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

Diff of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 733 by ph10, Tue Oct 11 10:29:36 2011 UTC revision 1359 by ph10, Tue Sep 3 10:10:59 2013 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2011 University of Cambridge             Copyright (c) 1997-2013 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 40  POSSIBILITY OF SUCH DAMAGE. Line 40  POSSIBILITY OF SUCH DAMAGE.
40    
41  /* This header contains definitions that are shared between the different  /* This header contains definitions that are shared between the different
42  modules, but which are not relevant to the exported API. This includes some  modules, but which are not relevant to the exported API. This includes some
43  functions whose names all begin with "_pcre_". */  functions whose names all begin with "_pcre_", "_pcre16_" or "_pcre32_"
44    depending on the PRIV macro. */
45    
46  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
47  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
# Line 51  functions whose names all begin with "_p Line 52  functions whose names all begin with "_p
52  #define PCRE_DEBUG  #define PCRE_DEBUG
53  #endif  #endif
54    
55  /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"  /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
 script prevents both being selected, but not everybody uses "configure". */  
56    
57  #if defined EBCDIC && defined SUPPORT_UTF8  #if !defined COMPILE_PCRE16 && !defined COMPILE_PCRE32
58  #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.  #define COMPILE_PCRE8
59  #endif  #endif
60    
61  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The  /* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The
62  "configure" script ensures this, but not everybody uses "configure". */  "configure" script ensures this, but not everybody uses "configure". */
63    
64  #if defined SUPPORT_UCP && !defined SUPPORT_UTF8  #if defined SUPPORT_UCP && !(defined SUPPORT_UTF)
65    #define SUPPORT_UTF 1
66    #endif
67    
68    /* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility
69    reasons with existing code. */
70    
71    #if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF)
72    #define SUPPORT_UTF 1
73    #endif
74    
75    /* Fixme: SUPPORT_UTF8 should be eventually disappear from the code.
76    Until then we define it if SUPPORT_UTF is defined. */
77    
78    #if defined SUPPORT_UTF && !(defined SUPPORT_UTF8)
79  #define SUPPORT_UTF8 1  #define SUPPORT_UTF8 1
80  #endif  #endif
81    
82    /* We do not support both EBCDIC and UTF-8/16/32 at the same time. The "configure"
83    script prevents both being selected, but not everybody uses "configure". */
84    
85    #if defined EBCDIC && defined SUPPORT_UTF
86    #error The use of both EBCDIC and SUPPORT_UTF is not supported.
87    #endif
88    
89  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
90  inline, and there are *still* stupid compilers about that don't like indented  inline, and there are *still* stupid compilers about that don't like indented
91  pre-processor statements, or at least there were when I first wrote this. After  pre-processor statements, or at least there were when I first wrote this. After
# Line 91  setjmp and stdarg are used is when NO_RE Line 112  setjmp and stdarg are used is when NO_RE
112  #include <stdlib.h>  #include <stdlib.h>
113  #include <string.h>  #include <string.h>
114    
115    /* Valgrind (memcheck) support */
116    
117    #ifdef SUPPORT_VALGRIND
118    #include <valgrind/memcheck.h>
119    #endif
120    
121  /* When compiling a DLL for Windows, the exported symbols have to be declared  /* When compiling a DLL for Windows, the exported symbols have to be declared
122  using some MS magic. I found some useful information on this web page:  using some MS magic. I found some useful information on this web page:
123  http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the  http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
# Line 158  set, we ensure here that it has no effec Line 185  set, we ensure here that it has no effec
185  #define PCRE_CALL_CONVENTION  #define PCRE_CALL_CONVENTION
186  #endif  #endif
187    
188  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
189  cannot determine these outside the compilation (e.g. by running a program as  cannot determine these outside the compilation (e.g. by running a program as
190  part of "configure") because PCRE is often cross-compiled for use on other  part of "configure") because PCRE is often cross-compiled for use on other
191  systems. Instead we make use of the maximum sizes that are available at  systems. Instead we make use of the maximum sizes that are available at
192  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
193    
194    typedef unsigned char pcre_uint8;
195    
196  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
197    typedef unsigned short pcre_uint16;  typedef unsigned short pcre_uint16;
198    typedef short pcre_int16;  typedef short pcre_int16;
199    #define PCRE_UINT16_MAX USHRT_MAX
200    #define PCRE_INT16_MAX SHRT_MAX
201  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
202    typedef unsigned int pcre_uint16;  typedef unsigned int pcre_uint16;
203    typedef int pcre_int16;  typedef int pcre_int16;
204    #define PCRE_UINT16_MAX UINT_MAX
205    #define PCRE_INT16_MAX INT_MAX
206  #else  #else
207    #error Cannot determine a type for 16-bit unsigned integers  #error Cannot determine a type for 16-bit integers
208  #endif  #endif
209    
210  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295U
211    typedef unsigned int pcre_uint32;  typedef unsigned int pcre_uint32;
212    typedef int pcre_int32;  typedef int pcre_int32;
213  #elif ULONG_MAX == 4294967295  #define PCRE_UINT32_MAX UINT_MAX
214    typedef unsigned long int pcre_uint32;  #define PCRE_INT32_MAX INT_MAX
215    typedef long int pcre_int32;  #elif ULONG_MAX == 4294967295UL
216    typedef unsigned long int pcre_uint32;
217    typedef long int pcre_int32;
218    #define PCRE_UINT32_MAX ULONG_MAX
219    #define PCRE_INT32_MAX LONG_MAX
220  #else  #else
221    #error Cannot determine a type for 32-bit unsigned integers  #error Cannot determine a type for 32-bit integers
222  #endif  #endif
223    
224  /* When checking for integer overflow in pcre_compile(), we need to handle  /* When checking for integer overflow in pcre_compile(), we need to handle
# Line 192  stdint.h is available, include it; it ma Line 229  stdint.h is available, include it; it ma
229  have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set  have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
230  by "configure". */  by "configure". */
231    
232  #if HAVE_STDINT_H  #if defined HAVE_STDINT_H
233  #include <stdint.h>  #include <stdint.h>
234  #elif HAVE_INTTYPES_H  #elif defined HAVE_INTTYPES_H
235  #include <inttypes.h>  #include <inttypes.h>
236  #endif  #endif
237    
# Line 206  by "configure". */ Line 243  by "configure". */
243    
244  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
245  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
246  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char * or short *,
247  should make things easier for callers. We define a short type for unsigned char  because that should make things easier for callers. This character type is
248  to save lots of typing. I tried "uchar", but it causes problems on Digital  called pcre_uchar.
249  Unix, where it is defined in sys/types, so use "uschar" instead. */  
250    The IN_UCHARS macro multiply its argument with the byte size of the current
251    pcre_uchar type. Useful for memcpy and such operations, whose require the
252    byte size of their input/output buffers.
253    
254    The MAX_255 macro checks whether its pcre_uchar input is less than 256.
255    
256    The TABLE_GET macro is designed for accessing elements of tables whose contain
257    exactly 256 items. When the character is able to contain more than 256
258    items, some check is needed before accessing these tables.
