 to n matches */
693     OP_TYPEMINUPTO, /* 58 */
694     OP_TYPEEXACT, /* 59 Exactly n matches */
695    
696     OP_TYPEPOSSTAR, /* 60 Possessified versions */
697     OP_TYPEPOSPLUS, /* 61 */
698     OP_TYPEPOSQUERY, /* 62 */
699     OP_TYPEPOSUPTO, /* 63 */
700    
701     OP_CRSTAR, /* 64 The maximizing and minimizing versions of */
702     OP_CRMINSTAR, /* 65 all these opcodes must come in pairs, with */
703     OP_CRPLUS, /* 66 the minimizing one second. These codes must */
704     OP_CRMINPLUS, /* 67 be in exactly the same order as those above. */
705     OP_CRQUERY, /* 68 These are for character classes and back refs */
706     OP_CRMINQUERY, /* 69 */
707     OP_CRRANGE, /* 70 These are different to the three sets above. */
708     OP_CRMINRANGE, /* 71 */
709    
710     OP_CLASS, /* 72 Match a character class, chars < 256 only */
711     OP_NCLASS, /* 73 Same, but the bitmap was created from a negative
712 nigel 77 class - the difference is relevant only when a UTF-8
713     character > 255 is encountered. */
714    
715 nigel 93 OP_XCLASS, /* 74 Extended class for handling UTF-8 chars within the
716 nigel 77 class. This does both positive and negative. */
717    
718 nigel 93 OP_REF, /* 75 Match a back reference */
719     OP_RECURSE, /* 76 Match a numbered subpattern (possibly recursive) */
720     OP_CALLOUT, /* 77 Call out to external function if provided */
721 nigel 77
722 nigel 93 OP_ALT, /* 78 Start of alternation */
723     OP_KET, /* 79 End of group that doesn't have an unbounded repeat */
724     OP_KETRMAX, /* 80 These two must remain together and in this */
725     OP_KETRMIN, /* 81 order. They are for groups the repeat for ever. */
726 nigel 77
727 nigel 93 /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
728 nigel 77
729 nigel 93 OP_ASSERT, /* 82 Positive lookahead */
730     OP_ASSERT_NOT, /* 83 Negative lookahead */
731     OP_ASSERTBACK, /* 84 Positive lookbehind */
732     OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */
733     OP_REVERSE, /* 86 Move pointer back - used in lookbehind assertions */
734 nigel 77
735 nigel 93 /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
736     as there's a test for >= ONCE for a subpattern that isn't an assertion. */
737 nigel 77
738 nigel 93 OP_ONCE, /* 87 Atomic group */
739     OP_BRA, /* 88 Start of non-capturing bracket */
740     OP_CBRA, /* 89 Start of capturing bracket */
741     OP_COND, /* 90 Conditional group */
742 nigel 77
743 nigel 93 /* These three must follow the previous three, in the same order. There's a
744     check for >= SBRA to distinguish the two sets. */
745 nigel 77
746 nigel 93 OP_SBRA, /* 91 Start of non-capturing bracket, check empty */
747     OP_SCBRA, /* 92 Start of capturing bracket, check empty */
748     OP_SCOND, /* 93 Conditional group, check empty */
749 nigel 77
750 nigel 93 OP_CREF, /* 94 Used to hold a capture number as condition */
751     OP_RREF, /* 95 Used to hold a recursion number as condition */
752     OP_DEF, /* 96 The DEFINE condition */
753    
754     OP_BRAZERO, /* 97 These two must remain together and in this */
755     OP_BRAMINZERO /* 98 order. */
756 nigel 77 };
757    
758    
759     /* This macro defines textual names for all the opcodes. These are used only
760     for debugging. The macro is referenced only in pcre_printint.c. */
761    
762     #define OP_NAME_LIST \
763     "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d", \
764     "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte", \
765 nigel 93 "notprop", "prop", "anynl", "extuni", \
766 nigel 77 "\\Z", "\\z", \
767     "Opt", "^", "$", "char", "charnc", "not", \
768     "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
769 nigel 93 "*+","++", "?+", "{", \
770 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
771 nigel 93 "*+","++", "?+", "{", \
772 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
773 nigel 93 "*+","++", "?+", "{", \
774 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", \
775     "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \
776     "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \
777 nigel 93 "AssertB", "AssertB not", "Reverse", \
778     "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond", \
779     "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
780 nigel 77
781    
782     /* This macro defines the length of fixed length operations in the compiled
783     regex. The lengths are used when searching for specific things, and also in the
784     debugging printing of a compiled regex. We use a macro so that it can be
785     defined close to the definitions of the opcodes themselves.