259    */
260    
261    #if defined COMPILE_PCRE8
262    
263  typedef unsigned char uschar;  typedef unsigned char pcre_uchar;
264    #define IN_UCHARS(x) (x)
265    #define MAX_255(c) 1
266    #define TABLE_GET(c, table, default) ((table)[c])
267    
268    #elif defined COMPILE_PCRE16
269    
270    #if USHRT_MAX != 65535
271    /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
272    pcre.h(.in) and disable (comment out) this message. */
273    #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
274    #endif
275    
276    typedef pcre_uint16 pcre_uchar;
277    #define UCHAR_SHIFT (1)
278    #define IN_UCHARS(x) ((x) << UCHAR_SHIFT)
279    #define MAX_255(c) ((c) <= 255u)
280    #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
281    
282    #elif defined COMPILE_PCRE32
283    
284    typedef pcre_uint32 pcre_uchar;
285    #define UCHAR_SHIFT (2)
286    #define IN_UCHARS(x) ((x) << UCHAR_SHIFT)
287    #define MAX_255(c) ((c) <= 255u)
288    #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
289    
290    #else
291    #error Unsupported compiling mode
292    #endif /* COMPILE_PCRE[8|16|32] */
293    
294  /* This is an unsigned int value that no character can ever have. UTF-8  /* This is an unsigned int value that no character can ever have. UTF-8
295  characters only go up to 0x7fffffff (though Unicode doesn't go beyond  characters only go up to 0x7fffffff (though Unicode doesn't go beyond
# Line 234  start/end of string field names are. */ Line 312  start/end of string field names are. */
312  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
313    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
314      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
315       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\       PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
316         utf8)) \         &(NLBLOCK->nllen), utf)) \
317      : \      : \
318      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
319       (p)[0] == NLBLOCK->nl[0] && \       RAWUCHARTEST(p) == NLBLOCK->nl[0] && \
320       (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \       (NLBLOCK->nllen == 1 || RAWUCHARTEST(p+1) == NLBLOCK->nl[1])       \
321      ) \      ) \
322    )    )
323    
# Line 248  start/end of string field names are. */ Line 326  start/end of string field names are. */
326  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
327    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
328      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
329       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \       PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
330         &(NLBLOCK->nllen), utf8)) \         &(NLBLOCK->nllen), utf)) \
331      : \      : \
332      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
333       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       RAWUCHARTEST(p - NLBLOCK->nllen) == NLBLOCK->nl[0] &&              \
334       (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \       (NLBLOCK->nllen == 1 || RAWUCHARTEST(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \
335      ) \      ) \
336    )    )
337    
# Line 267  used for the external interface and appe Line 345  used for the external interface and appe
345  must begin with PCRE_. */  must begin with PCRE_. */
346    
347  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
348  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
349  #else  #else
350  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
351  #endif  #endif
352    
   
   
353  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
354  values. */  values. */
355    
356  #include "pcre.h"  #include "pcre.h"
357  #include "ucp.h"  #include "ucp.h"
358    
359    #ifdef COMPILE_PCRE32
360    /* Assert that the public PCRE_UCHAR32 is a 32-bit type */
361    typedef int __assert_pcre_uchar32_size[sizeof(PCRE_UCHAR32) == 4 ? 1 : -1];
362    #endif
363    
364  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
365  need to have their names changed. PCRE must be compiled with the -DVPCOMPAT  need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
366  option on the command line. */  option on the command line. */
# Line 343  The macros are controlled by the value o Line 422  The macros are controlled by the value o
422  the config.h file, but can be overridden by using -D on the command line. This  the config.h file, but can be overridden by using -D on the command line. This
423  is automated on Unix systems via the "configure" command. */  is automated on Unix systems via the "configure" command. */
424    
425    #if defined COMPILE_PCRE8
426    
427  #if LINK_SIZE == 2  #if LINK_SIZE == 2
428    
429  #define PUT(a,n,d)   \  #define PUT(a,n,d)   \
# Line 379  is automated on Unix systems via the "co Line 460  is automated on Unix systems via the "co
460  #define GET(a,n) \  #define GET(a,n) \
461    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
462    
463  #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */  /* Keep it positive */
464    #define MAX_PATTERN_SIZE (1 << 30)
465    
466    #else
467    #error LINK_SIZE must be either 2, 3, or 4
468    #endif
469    
470    #elif defined COMPILE_PCRE16
471    
472    #if LINK_SIZE == 2
473    
474    /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
475    #undef LINK_SIZE
476    #define LINK_SIZE 1
477    
478    #define PUT(a,n,d)   \
479      (a[n] = (d))
480    
481    #define GET(a,n) \
482      (a[n])
483    
484    #define MAX_PATTERN_SIZE (1 << 16)
485    
486    #elif LINK_SIZE == 3 || LINK_SIZE == 4
487    
488    /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
489    #undef LINK_SIZE
490    #define LINK_SIZE 2
491    
492    #define PUT(a,n,d)   \
493      (a[n] = (d) >> 16), \
494      (a[(n)+1] = (d) & 65535)
495    
496    #define GET(a,n) \
497      (((a)[n] << 16) | (a)[(n)+1])
498    
499    /* Keep it positive */
500    #define MAX_PATTERN_SIZE (1 << 30)
501    
502  #else  #else
503  #error LINK_SIZE must be either 2, 3, or 4  #error LINK_SIZE must be either 2, 3, or 4
504  #endif  #endif
505    
506    #elif defined COMPILE_PCRE32
507    
508    /* Only supported LINK_SIZE is 4 */
509    /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
510    #undef LINK_SIZE
511    #define LINK_SIZE 1
512    
513    #define PUT(a,n,d)   \
514      (a[n] = (d))
515    
516    #define GET(a,n) \
517      (a[n])
518    
519    /* Keep it positive */
520    #define MAX_PATTERN_SIZE (1 << 30)
521    
522    #else
523    #error Unsupported compiling mode
524    #endif /* COMPILE_PCRE[8|16|32] */
525    
526  /* Convenience macro defined in terms of the others */  /* Convenience macro defined in terms of the others */
527    
# Line 396  is automated on Unix systems via the "co Line 532  is automated on Unix systems via the "co
532  offsets changes. There are used for repeat counts and for other things such as  offsets changes. There are used for repeat counts and for other things such as
533  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
534    
535    #if defined COMPILE_PCRE8
536    
537    #define IMM2_SIZE 2
538    
539  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
540    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
541    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
542    
543    /* For reasons that I do not understand, the expression in this GET2 macro is
544    treated by gcc as a signed expression, even when a is declared as unsigned. It
545    seems that any kind of arithmetic results in a signed value. */
546    
547  #define GET2(a,n) \  #define GET2(a,n) \
548    (((a)[n] << 8) | (a)[(n)+1])    (unsigned int)(((a)[n] << 8) | (a)[(n)+1])
549    
550  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  #elif defined COMPILE_PCRE16
551    
552    #define IMM2_SIZE 1
553    
554  /* When UTF-8 encoding is being used, a character is no longer just a single  #define PUT2(a,n,d)   \
555  byte. The macros for character handling generate simple sequences when used in     a[n] = d
556  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is  
557  not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should  #define GET2(a,n) \
558  never be called in byte mode. To make sure they can never even appear when     a[n]
559  UTF-8 support is omitted, we don't even define them. */  
560    #elif defined COMPILE_PCRE32
561    
562    #define IMM2_SIZE 1
563    
564  #ifndef SUPPORT_UTF8  #define PUT2(a,n,d)   \
565       a[n] = d
566    
567    #define GET2(a,n) \
568       a[n]
569    
570    #else
571    #error Unsupported compiling mode
572    #endif /* COMPILE_PCRE[8|16|32] */
573    
574    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
575    
576    /* The maximum length of a MARK name is currently one data unit; it may be
577    changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */
578    
579    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
580    #define MAX_MARK ((1u << 16) - 1)
581    #else
582    #define MAX_MARK ((1u << 8) - 1)
583    #endif
584    
585    /* When UTF encoding is being used, a character is no longer just a single
586    byte. The macros for character handling generate simple sequences when used in
587    character-mode, and more complicated ones for UTF characters. GETCHARLENTEST
588    and other macros are not used when UTF is not supported, so they are not
589    defined. To make sure they can never even appear when UTF support is omitted,
590    we don't even define them. */
591    
592    #ifndef SUPPORT_UTF
593    
594    /* #define MAX_VALUE_FOR_SINGLE_CHAR */
595    /* #define HAS_EXTRALEN(c) */
596    /* #define GET_EXTRALEN(c) */
597    /* #define NOT_FIRSTCHAR(c) */
598  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
599  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
600  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
601  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
602  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
603    #define RAWUCHAR(eptr) (*(eptr))
604    #define RAWUCHARINC(eptr) (*(eptr)++)
605    #define RAWUCHARTEST(eptr) (*(eptr))
606    #define RAWUCHARINCTEST(eptr) (*(eptr)++)
607  /* #define GETCHARLENTEST(c, eptr, len) */  /* #define GETCHARLENTEST(c, eptr, len) */
608  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
609    /* #define FORWARDCHAR(eptr) */
610    /* #define ACROSSCHAR(condition, eptr, action) */
611    
612  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF */
613    
614  /* These macros were originally written in the form of loops that used data  /* Tests whether the code point needs extra characters to decode. */
615  from the tables whose names start with _pcre_utf8_table. They were rewritten by  
616  a user so as not to use loops, because in some environments this gives a  #define HASUTF8EXTRALEN(c) ((c) >= 0xc0)
 significant performance advantage, and it seems never to do any harm. */  
617    
618  /* Base macro to pick up the remaining bytes of a UTF-8 character, not  /* Base macro to pick up the remaining bytes of a UTF-8 character, not
619  advancing the pointer. */  advancing the pointer. */
# Line 451  advancing the pointer. */ Line 637  advancing the pointer. */
637            ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \            ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
638      }      }
639    
 /* Get the next UTF-8 character, not advancing the pointer. This is called when  
 we know we are in UTF-8 mode. */  
   
 #define GETCHAR(c, eptr) \  
   c = *eptr; \  
   if (c >= 0xc0) GETUTF8(c, eptr);  
   
 /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the  
 pointer. */  
   
 #define GETCHARTEST(c, eptr) \  
   c = *eptr; \  
   if (utf8 && c >= 0xc0) GETUTF8(c, eptr);  
   
640  /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing  /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
641  the pointer. */  the pointer. */
642    
# Line 499  the pointer. */ Line 671  the pointer. */
671        } \        } \
672      }      }
673    
674    #if defined COMPILE_PCRE8
675    
676    /* These macros were originally written in the form of loops that used data
677    from the tables whose names start with PRIV(utf8_table). They were rewritten by
678    a user so as not to use loops, because in some environments this gives a
679    significant performance advantage, and it seems never to do any harm. */
680    
681    /* Tells the biggest code point which can be encoded as a single character. */
682    
683    #define MAX_VALUE_FOR_SINGLE_CHAR 127
684    
685    /* Tests whether the code point needs extra characters to decode. */
686    
687    #define HAS_EXTRALEN(c) ((c) >= 0xc0)
688    
689    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
690    Otherwise it has an undefined behaviour. */
691    
692    #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
693    
694    /* Returns TRUE, if the given character is not the first character
695    of a UTF sequence. */
696    
697    #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
698    
699    /* Get the next UTF-8 character, not advancing the pointer. This is called when
700    we know we are in UTF-8 mode. */
701    
702    #define GETCHAR(c, eptr) \
703      c = *eptr; \
704      if (c >= 0xc0) GETUTF8(c, eptr);
705    
706    /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
707    pointer. */
708    
709    #define GETCHARTEST(c, eptr) \
710      c = *eptr; \
711      if (utf && c >= 0xc0) GETUTF8(c, eptr);
712    
713  /* Get the next UTF-8 character, advancing the pointer. This is called when we  /* Get the next UTF-8 character, advancing the pointer. This is called when we
714  know we are in UTF-8 mode. */  know we are in UTF-8 mode. */
715    
# Line 511  This is called when we don't know if we Line 722  This is called when we don't know if we
722    
723  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
724    c = *eptr++; \    c = *eptr++; \
725    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);    if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
726    
727  /* Base macro to pick up the remaining bytes of a UTF-8 character, not  /* Base macro to pick up the remaining bytes of a UTF-8 character, not
728  advancing the pointer, incrementing the length. */  advancing the pointer, incrementing the length. */
# Line 563  do not know if we are in UTF-8 mode. */ Line 774  do not know if we are in UTF-8 mode. */
774    
775  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
776    c = *eptr; \    c = *eptr; \
777    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
778    
779    /* Returns the next uchar, not advancing the pointer. This is called when
780    we know we are in UTF mode. */
781    
782    #define RAWUCHAR(eptr) \
783      (*(eptr))
784    
785    /* Returns the next uchar, advancing the pointer. This is called when
786    we know we are in UTF mode. */
787    
788    #define RAWUCHARINC(eptr) \
789      (*((eptr)++))
790    
791    /* Returns the next uchar, testing for UTF mode, and not advancing the
792    pointer. */
793    
794    #define RAWUCHARTEST(eptr) \
795      (*(eptr))
796    
797    /* Returns the next uchar, testing for UTF mode, advancing the
798    pointer. */
799    
800    #define RAWUCHARINCTEST(eptr) \
801      (*((eptr)++))
802    
803  /* If the pointer is not at the start of a character, move it back until  /* If the pointer is not at the start of a character, move it back until
804  it is. This is called only in UTF-8 mode - we don't put a test within the macro  it is. This is called only in UTF-8 mode - we don't put a test within the macro
# Line 571  because almost all calls are already wit Line 806  because almost all calls are already wit
806    
807  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
808    
809  #endif  /* SUPPORT_UTF8 */  /* Same as above, just in the other direction. */
810    #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
811    
812    /* Same as above, but it allows a fully customizable form. */
813    #define ACROSSCHAR(condition, eptr, action) \
814      while((condition) && ((eptr) & 0xc0) == 0x80) action
815    
816    #elif defined COMPILE_PCRE16
817    
818    /* Tells the biggest code point which can be encoded as a single character. */
819    
820    #define MAX_VALUE_FOR_SINGLE_CHAR 65535
821    
822    /* Tests whether the code point needs extra characters to decode. */
823    
824    #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
825    
826    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
827    Otherwise it has an undefined behaviour. */
828    
829    #define GET_EXTRALEN(c) 1
830    
831    /* Returns TRUE, if the given character is not the first character
832    of a UTF sequence. */
833    
834    #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
835    
836    /* Base macro to pick up the low surrogate of a UTF-16 character, not
837    advancing the pointer. */
838    
839    #define GETUTF16(c, eptr) \
840       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
841    
842    /* Get the next UTF-16 character, not advancing the pointer. This is called when
843    we know we are in UTF-16 mode. */
844    
845    #define GETCHAR(c, eptr) \
846      c = *eptr; \
847      if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
848    
849    /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
850    pointer. */
851    
852    #define GETCHARTEST(c, eptr) \
853      c = *eptr; \
854      if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
855    
856    /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
857    the pointer. */
858    
859    #define GETUTF16INC(c, eptr) \
860       { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
861    
862  /* In case there is no definition of offsetof() provided - though any proper  /* Get the next UTF-16 character, advancing the pointer. This is called when we
863  Standard C system should have one. */  know we are in UTF-16 mode. */
864    
865  #ifndef offsetof  #define GETCHARINC(c, eptr) \
866  #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))    c = *eptr++; \
867      if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
868    
869    /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
870    This is called when we don't know if we are in UTF-16 mode. */
871    
872    #define GETCHARINCTEST(c, eptr) \
873      c = *eptr++; \
874      if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
875    
876    /* Base macro to pick up the low surrogate of a UTF-16 character, not
877    advancing the pointer, incrementing the length. */
878    
879    #define GETUTF16LEN(c, eptr, len) \
880       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
881    
882    /* Get the next UTF-16 character, not advancing the pointer, incrementing
883    length if there is a low surrogate. This is called when we know we are in
884    UTF-16 mode. */
885    
886    #define GETCHARLEN(c, eptr, len) \
887      c = *eptr; \
888      if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
889    
890    /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
891    pointer, incrementing length if there is a low surrogate. This is called when
892    we do not know if we are in UTF-16 mode. */
893    
894    #define GETCHARLENTEST(c, eptr, len) \
895      c = *eptr; \
896      if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
897    
898    /* Returns the next uchar, not advancing the pointer. This is called when
899    we know we are in UTF mode. */
900    
901    #define RAWUCHAR(eptr) \
902      (*(eptr))
903    
904    /* Returns the next uchar, advancing the pointer. This is called when
905    we know we are in UTF mode. */
906    
907    #define RAWUCHARINC(eptr) \
908      (*((eptr)++))
909    
910    /* Returns the next uchar, testing for UTF mode, and not advancing the
911    pointer. */
912    
913    #define RAWUCHARTEST(eptr) \
914      (*(eptr))
915    
916    /* Returns the next uchar, testing for UTF mode, advancing the
917    pointer. */
918    
919    #define RAWUCHARINCTEST(eptr) \
920      (*((eptr)++))
921    
922    /* If the pointer is not at the start of a character, move it back until
923    it is. This is called only in UTF-16 mode - we don't put a test within the
924    macro because almost all calls are already within a block of UTF-16 only
925    code. */
926    
927    #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
928    
929    /* Same as above, just in the other direction. */
930    #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
931    
932    /* Same as above, but it allows a fully customizable form. */
933    #define ACROSSCHAR(condition, eptr, action) \
934      if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
935    
936    #elif defined COMPILE_PCRE32
937    
938    /* These are trivial for the 32-bit library, since all UTF-32 characters fit
939    into one pcre_uchar unit. */
940    #define MAX_VALUE_FOR_SINGLE_CHAR (0x10ffffu)
941    #define HAS_EXTRALEN(c) (0)
942    #define GET_EXTRALEN(c) (0)
943    #define NOT_FIRSTCHAR(c) (0)
944    
945    /* Get the next UTF-32 character, not advancing the pointer. This is called when
946    we know we are in UTF-32 mode. */
947    
948    #define GETCHAR(c, eptr) \
949      c = *(eptr);
950    
951    /* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the
952    pointer. */
953    
954    #define GETCHARTEST(c, eptr) \
955      c = *(eptr);
956    
957    /* Get the next UTF-32 character, advancing the pointer. This is called when we
958    know we are in UTF-32 mode. */
959    
960    #define GETCHARINC(c, eptr) \
961      c = *((eptr)++);
962    
963    /* Get the next character, testing for UTF-32 mode, and advancing the pointer.
964    This is called when we don't know if we are in UTF-32 mode. */
965    
966    #define GETCHARINCTEST(c, eptr) \
967      c = *((eptr)++);
968    
969    /* Get the next UTF-32 character, not advancing the pointer, not incrementing
970    length (since all UTF-32 is of length 1). This is called when we know we are in
971    UTF-32 mode. */
972    
973    #define GETCHARLEN(c, eptr, len) \
974      GETCHAR(c, eptr)
975    
976    /* Get the next UTF-32character, testing for UTF-32 mode, not advancing the
977    pointer, not incrementing the length (since all UTF-32 is of length 1).
978    This is called when we do not know if we are in UTF-32 mode. */
979    
980    #define GETCHARLENTEST(c, eptr, len) \
981      GETCHARTEST(c, eptr)
982    
983    /* Returns the next uchar, not advancing the pointer. This is called when
984    we know we are in UTF mode. */
985    
986    #define RAWUCHAR(eptr) \
987      (*(eptr))
988    
989    /* Returns the next uchar, advancing the pointer. This is called when
990    we know we are in UTF mode. */
991    
992    #define RAWUCHARINC(eptr) \
993      (*((eptr)++))
994    
995    /* Returns the next uchar, testing for UTF mode, and not advancing the
996    pointer. */
997    
998    #define RAWUCHARTEST(eptr) \
999      (*(eptr))
1000    
1001    /* Returns the next uchar, testing for UTF mode, advancing the
1002    pointer. */
1003    
1004    #define RAWUCHARINCTEST(eptr) \
1005      (*((eptr)++))
1006    
1007    /* If the pointer is not at the start of a character, move it back until
1008    it is. This is called only in UTF-32 mode - we don't put a test within the
1009    macro because almost all calls are already within a block of UTF-32 only
1010    code.