786    
787     As things have been extended, some of these are no longer fixed lenths, but are
788     minima instead. For example, the length of a single-character repeat may vary
789     in UTF-8 mode. The code that uses this table must know about such things. */
790    
791     #define OP_LENGTHS \
792     1, /* End */ \
793     1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \
794     1, 1, /* Any, Anybyte */ \
795 nigel 93 3, 3, 1, 1, /* NOTPROP, PROP, EXTUNI, ANYNL */ \
796 nigel 77 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \
797     2, /* Char - the minimum length */ \
798     2, /* Charnc - the minimum length */ \
799     2, /* not */ \
800     /* Positive single-char repeats ** These are */ \
801     2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
802     4, 4, 4, /* upto, minupto, exact ** UTF-8 mode */ \
803 nigel 93 2, 2, 2, 4, /* *+, ++, ?+, upto+ */ \
804 nigel 77 /* Negative single-char repeats - only for chars < 256 */ \
805     2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
806     4, 4, 4, /* NOT upto, minupto, exact */ \
807 nigel 93 2, 2, 2, 4, /* Possessive *, +, ?, upto */ \
808 nigel 77 /* Positive type repeats */ \
809     2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
810     4, 4, 4, /* Type upto, minupto, exact */ \
811 nigel 93 2, 2, 2, 4, /* Possessive *+, ++, ?+, upto+ */ \
812 nigel 77 /* Character class & ref repeats */ \
813     1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
814     5, 5, /* CRRANGE, CRMINRANGE */ \
815     33, /* CLASS */ \
816     33, /* NCLASS */ \
817     0, /* XCLASS - variable length */ \
818     3, /* REF */ \
819     1+LINK_SIZE, /* RECURSE */ \
820     2+2*LINK_SIZE, /* CALLOUT */ \
821     1+LINK_SIZE, /* Alt */ \
822     1+LINK_SIZE, /* Ket */ \
823     1+LINK_SIZE, /* KetRmax */ \
824     1+LINK_SIZE, /* KetRmin */ \
825     1+LINK_SIZE, /* Assert */ \
826     1+LINK_SIZE, /* Assert not */ \
827     1+LINK_SIZE, /* Assert behind */ \
828     1+LINK_SIZE, /* Assert behind not */ \
829     1+LINK_SIZE, /* Reverse */ \
830 nigel 93 1+LINK_SIZE, /* ONCE */ \
831     1+LINK_SIZE, /* BRA */ \
832     3+LINK_SIZE, /* CBRA */ \
833 nigel 77 1+LINK_SIZE, /* COND */ \
834 nigel 93 1+LINK_SIZE, /* SBRA */ \
835     3+LINK_SIZE, /* SCBRA */ \
836     1+LINK_SIZE, /* SCOND */ \
837 nigel 77 3, /* CREF */ \
838 nigel 93 3, /* RREF */ \
839     1, /* DEF */ \
840 nigel 77 1, 1, /* BRAZERO, BRAMINZERO */ \
841    
842    
843 nigel 93 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
844 nigel 77
845 nigel 93 #define RREF_ANY 0xffff
846 nigel 77
847     /* Error code numbers. They are given names so that they can more easily be
848     tracked. */
849    
850     enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
851     ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
852     ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
853     ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
854 nigel 91 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
855 nigel 93 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };
856 nigel 77
857     /* The real format of the start of the pcre block; the index of names and the
858     code vector run on as long as necessary after the end. We store an explicit
859     offset to the name table so that if a regex is compiled on one host, saved, and
860     then run on another where the size of pointers is different, all might still
861     be well. For the case of compiled-on-4 and run-on-8, we include an extra
862     pointer that is always NULL. For future-proofing, a few dummy fields were
863     originally included - even though you can never get this planning right - but
864     there is only one left now.