1011    These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */
1012    
1013    #define BACKCHAR(eptr) do { } while (0)
1014    
1015    /* Same as above, just in the other direction. */
1016    #define FORWARDCHAR(eptr) do { } while (0)
1017    
1018    /* Same as above, but it allows a fully customizable form. */
1019    #define ACROSSCHAR(condition, eptr, action) do { } while (0)
1020    
1021    #else
1022    #error Unsupported compiling mode
1023    #endif /* COMPILE_PCRE[8|16|32] */
1024    
1025    #endif  /* SUPPORT_UTF */
1026    
1027    /* Tests for Unicode horizontal and vertical whitespace characters must check a
1028    number of different values. Using a switch statement for this generates the
1029    fastest code (no loop, no memory access), and there are several places in the
1030    interpreter code where this happens. In order to ensure that all the case lists
1031    remain in step, we use macros so that there is only one place where the lists
1032    are defined.
1033    
1034    These values are also required as lists in pcre_compile.c when processing \h,
1035    \H, \v and \V in a character class. The lists are defined in pcre_tables.c, but
1036    macros that define the values are here so that all the definitions are
1037    together. The lists must be in ascending character order, terminated by
1038    NOTACHAR (which is 0xffffffff).
1039    
1040    Any changes should ensure that the various macros are kept in step with each
1041    other. NOTE: The values also appear in pcre_jit_compile.c. */
1042    
1043    /* ------ ASCII/Unicode environments ------ */
1044    
1045    #ifndef EBCDIC
1046    
1047    #define HSPACE_LIST \
1048      CHAR_HT, CHAR_SPACE, 0xa0, \
1049      0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \
1050      0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \
1051      NOTACHAR
1052    
1053    #define HSPACE_MULTIBYTE_CASES \
1054      case 0x1680:  /* OGHAM SPACE MARK */ \
1055      case 0x180e:  /* MONGOLIAN VOWEL SEPARATOR */ \
1056      case 0x2000:  /* EN QUAD */ \
1057      case 0x2001:  /* EM QUAD */ \
1058      case 0x2002:  /* EN SPACE */ \
1059      case 0x2003:  /* EM SPACE */ \
1060      case 0x2004:  /* THREE-PER-EM SPACE */ \
1061      case 0x2005:  /* FOUR-PER-EM SPACE */ \
1062      case 0x2006:  /* SIX-PER-EM SPACE */ \
1063      case 0x2007:  /* FIGURE SPACE */ \
1064      case 0x2008:  /* PUNCTUATION SPACE */ \
1065      case 0x2009:  /* THIN SPACE */ \
1066      case 0x200A:  /* HAIR SPACE */ \
1067      case 0x202f:  /* NARROW NO-BREAK SPACE */ \
1068      case 0x205f:  /* MEDIUM MATHEMATICAL SPACE */ \
1069      case 0x3000   /* IDEOGRAPHIC SPACE */
1070    
1071    #define HSPACE_BYTE_CASES \
1072      case CHAR_HT: \
1073      case CHAR_SPACE: \
1074      case 0xa0     /* NBSP */
1075    
1076    #define HSPACE_CASES \
1077      HSPACE_BYTE_CASES: \
1078      HSPACE_MULTIBYTE_CASES
1079    
1080    #define VSPACE_LIST \
1081      CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR
1082    
1083    #define VSPACE_MULTIBYTE_CASES \
1084      case 0x2028:    /* LINE SEPARATOR */ \
1085      case 0x2029     /* PARAGRAPH SEPARATOR */
1086    
1087    #define VSPACE_BYTE_CASES \
1088      case CHAR_LF: \
1089      case CHAR_VT: \
1090      case CHAR_FF: \
1091      case CHAR_CR: \
1092      case CHAR_NEL
1093    
1094    #define VSPACE_CASES \
1095      VSPACE_BYTE_CASES: \
1096      VSPACE_MULTIBYTE_CASES
1097    
1098    /* ------ EBCDIC environments ------ */
1099    
1100    #else
1101    #define HSPACE_LIST CHAR_HT, CHAR_SPACE
1102    
1103    #define HSPACE_BYTE_CASES \
1104      case CHAR_HT: \
1105      case CHAR_SPACE
1106    
1107    #define HSPACE_CASES HSPACE_BYTE_CASES
1108    
1109    #ifdef EBCDIC_NL25
1110    #define VSPACE_LIST \
1111      CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR
1112    #else
1113    #define VSPACE_LIST \
1114      CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR
1115  #endif  #endif
1116    
1117    #define VSPACE_BYTE_CASES \
1118      case CHAR_LF: \
1119      case CHAR_VT: \
1120      case CHAR_FF: \
1121      case CHAR_CR: \
1122      case CHAR_NEL
1123    
1124    #define VSPACE_CASES VSPACE_BYTE_CASES
1125    #endif  /* EBCDIC */
1126    
1127    /* ------ End of whitespace macros ------ */
1128    
1129    
1130    
1131  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
1132  live at the top end of the options word, but that got almost full, so now they  live at the top end of the options word, but that got almost full, so they were
1133  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as  moved to a 16-bit flags word - which got almost full, so now they are in a
1134  the restrictions on partial matching have been lifted. It remains for backwards  32-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as the
1135    restrictions on partial matching have been lifted. It remains for backwards
1136  compatibility. */  compatibility. */
1137    
1138  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_MODE8         0x00000001  /* compiled in 8 bit mode */
1139  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_MODE16        0x00000002  /* compiled in 16 bit mode */
1140  #define PCRE_REQCHSET      0x0004  /* req_byte is set */  #define PCRE_MODE32        0x00000004  /* compiled in 32 bit mode */
1141  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #define PCRE_FIRSTSET      0x00000010  /* first_char is set */
1142  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_FCH_CASELESS  0x00000020  /* caseless first char */
1143  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #define PCRE_REQCHSET      0x00000040  /* req_byte is set */
1144  #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */  #define PCRE_RCH_CASELESS  0x00000080  /* caseless requested char */
1145    #define PCRE_STARTLINE     0x00000100  /* start after \n for multiline */
1146    #define PCRE_NOPARTIAL     0x00000200  /* can't use partial with this regex */
1147    #define PCRE_JCHANGED      0x00000400  /* j option used in regex */
1148    #define PCRE_HASCRORLF     0x00000800  /* explicit \r or \n in pattern */
1149    #define PCRE_HASTHEN       0x00001000  /* pattern contains (*THEN) */
1150    #define PCRE_MLSET         0x00002000  /* match limit set by regex */
1151    #define PCRE_RLSET         0x00004000  /* recursion limit set by regex */
1152    #define PCRE_MATCH_EMPTY   0x00008000  /* pattern can match empty string */
1153    
1154    #if defined COMPILE_PCRE8
1155    #define PCRE_MODE          PCRE_MODE8
1156    #elif defined COMPILE_PCRE16
1157    #define PCRE_MODE          PCRE_MODE16
1158    #elif defined COMPILE_PCRE32
1159    #define PCRE_MODE          PCRE_MODE32
1160    #endif
1161    #define PCRE_MODE_MASK     (PCRE_MODE8 | PCRE_MODE16 | PCRE_MODE32)
1162    
1163  /* Flags for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
1164    
# Line 612  time, run time, or study time, respectiv Line 1176  time, run time, or study time, respectiv
1176     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
1177     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
1178     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1179     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
1180    
1181  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
1182    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
# Line 626  time, run time, or study time, respectiv Line 1190  time, run time, or study time, respectiv
1190     PCRE_NO_START_OPTIMIZE)     PCRE_NO_START_OPTIMIZE)
1191    
1192  #define PUBLIC_STUDY_OPTIONS \  #define PUBLIC_STUDY_OPTIONS \
1193     PCRE_STUDY_JIT_COMPILE     (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
1194        PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)
1195    
1196    #define PUBLIC_JIT_EXEC_OPTIONS \
1197       (PCRE_NO_UTF8_CHECK|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|\
1198        PCRE_NOTEMPTY_ATSTART|PCRE_PARTIAL_SOFT|PCRE_PARTIAL_HARD)
1199    
1200  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. */
 to detect whether a pattern was compiled on a host of different endianness. */  
1201    
1202  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
1203    
1204  /* Negative values for the firstchar and reqchar variables */  /* This variable is used to detect a loaded regular expression
1205    in different endianness. */
1206    
1207  #define REQ_UNSET (-2)  #define REVERSED_MAGIC_NUMBER  0x45524350UL   /* 'ERCP' */
 #define REQ_NONE  (-1)  
1208    
1209  /* The maximum remaining length of subject we are prepared to search for a  /* The maximum remaining length of subject we are prepared to search for a
1210  req_byte match. */  req_byte match. */
1211    
1212  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
1213    
 /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a  
 variable-length repeat, or a anything other than literal characters. */  
   
 #define REQ_CASELESS 0x0100    /* indicates caselessness */  
 #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  
   
1214  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
1215  environments where these macros are defined elsewhere. Unfortunately, there  environments where these macros are defined elsewhere. Unfortunately, there
1216  is no way to do the same for the typedef. */  is no way to do the same for the typedef. */
# Line 677  for) in a minority area (EBCDIC platform Line 1239  for) in a minority area (EBCDIC platform
1239  application that did need both could compile two versions of the library, using  application that did need both could compile two versions of the library, using
1240  macros to give the functions distinct names. */  macros to give the functions distinct names. */
1241    
1242  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF
1243    
1244  /* UTF-8 support is not enabled; use the platform-dependent character literals  /* UTF-8 support is not enabled; use the platform-dependent character literals
1245  so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */  so that PCRE works in both ASCII and EBCDIC environments, but only in non-UTF
1246    mode. Newline characters are problematic in EBCDIC. Though it has CR and LF
1247    characters, a common practice has been to use its NL (0x15) character as the
1248    line terminator in C-like processing environments. However, sometimes the LF
1249    (0x25) character is used instead, according to this Unicode document:
1250    
1251    http://unicode.org/standard/reports/tr13/tr13-5.html
1252    
1253    PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25
1254    instead. Whichever is *not* chosen is defined as NEL.
1255    
1256    In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the
1257    same code point. */
1258    
1259    #ifdef EBCDIC
1260    
1261    #ifndef EBCDIC_NL25
1262    #define CHAR_NL                     '\x15'
1263    #define CHAR_NEL                    '\x25'
1264    #define STR_NL                      "\x15"
1265    #define STR_NEL                     "\x25"
1266    #else
1267    #define CHAR_NL                     '\x25'
1268    #define CHAR_NEL                    '\x15'
1269    #define STR_NL                      "\x25"
1270    #define STR_NEL                     "\x15"
1271    #endif
1272    
1273    #define CHAR_LF                     CHAR_NL
1274    #define STR_LF                      STR_NL
1275    
1276    #define CHAR_ESC                    '\047'
1277    #define CHAR_DEL                    '\007'
1278    #define STR_ESC                     "\047"
1279    #define STR_DEL                     "\007"
1280    
1281    #else  /* Not EBCDIC */
1282    
1283    /* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for
1284    compatibility. NEL is the Unicode newline character; make sure it is
1285    a positive value. */
1286    
1287    #define CHAR_LF                     '\n'
1288    #define CHAR_NL                     CHAR_LF
1289    #define CHAR_NEL                    ((unsigned char)'\x85')
1290    #define CHAR_ESC                    '\033'
1291    #define CHAR_DEL                    '\177'
1292    
1293    #define STR_LF                      "\n"
1294    #define STR_NL                      STR_LF
1295    #define STR_NEL                     "\x85"
1296    #define STR_ESC                     "\033"
1297    #define STR_DEL                     "\177"
1298    
1299    #endif  /* EBCDIC */
1300    
1301    /* The remaining definitions work in both environments. */
1302    
1303    #define CHAR_NULL                   '\0'
1304  #define CHAR_HT                     '\t'  #define CHAR_HT                     '\t'
1305  #define CHAR_VT                     '\v'  #define CHAR_VT                     '\v'
1306  #define CHAR_FF                     '\f'  #define CHAR_FF                     '\f'
1307  #define CHAR_CR                     '\r'  #define CHAR_CR                     '\r'
 #define CHAR_NL                     '\n'  
1308  #define CHAR_BS                     '\b'  #define CHAR_BS                     '\b'
1309  #define CHAR_BEL                    '\a'  #define CHAR_BEL                    '\a'
 #ifdef EBCDIC  
 #define CHAR_ESC                    '\047'  
 #define CHAR_DEL                    '\007'  
 #else  
 #define CHAR_ESC                    '\033'  
 #define CHAR_DEL                    '\177'  
 #endif  
1310    
1311  #define CHAR_SPACE                  ' '  #define CHAR_SPACE                  ' '
1312  #define CHAR_EXCLAMATION_MARK       '!'  #define CHAR_EXCLAMATION_MARK       '!'
# Line 797  so that PCRE works on both ASCII and EBC Line 1408  so that PCRE works on both ASCII and EBC
1408  #define STR_VT                      "\v"  #define STR_VT                      "\v"
1409  #define STR_FF                      "\f"  #define STR_FF                      "\f"
1410  #define STR_CR                      "\r"  #define STR_CR                      "\r"
 #define STR_NL                      "\n"  
1411  #define STR_BS                      "\b"  #define STR_BS                      "\b"
1412  #define STR_BEL                     "\a"  #define STR_BEL                     "\a"
 #ifdef EBCDIC  
 #define STR_ESC                     "\047"  
 #define STR_DEL                     "\007"  
 #else  
 #define STR_ESC                     "\033"  
 #define STR_DEL                     "\177"  
 #endif  
1413    
1414  #define STR_SPACE                   " "  #define STR_SPACE                   " "
1415  #define STR_EXCLAMATION_MARK        "!"  #define STR_EXCLAMATION_MARK        "!"
# Line 938  so that PCRE works on both ASCII and EBC Line 1541  so that PCRE works on both ASCII and EBC
1541  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1542  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1543  #define STRING_UTF8_RIGHTPAR           "UTF8)"  #define STRING_UTF8_RIGHTPAR           "UTF8)"
1544    #define STRING_UTF16_RIGHTPAR          "UTF16)"
1545    #define STRING_UTF32_RIGHTPAR          "UTF32)"
1546    #define STRING_UTF_RIGHTPAR            "UTF)"
1547  #define STRING_UCP_RIGHTPAR            "UCP)"  #define STRING_UCP_RIGHTPAR            "UCP)"
1548  #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"  #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1549    #define STRING_LIMIT_MATCH_EQ          "LIMIT_MATCH="
1550    #define STRING_LIMIT_RECURSION_EQ      "LIMIT_RECURSION="
1551    
1552  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF */
1553    
1554  /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This  /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1555  works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode  works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
# Line 951  only. */ Line 1559  only. */
1559  #define CHAR_VT                     '\013'  #define CHAR_VT                     '\013'
1560  #define CHAR_FF                     '\014'  #define CHAR_FF                     '\014'
1561  #define CHAR_CR                     '\015'  #define CHAR_CR                     '\015'
1562  #define CHAR_NL                     '\012'  #define CHAR_LF                     '\012'
1563    #define CHAR_NL                     CHAR_LF
1564    #define CHAR_NEL                    ((unsigned char)'\x85')
1565  #define CHAR_BS                     '\010'  #define CHAR_BS                     '\010'
1566  #define CHAR_BEL                    '\007'  #define CHAR_BEL                    '\007'
1567  #define CHAR_ESC                    '\033'  #define CHAR_ESC                    '\033'
1568  #define CHAR_DEL                    '\177'  #define CHAR_DEL                    '\177'
1569    
1570    #define CHAR_NULL                   '\0'
1571  #define CHAR_SPACE                  '\040'  #define CHAR_SPACE                  '\040'
1572  #define CHAR_EXCLAMATION_MARK       '\041'  #define CHAR_EXCLAMATION_MARK       '\041'
1573  #define CHAR_QUOTATION_MARK         '\042'  #define CHAR_QUOTATION_MARK         '\042'
# Line 1193  only. */ Line 1804  only. */
1804  #define STRING_BSR_ANYCRLF_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_BSR_ANYCRLF_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1805  #define STRING_BSR_UNICODE_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS  #define STRING_BSR_UNICODE_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1806  #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1807    #define STRING_UTF16_RIGHTPAR          STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1808    #define STRING_UTF32_RIGHTPAR          STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS
1809    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
1810  #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS  #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1811  #define STRING_NO_START_OPT_RIGHTPAR   STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS  #define STRING_NO_START_OPT_RIGHTPAR   STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS
1812    #define STRING_LIMIT_MATCH_EQ          STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_M STR_A STR_T STR_C STR_H STR_EQUALS_SIGN
1813    #define STRING_LIMIT_RECURSION_EQ      STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_R STR_E STR_C STR_U STR_R STR_S STR_I STR_O STR_N STR_EQUALS_SIGN
1814    
1815  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
1816    
1817  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1818    
# Line 1209  only. */ Line 1825  only. */
1825  #endif  #endif
1826    
1827  #ifndef ESC_n  #ifndef ESC_n
1828  #define ESC_n CHAR_NL  #define ESC_n CHAR_LF
1829  #endif  #endif
1830    
1831  #ifndef ESC_r  #ifndef ESC_r
# Line 1234  only. */ Line 1850  only. */
1850  #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */  #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1851  #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */  #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1852  #define PT_WORD       8    /* Word - L plus N plus underscore */  #define PT_WORD       8    /* Word - L plus N plus underscore */
1853    #define PT_CLIST      9    /* Pseudo-property: match character list */
1854    #define PT_UCNC      10    /* Universal Character nameable character */
1855    
1856  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1857  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1858    
1859  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1860  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
# Line 1249  contain UTF-8 characters with values gre Line 1867  contain UTF-8 characters with values gre
1867    
1868  /* These are escaped items that aren't just an encoding of a particular data  /* These are escaped items that aren't just an encoding of a particular data
1869  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
1870  their negation. Also, they must appear in the same order as in the opcode  0 for a data character.  Also, they must appear in the same order as in the opcode
1871  definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it  definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1872  corresponds to "." in DOTALL mode rather than an escape sequence. It is also  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1873  used for [^] in JavaScript compatibility mode. In non-DOTALL mode, "." behaves  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1874  like \N.  non-DOTALL mode, "." behaves like \N.