865    
866     NOTE NOTE NOTE:
867     Because people can now save and re-use compiled patterns, any additions to this
868     structure should be made at the end, and something earlier (e.g. a new
869     flag in the options or one of the dummy fields) should indicate that the new
870     fields are present. Currently PCRE always sets the dummy fields to zero.
871     NOTE NOTE NOTE:
872     */
873    
874     typedef struct real_pcre {
875     pcre_uint32 magic_number;
876     pcre_uint32 size; /* Total that was malloced */
877     pcre_uint32 options;
878     pcre_uint32 dummy1; /* For future use, maybe */
879    
880     pcre_uint16 top_bracket;
881     pcre_uint16 top_backref;
882     pcre_uint16 first_byte;
883     pcre_uint16 req_byte;
884     pcre_uint16 name_table_offset; /* Offset to name table that follows */
885     pcre_uint16 name_entry_size; /* Size of any name items */
886     pcre_uint16 name_count; /* Number of name items */
887     pcre_uint16 ref_count; /* Reference count */
888    
889     const unsigned char *tables; /* Pointer to tables or NULL for std */
890     const unsigned char *nullpad; /* NULL padding */
891     } real_pcre;
892    
893     /* The format of the block used to store data from pcre_study(). The same
894     remark (see NOTE above) about extending this structure applies. */
895    
896     typedef struct pcre_study_data {
897     pcre_uint32 size; /* Total that was malloced */
898     pcre_uint32 options;
899     uschar start_bits[32];
900     } pcre_study_data;
901    
902     /* Structure for passing "static" information around between the functions
903     doing the compiling, so that they are thread-safe. */
904    
905     typedef struct compile_data {
906     const uschar *lcc; /* Points to lower casing table */
907     const uschar *fcc; /* Points to case-flipping table */
908     const uschar *cbits; /* Points to character type table */
909     const uschar *ctypes; /* Points to table of type maps */
910 nigel 93 const uschar *start_workspace;/* The start of working space */
911 nigel 77 const uschar *start_code; /* The start of the compiled code */
912     const uschar *start_pattern; /* The start of the pattern */
913 nigel 93 const uschar *end_pattern; /* The end of the pattern */
914     uschar *hwm; /* High watermark of workspace */
915 nigel 77 uschar *name_table; /* The name/number table */
916     int names_found; /* Number of entries so far */
917     int name_entry_size; /* Size of each entry */
918 nigel 93 int bracount; /* Count of capturing parens */
919 nigel 77 int top_backref; /* Maximum back reference */
920     unsigned int backref_map; /* Bitmap of low back refs */
921 nigel 93 int external_options; /* External (initial) options */
922 nigel 77 int req_varyopt; /* "After variable item" flag for reqbyte */
923     BOOL nopartial; /* Set TRUE if partial won't work */
924 nigel 93 int nltype; /* Newline type */
925     int nllen; /* Newline string length */
926     uschar nl[4]; /* Newline string when fixed length */
927 nigel 77 } compile_data;
928    
929     /* Structure for maintaining a chain of pointers to the currently incomplete
930     branches, for testing for left recursion. */
931    
932     typedef struct branch_chain {
933     struct branch_chain *outer;
934     uschar *current;
935     } branch_chain;
936    
937     /* Structure for items in a linked list that represents an explicit recursive
938     call within the pattern. */
939    
940     typedef struct recursion_info {
941     struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
942     int group_num; /* Number of group that was called */
943     const uschar *after_call; /* "Return value": points after the call in the expr */