1875    
1876  The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.  The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1877  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.  when PCRE_UCP is set and replacement of \d etc by \p sequences is required.
1878  They must be contiguous, and remain in order so that the replacements can be  They must be contiguous, and remain in order so that the replacements can be
1879  looked up from a table.  looked up from a table.
1880    
1881  The final escape must be ESC_REF as subsequent values are used for  Negative numbers are used to encode a backreference (\1, \2, \3, etc.) in
1882  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape  check_escape(). There are two tests in the code for an escape
1883  greater than ESC_b and less than ESC_Z to detect the types that may be  greater than ESC_b and less than ESC_Z to detect the types that may be
1884  repeated. These are the types that consume characters. If any new escapes are  repeated. These are the types that consume characters. If any new escapes are
1885  put in between that don't consume a character, that code will have to change.  put in between that don't consume a character, that code will have to change.
# Line 1271  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E Line 1889  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1889         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1890         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1891         ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_E, ESC_Q, ESC_g, ESC_k,
1892         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
        ESC_REF };  
1893    
1894  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
1895  OP_EOD must correspond in order to the list of escapes immediately above.  OP_EOD must correspond in order to the list of escapes immediately above.
# Line 1298  enum { Line 1915  enum {
1915    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1916    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1917    
1918    OP_ANY,            /* 12 Match any character except newline */    OP_ANY,            /* 12 Match any character except newline (\N) */
1919    OP_ALLANY,         /* 13 Match any character */    OP_ALLANY,         /* 13 Match any character */
1920    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1921    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
# Line 1309  enum { Line 1926  enum {
1926    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1927    OP_VSPACE,         /* 21 \v (vertical whitespace) */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1928    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1929    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */    OP_EODN,           /* 23 End of data or \n at end of data (\Z) */
1930    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data (\z) */
1931    
1932    OP_CIRC,           /* 25 Start of line - not multiline */    OP_CIRC,           /* 25 Start of line - not multiline */
1933    OP_CIRCM,          /* 26 Start of line - multiline */    OP_CIRCM,          /* 26 Start of line - multiline */
# Line 1433  enum { Line 2050  enum {
2050    OP_CLASS,          /* 106 Match a character class, chars < 256 only */    OP_CLASS,          /* 106 Match a character class, chars < 256 only */
2051    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
2052                                class - the difference is relevant only when a                                class - the difference is relevant only when a
2053                                UTF-8 character > 255 is encountered. */                                character > 255 is encountered. */
2054    OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
2055                                class. This does both positive and negative. */                                class. This does both positive and negative. */
2056    OP_REF,            /* 109 Match a back reference, casefully */    OP_REF,            /* 109 Match a back reference, casefully */
2057    OP_REFI,           /* 110 Match a back reference, caselessly */    OP_REFI,           /* 110 Match a back reference, caselessly */
# Line 1591  in UTF-8 mode. The code that uses this t Line 2208  in UTF-8 mode. The code that uses this t
2208    2,                             /* noti                                   */ \    2,                             /* noti                                   */ \
2209    /* Positive single-char repeats                             ** These are */ \    /* Positive single-char repeats                             ** These are */ \
2210    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
2211    4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
2212    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE,                   /* exact                                  */ \
2213      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
2214    2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \    2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
2215    4, 4, 4,                       /* upto I, minupto I, exact I             */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
2216    2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \    2+IMM2_SIZE,                   /* exact I                                */ \
2217      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
2218    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
2219    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
2220    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
2221    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
2222      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
2223    2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \    2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
2224    4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
2225    2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \    2+IMM2_SIZE,                   /* NOT exact I                            */ \
2226      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
2227    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
2228    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
2229    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
2230    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
2231      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
2232    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
2233    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
2234    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
2235   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
2236   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
2237    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
2238    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
2239    3,                             /* REFI                                   */ \    1+IMM2_SIZE,                   /* REFI                                   */ \
2240    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
2241    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
2242    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
# Line 1631  in UTF-8 mode. The code that uses this t Line 2253  in UTF-8 mode. The code that uses this t
2253    1+LINK_SIZE,                   /* ONCE_NC                                */ \    1+LINK_SIZE,                   /* ONCE_NC                                */ \
2254    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
2255    1+LINK_SIZE,                   /* BRAPOS                                 */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
2256    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
2257    3+LINK_SIZE,                   /* CBRAPOS                                */ \    1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
2258    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
2259    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
2260    1+LINK_SIZE,                   /* SBRAPOS                                */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
2261    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
2262    3+LINK_SIZE,                   /* SCBRAPOS                               */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
2263    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
2264    3, 3,                          /* CREF, NCREF                            */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
2265    3, 3,                          /* RREF, NRREF                            */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
2266    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
2267    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
2268    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
2269    1, 3,                          /* SKIP, SKIP_ARG                         */ \    1, 3,                          /* SKIP, SKIP_ARG                         */ \
2270    1, 3,                          /* THEN, THEN_ARG                         */ \    1, 3,                          /* THEN, THEN_ARG                         */ \
2271    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
2272    3, 1                           /* CLOSE, SKIPZERO  */    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
2273    
2274  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
2275  condition. */  condition. */
# Line 1665  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 2287  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
2287         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2288         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2289         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2290         ERRCOUNT };         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERRCOUNT };
2291    
2292    /* JIT compiling modes. The function list is indexed by them. */
2293    enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2294           JIT_NUMBER_OF_COMPILE_MODES };
2295    
2296  /* The real format of the start of the pcre block; the index of names and the  /* The real format of the start of the pcre block; the index of names and the
2297  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
2298  offset to the name table so that if a regex is compiled on one host, saved, and  offset to the name table so that if a regex is compiled on one host, saved, and
2299  then run on another where the size of pointers is different, all might still  then run on another where the size of pointers is different, all might still
2300  be well. For the case of compiled-on-4 and run-on-8, we include an extra  be well.
2301  pointer that is always NULL. For future-proofing, a few dummy fields were  
2302  originally included - even though you can never get this planning right - but  The size of the structure must be a multiple of 8 bytes. For the case of
2303  there is only one left now.  compiled-on-4 and run-on-8, we include an extra pointer that is always NULL so
2304    that there are an even number of pointers which therefore are a multiple of 8
2305  NOTE NOTE NOTE:  bytes.
2306  Because people can now save and re-use compiled patterns, any additions to this  
2307  structure should be made at the end, and something earlier (e.g. a new  It is necessary to fork the struct for the 32 bit library, since it needs to
2308  flag in the options or one of the dummy fields) should indicate that the new  use pcre_uint32 for first_char and req_char. We can't put an ifdef inside the
2309  fields are present. Currently PCRE always sets the dummy fields to zero.  typedef because pcretest needs access to the struct of the 8-, 16- and 32-bit
2310  NOTE NOTE NOTE  variants.
2311    
2312    *** WARNING ***
2313    When new fields are added to these structures, remember to adjust the code in
2314    pcre_byte_order.c that is concerned with swapping the byte order of the fields
2315    when a compiled regex is reloaded on a host with different endianness.
2316    *** WARNING ***
2317    There is also similar byte-flipping code in pcretest.c, which is used for
2318    testing the byte-flipping features. It must also be kept in step.