944 nigel 87 USPTR save_start; /* Old value of md->start_match */
945 nigel 77 int *offset_save; /* Pointer to start of saved offsets */
946     int saved_max; /* Number of saved offsets */
947     } recursion_info;
948    
949     /* When compiling in a mode that doesn't use recursive calls to match(),
950     a structure is used to remember local variables on the heap. It is defined in
951 nigel 91 pcre_exec.c, close to the match() function, so that it is easy to keep it in
952     step with any changes of local variable. However, the pointer to the current
953     frame must be saved in some "static" place over a longjmp(). We declare the
954     structure here so that we can put a pointer in the match_data structure. NOTE:
955     This isn't used for a "normal" compilation of pcre. */
956 nigel 77
957     struct heapframe;
958    
959 nigel 93 /* Structure for building a chain of data for holding the values of the subject
960     pointer at the start of each subpattern, so as to detect when an empty string
961     has been matched by a subpattern - to break infinite loops. */
962    
963     typedef struct eptrblock {
964     struct eptrblock *epb_prev;
965     USPTR epb_saved_eptr;
966     } eptrblock;
967    
968    
969 nigel 77 /* Structure for passing "static" information around between the functions
970     doing traditional NFA matching, so that they are thread-safe. */
971    
972     typedef struct match_data {
973 nigel 87 unsigned long int match_call_count; /* As it says */
974     unsigned long int match_limit; /* As it says */
975     unsigned long int match_limit_recursion; /* As it says */
976 nigel 77 int *offset_vector; /* Offset vector */
977     int offset_end; /* One past the end */
978     int offset_max; /* The maximum usable for return data */
979 nigel 93 int nltype; /* Newline type */
980     int nllen; /* Newline string length */
981     uschar nl[4]; /* Newline string when fixed */
982 nigel 77 const uschar *lcc; /* Points to lower casing table */
983     const uschar *ctypes; /* Points to table of type maps */
984     BOOL offset_overflow; /* Set if too many extractions */
985     BOOL notbol; /* NOTBOL flag */
986     BOOL noteol; /* NOTEOL flag */
987     BOOL utf8; /* UTF8 flag */
988     BOOL endonly; /* Dollar not before final \n */
989     BOOL notempty; /* Empty string match not wanted */
990     BOOL partial; /* PARTIAL flag */
991     BOOL hitend; /* Hit the end of the subject at some point */
992     const uschar *start_code; /* For use when recursing */
993 nigel 87 USPTR start_subject; /* Start of the subject string */
994     USPTR end_subject; /* End of the subject string */
995     USPTR start_match; /* Start of this match attempt */
996     USPTR end_match_ptr; /* Subject position at end match */
997 nigel 77 int end_offset_top; /* Highwater mark at end of match */
998     int capture_last; /* Most recent capture number */
999     int start_offset; /* The start offset value */
1000 nigel 93 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
1001     int eptrn; /* Next free eptrblock */
1002 nigel 77 recursion_info *recursive; /* Linked list of recursion data */
1003     void *callout_data; /* To pass back to callouts */
1004     struct heapframe *thisframe; /* Used only when compiling for no recursion */
1005     } match_data;
1006    
1007     /* A similar structure is used for the same purpose by the DFA matching
1008     functions. */
1009    
1010     typedef struct dfa_match_data {
1011     const uschar *start_code; /* Start of the compiled pattern */
1012     const uschar *start_subject; /* Start of the subject string */
1013     const uschar *end_subject; /* End of subject string */
1014     const uschar *tables; /* Character tables */
1015     int moptions; /* Match options */