2319    *** WARNING ***
2320  */  */
2321    
2322  typedef struct real_pcre {  typedef struct real_pcre8_or_16 {
2323    pcre_uint32 magic_number;    pcre_uint32 magic_number;
2324    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2325    pcre_uint32 options;            /* Public options */    pcre_uint32 options;            /* Public options */
2326    pcre_uint16 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
2327    pcre_uint16 dummy1;             /* For future use */    pcre_uint32 limit_match;        /* Limit set from regex */
2328    pcre_uint16 top_bracket;    pcre_uint32 limit_recursion;    /* Limit set from regex */
2329    pcre_uint16 top_backref;    pcre_uint16 first_char;         /* Starting character */
2330    pcre_uint16 first_byte;    pcre_uint16 req_char;           /* This character must be seen */
2331    pcre_uint16 req_byte;    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
2332      pcre_uint16 top_bracket;        /* Highest numbered group */
2333      pcre_uint16 top_backref;        /* Highest numbered back reference */
2334    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
2335    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
2336    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
2337    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
2338      pcre_uint16 dummy1;             /* To ensure size is a multiple of 8 */
2339      pcre_uint16 dummy2;             /* To ensure size is a multiple of 8 */
2340      pcre_uint16 dummy3;             /* To ensure size is a multiple of 8 */
2341      const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
2342      void             *nullpad;      /* NULL padding */
2343    } real_pcre8_or_16;
2344    
2345    const unsigned char *tables;    /* Pointer to tables or NULL for std */  typedef struct real_pcre8_or_16 real_pcre;
2346    const unsigned char *nullpad;   /* NULL padding */  typedef struct real_pcre8_or_16 real_pcre16;
2347  } real_pcre;  
2348    typedef struct real_pcre32 {
2349      pcre_uint32 magic_number;
2350      pcre_uint32 size;               /* Total that was malloced */
2351      pcre_uint32 options;            /* Public options */
2352      pcre_uint32 flags;              /* Private flags */
2353      pcre_uint32 limit_match;        /* Limit set from regex */
2354      pcre_uint32 limit_recursion;    /* Limit set from regex */
2355      pcre_uint32 first_char;         /* Starting character */
2356      pcre_uint32 req_char;           /* This character must be seen */
2357      pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
2358      pcre_uint16 top_bracket;        /* Highest numbered group */
2359      pcre_uint16 top_backref;        /* Highest numbered back reference */
2360      pcre_uint16 name_table_offset;  /* Offset to name table that follows */
2361      pcre_uint16 name_entry_size;    /* Size of any name items */
2362      pcre_uint16 name_count;         /* Number of name items */
2363      pcre_uint16 ref_count;          /* Reference count */
2364      pcre_uint16 dummy;              /* To ensure size is a multiple of 8 */
2365      const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
2366      void             *nullpad;      /* NULL padding */
2367    } real_pcre32;
2368    
2369    #if defined COMPILE_PCRE8
2370    #define REAL_PCRE real_pcre
2371    #elif defined COMPILE_PCRE16
2372    #define REAL_PCRE real_pcre16
2373    #elif defined COMPILE_PCRE32
2374    #define REAL_PCRE real_pcre32
2375    #endif
2376    
2377    /* Assert that the size of REAL_PCRE is divisible by 8 */
2378    typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1];
2379    
2380    /* Needed in pcretest to access some fields in the real_pcre* structures
2381     * directly. They're unified for 8/16/32 bits since the structs only differ
2382     * after these fields; if that ever changes, need to fork those defines into
2383     * 8/16 and 32 bit versions. */
2384    #define REAL_PCRE_MAGIC(re)     (((REAL_PCRE*)re)->magic_number)
2385    #define REAL_PCRE_SIZE(re)      (((REAL_PCRE*)re)->size)
2386    #define REAL_PCRE_OPTIONS(re)   (((REAL_PCRE*)re)->options)
2387    #define REAL_PCRE_FLAGS(re)     (((REAL_PCRE*)re)->flags)
2388    
2389  /* The format of the block used to store data from pcre_study(). The same  /* The format of the block used to store data from pcre_study(). The same
2390  remark (see NOTE above) about extending this structure applies. */  remark (see NOTE above) about extending this structure applies. */
# Line 1709  remark (see NOTE above) about extending Line 2392  remark (see NOTE above) about extending
2392  typedef struct pcre_study_data {  typedef struct pcre_study_data {
2393    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2394    pcre_uint32 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
2395    uschar start_bits[32];          /* Starting char bits */    pcre_uint8 start_bits[32];      /* Starting char bits */
2396    pcre_uint32 minlength;          /* Minimum subject length */    pcre_uint32 minlength;          /* Minimum subject length */
2397  } pcre_study_data;  } pcre_study_data;
2398    
# Line 1724  typedef struct open_capitem { Line 2407  typedef struct open_capitem {
2407    pcre_uint16 flag;             /* Set TRUE if recursive back ref */    pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2408  } open_capitem;  } open_capitem;
2409    
2410    /* Structure for building a list of named groups during the first pass of
2411    compiling. */
2412    
2413    typedef struct named_group {
2414      const pcre_uchar  *name;          /* Points to the name in the pattern */
2415      int                length;        /* Length of the name */
2416      pcre_uint32        number;        /* Group number */
2417    } named_group;
2418    
2419  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2420  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2421    
2422  typedef struct compile_data {  typedef struct compile_data {
2423    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
2424    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
2425    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
2426    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
2427    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
2428    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
2429    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2430    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
2431    open_capitem *open_caps;      /* Chain of open capture items */    pcre_uchar *hwm;                  /* High watermark of workspace */
2432    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
2433    uschar *name_table;           /* The name/number table */    named_group *named_groups;        /* Points to vector in pre-compile */
2434    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
2435    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
2436    int  bracount;                /* Count of capturing parens as we compile */    int  name_entry_size;             /* Size of each entry */
2437    int  final_bracount;          /* Saved value after first pass */    int  named_group_list_size;       /* Number of entries in the list */
2438    int  top_backref;             /* Maximum back reference */    int  workspace_size;              /* Size of workspace */
2439    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int bracount;            /* Count of capturing parens as we compile */
2440    int  assert_depth;            /* Depth of nested assertions */    int  final_bracount;              /* Saved value after first pass */
2441    int  external_options;        /* External (initial) options */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2442    int  external_flags;          /* External flag bits to be set */    int  top_backref;                 /* Maximum back reference */
2443    int  req_varyopt;             /* "After variable item" flag for reqbyte */    unsigned int backref_map;         /* Bitmap of low back refs */
2444    BOOL had_accept;              /* (*ACCEPT) encountered */    int  assert_depth;                /* Depth of nested assertions */
2445    BOOL check_lookbehind;        /* Lookbehinds need later checking */    pcre_uint32 external_options;     /* External (initial) options */
2446    int  nltype;                  /* Newline type */    pcre_uint32 external_flags;       /* External flag bits to be set */
2447    int  nllen;                   /* Newline string length */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
2448    uschar nl[4];                 /* Newline string when fixed length */    BOOL had_accept;                  /* (*ACCEPT) encountered */
2449      BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
2450      BOOL check_lookbehind;            /* Lookbehinds need later checking */
2451      int  nltype;                      /* Newline type */
2452      int  nllen;                       /* Newline string length */
2453      pcre_uchar nl[4];                 /* Newline string when fixed length */
2454  } compile_data;  } compile_data;
2455    
2456  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 1761  branches, for testing for left recursion Line 2458  branches, for testing for left recursion
2458    
2459  typedef struct branch_chain {  typedef struct branch_chain {
2460    struct branch_chain *outer;    struct branch_chain *outer;
2461    uschar *current_branch;    pcre_uchar *current_branch;
2462  } branch_chain;  } branch_chain;
2463    
2464  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
# Line 1769  call within the pattern; used by pcre_ex Line 2466  call within the pattern; used by pcre_ex
2466    
2467  typedef struct recursion_info {  typedef struct recursion_info {
2468    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2469    int group_num;                  /* Number of group that was called */    unsigned int group_num;         /* Number of group that was called */
2470    int *offset_save;               /* Pointer to start of saved offsets */    int *offset_save;               /* Pointer to start of saved offsets */
2471    int saved_max;                  /* Number of saved offsets */    int saved_max;                  /* Number of saved offsets */
2472    USPTR subject_position;         /* Position at start of recursion */    int saved_capture_last;         /* Last capture number */
2473      PCRE_PUCHAR subject_position;   /* Position at start of recursion */
2474  } recursion_info;  } recursion_info;
2475    
2476  /* A similar structure for pcre_dfa_exec(). */  /* A similar structure for pcre_dfa_exec(). */
# Line 1780  typedef struct recursion_info { Line 2478  typedef struct recursion_info {
2478  typedef struct dfa_recursion_info {  typedef struct dfa_recursion_info {
2479    struct dfa_recursion_info *prevrec;    struct dfa_recursion_info *prevrec;
2480    int group_num;    int group_num;
2481    USPTR subject_position;    PCRE_PUCHAR subject_position;
2482  } dfa_recursion_info;  } dfa_recursion_info;
2483    
2484  /* Structure for building a chain of data for holding the values of the subject  /* Structure for building a chain of data for holding the values of the subject
# Line 1790  pcre_exec(). */ Line 2488  pcre_exec(). */
2488    
2489  typedef struct eptrblock {  typedef struct eptrblock {
2490    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
2491    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
2492  } eptrblock;  } eptrblock;
2493    
2494    
# Line 1801  typedef struct match_data { Line 2499  typedef struct match_data {
2499    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
2500    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
2501    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
2502    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
2503    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
2504    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
2505    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
2506    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