1016     int poptions; /* Pattern options */
1017 nigel 93 int nltype; /* Newline type */
1018     int nllen; /* Newline string length */
1019     uschar nl[4]; /* Newline string when fixed */
1020 nigel 77 void *callout_data; /* To pass back to callouts */
1021     } dfa_match_data;
1022    
1023     /* Bit definitions for entries in the pcre_ctypes table. */
1024    
1025     #define ctype_space 0x01
1026     #define ctype_letter 0x02
1027     #define ctype_digit 0x04
1028     #define ctype_xdigit 0x08
1029     #define ctype_word 0x10 /* alphameric or '_' */
1030     #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
1031    
1032     /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
1033     of bits for a class map. Some classes are built by combining these tables. */
1034    
1035     #define cbit_space 0 /* [:space:] or \s */
1036     #define cbit_xdigit 32 /* [:xdigit:] */
1037     #define cbit_digit 64 /* [:digit:] or \d */
1038     #define cbit_upper 96 /* [:upper:] */
1039     #define cbit_lower 128 /* [:lower:] */
1040     #define cbit_word 160 /* [:word:] or \w */
1041     #define cbit_graph 192 /* [:graph:] */
1042     #define cbit_print 224 /* [:print:] */
1043     #define cbit_punct 256 /* [:punct:] */
1044     #define cbit_cntrl 288 /* [:cntrl:] */
1045     #define cbit_length 320 /* Length of the cbits table */
1046    
1047     /* Offsets of the various tables from the base tables pointer, and
1048     total length. */
1049    
1050     #define lcc_offset 0
1051     #define fcc_offset 256
1052     #define cbits_offset 512
1053     #define ctypes_offset (cbits_offset + cbit_length)
1054     #define tables_length (ctypes_offset + 256)
1055    
1056 nigel 87 /* Layout of the UCP type table that translates property names into types and
1057     codes. */
1058 nigel 77
1059     typedef struct {
1060     const char *name;
1061 nigel 87 pcre_uint16 type;
1062     pcre_uint16 value;
1063 nigel 77 } ucp_type_table;
1064    
1065    
1066     /* Internal shared data tables. These are tables that are used by more than one
1067     of the exported public functions. They have to be "external" in the C sense,
1068     but are not part of the PCRE public API. The data for these tables is in the
1069     pcre_tables.c module. */
1070    
1071     extern const int _pcre_utf8_table1[];
1072     extern const int _pcre_utf8_table2[];
1073     extern const int _pcre_utf8_table3[];
1074     extern const uschar _pcre_utf8_table4[];
1075    
1076     extern const int _pcre_utf8_table1_size;
1077    
1078     extern const ucp_type_table _pcre_utt[];
1079     extern const int _pcre_utt_size;
1080    
1081     extern const uschar _pcre_default_tables[];
1082    
1083     extern const uschar _pcre_OP_lengths[];
1084    
1085    
1086     /* Internal shared functions. These are functions that are used by more than
1087     one of the exported public functions. They have to be "external" in the C
1088     sense, but are not part of the PCRE public API. */
1089    
1090 nigel 93 extern BOOL _pcre_is_newline(const uschar *, const uschar *, int *,
1091     BOOL);
1092     extern int _pcre_ord2utf8(int, uschar *);
1093     extern real_pcre *_pcre_try_flipped(const real_pcre *, real_pcre *,
1094     const pcre_study_data *, pcre_study_data *);
1095     extern int _pcre_ucp_findprop(const unsigned int, int *, int *);
1096     extern unsigned int _pcre_ucp_othercase(const unsigned int);
1097     extern int _pcre_valid_utf8(const uschar *, int);
1098     extern BOOL _pcre_was_newline(const uschar *, const uschar *, int *,
1099     BOOL);
1100     extern BOOL _pcre_xclass(int, const uschar *);
1101 nigel 77
1102 nigel 85 #endif
1103    
1104 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