2507    int    name_count;            /* Number of names in name table */    int    name_count;              /* Number of names in name table */
2508    int    name_entry_size;       /* Size of entry in names table */    int    name_entry_size;         /* Size of entry in names table */
2509    uschar *name_table;           /* Table of names */    unsigned int skip_arg_count;    /* For counting SKIP_ARGs */
2510    uschar nl[4];                 /* Newline string when fixed */    unsigned int ignore_skip_arg;   /* For re-run when SKIP arg name not found */
2511    const  uschar *lcc;           /* Points to lower casing table */    pcre_uchar *name_table;         /* Table of names */
2512    const  uschar *ctypes;        /* Points to table of type maps */    pcre_uchar nl[4];               /* Newline string when fixed */
2513    BOOL   offset_overflow;       /* Set if too many extractions */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2514    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2515    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2516    BOOL   utf8;                  /* UTF8 flag */    BOOL   notbol;                  /* NOTBOL flag */
2517    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   noteol;                  /* NOTEOL flag */
2518    BOOL   use_ucp;               /* PCRE_UCP flag */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2519    BOOL   endonly;               /* Dollar not before final \n */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2520    BOOL   notempty;              /* Empty string match not wanted */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2521    BOOL   notempty_atstart;      /* Empty string match at start not wanted */    BOOL   endonly;                 /* Dollar not before final \n */
2522    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   notempty;                /* Empty string match not wanted */
2523    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2524    BOOL   hasthen;               /* Pattern contains (*THEN) */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2525    const  uschar *start_code;    /* For use when recursing */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2526    USPTR  start_subject;         /* Start of the subject string */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
2527    USPTR  end_subject;           /* End of the subject string */    const  pcre_uchar *start_code;  /* For use when recursing */
2528    USPTR  start_match_ptr;       /* Start of matched string */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2529    USPTR  end_match_ptr;         /* Subject position at end match */    PCRE_PUCHAR end_subject;        /* End of the subject string */
2530    USPTR  start_used_ptr;        /* Earliest consulted character */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2531    int    partial;               /* PARTIAL options */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2532    int    end_offset_top;        /* Highwater mark at end of match */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2533    int    capture_last;          /* Most recent capture number */    int    partial;                 /* PARTIAL options */
2534    int    start_offset;          /* The start offset value */    int    end_offset_top;          /* Highwater mark at end of match */
2535    int    match_function_type;   /* Set for certain special calls of MATCH() */    pcre_int32 capture_last;        /* Most recent capture number + overflow flag */
2536    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    int    start_offset;            /* The start offset value */
2537    int    eptrn;                 /* Next free eptrblock */    int    match_function_type;     /* Set for certain special calls of MATCH() */
2538    recursion_info *recursive;    /* Linked list of recursion data */    eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2539    void  *callout_data;          /* To pass back to callouts */    int    eptrn;                   /* Next free eptrblock */
2540    const  uschar *mark;          /* Mark pointer to pass back */    recursion_info *recursive;      /* Linked list of recursion data */
2541    const  uschar *once_target;   /* Where to back up to for atomic groups */    void  *callout_data;            /* To pass back to callouts */
2542      const  pcre_uchar *mark;        /* Mark pointer to pass back on success */
2543      const  pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2544      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2545    #ifdef NO_RECURSE
2546      void  *match_frames_base;       /* For remembering malloc'd frames */
2547    #endif
2548  } match_data;  } match_data;
2549    
2550  /* A similar structure is used for the same purpose by the DFA matching  /* A similar structure is used for the same purpose by the DFA matching
2551  functions. */  functions. */
2552    
2553  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2554    const uschar *start_code;      /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2555    const uschar *start_subject;   /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2556    const uschar *end_subject;     /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2557    const uschar *start_used_ptr;  /* Earliest consulted character */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2558    const uschar *tables;          /* Character tables */    const pcre_uint8 *tables;         /* Character tables */
2559    int   start_offset;            /* The start offset value */    int   start_offset;               /* The start offset value */
2560    int   moptions;                /* Match options */    int   moptions;                   /* Match options */
2561    int   poptions;                /* Pattern options */    int   poptions;                   /* Pattern options */
2562    int    nltype;                 /* Newline type */    int   nltype;                     /* Newline type */
2563    int    nllen;                  /* Newline string length */    int   nllen;                      /* Newline string length */
2564    uschar nl[4];                  /* Newline string when fixed */    pcre_uchar nl[4];                 /* Newline string when fixed */
2565    void  *callout_data;           /* To pass back to callouts */    void *callout_data;               /* To pass back to callouts */
2566    dfa_recursion_info *recursive; /* Linked list of recursion data */    dfa_recursion_info *recursive;    /* Linked list of recursion data */
2567  } dfa_match_data;  } dfa_match_data;
2568    
2569  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1895  total length. */ Line 2599  total length. */
2599  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2600  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2601    
2602    /* Internal function and data prefixes. */
2603    
2604    #if defined COMPILE_PCRE8
2605    #ifndef PUBL
2606    #define PUBL(name) pcre_##name
2607    #endif
2608    #ifndef PRIV
2609    #define PRIV(name) _pcre_##name
2610    #endif
2611    #elif defined COMPILE_PCRE16
2612    #ifndef PUBL
2613    #define PUBL(name) pcre16_##name
2614    #endif
2615    #ifndef PRIV
2616    #define PRIV(name) _pcre16_##name
2617    #endif
2618    #elif defined COMPILE_PCRE32
2619    #ifndef PUBL
2620    #define PUBL(name) pcre32_##name
2621    #endif
2622    #ifndef PRIV
2623    #define PRIV(name) _pcre32_##name
2624    #endif
2625    #else
2626    #error Unsupported compiling mode
2627    #endif /* COMPILE_PCRE[8|16|32] */
2628    
2629  /* Layout of the UCP type table that translates property names into types and  /* Layout of the UCP type table that translates property names into types and
2630  codes. Each entry used to point directly to a name, but to reduce the number of  codes. Each entry used to point directly to a name, but to reduce the number of
2631  relocations in shared libraries, it now has an offset into a single string  relocations in shared libraries, it now has an offset into a single string
# Line 1912  of the exported public functions. They h Line 2643  of the exported public functions. They h
2643  but are not part of the PCRE public API. The data for these tables is in the  but are not part of the PCRE public API. The data for these tables is in the
2644  pcre_tables.c module. */  pcre_tables.c module. */
2645    
2646  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
2647  extern const int    _pcre_utf8_table2[];  extern const int            PRIV(utf8_table1)[];
2648  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table1_size);
2649  extern const uschar _pcre_utf8_table4[];  extern const int            PRIV(utf8_table2)[];
2650    extern const int            PRIV(utf8_table3)[];
2651  #ifdef SUPPORT_JIT  extern const pcre_uint8     PRIV(utf8_table4)[];
2652  extern const uschar _pcre_utf8_char_sizes[];  #endif /* COMPILE_PCRE8 */
2653  #endif  
2654    extern const char           PRIV(utt_names)[];
2655    extern const ucp_type_table PRIV(utt)[];
2656    extern const int            PRIV(utt_size);
2657    
2658  extern const int    _pcre_utf8_table1_size;  extern const pcre_uint8     PRIV(OP_lengths)[];
2659    extern const pcre_uint8     PRIV(default_tables)[];
2660    
2661  extern const char   _pcre_utt_names[];  extern const pcre_uint32    PRIV(hspace_list)[];
2662  extern const ucp_type_table _pcre_utt[];  extern const pcre_uint32    PRIV(vspace_list)[];
 extern const int _pcre_utt_size;  
   
 extern const uschar _pcre_default_tables[];  
   
 extern const uschar _pcre_OP_lengths[];  
2663    
2664    
2665  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2666  one of the exported public functions. They have to be "external" in the C  one of the exported public functions. They have to be "external" in the C
2667  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2668    
2669  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);  /* String comparison functions. */
2670  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);  #if defined COMPILE_PCRE8
2671  extern int           _pcre_ord2utf8(int, uschar *);  
2672  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
2673                         const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
2674  extern int           _pcre_valid_utf8(USPTR, int, int *);  #define STRCMP_UC_C8(str1, str2) \
2675  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);    strcmp((char *)(str1), (str2))
2676  extern BOOL          _pcre_xclass(int, const uschar *);  #define STRNCMP_UC_UC(str1, str2, num) \
2677      strncmp((char *)(str1), (char *)(str2), (num))
2678    #define STRNCMP_UC_C8(str1, str2, num) \
2679      strncmp((char *)(str1), (str2), (num))
2680    #define STRLEN_UC(str) strlen((const char *)str)
2681    
2682    #elif defined COMPILE_PCRE16 || defined COMPILE_PCRE32
2683    
2684    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2685                               const pcre_uchar *);
2686    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2687                               const char *);
2688    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2689                               const pcre_uchar *, unsigned int num);
2690    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2691                               const char *, unsigned int num);
2692    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2693    
2694    #define STRCMP_UC_UC(str1, str2) \
2695      PRIV(strcmp_uc_uc)((str1), (str2))
2696    #define STRCMP_UC_C8(str1, str2) \
2697      PRIV(strcmp_uc_c8)((str1), (str2))
2698    #define STRNCMP_UC_UC(str1, str2, num) \
2699      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2700    #define STRNCMP_UC_C8(str1, str2, num) \
2701      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2702    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2703    
2704    #endif /* COMPILE_PCRE[8|16|32] */
2705    
2706    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
2707    
2708    #define STRCMP_UC_UC_TEST(str1, str2) STRCMP_UC_UC(str1, str2)
2709    #define STRCMP_UC_C8_TEST(str1, str2) STRCMP_UC_C8(str1, str2)
2710    
2711    #elif defined COMPILE_PCRE32
2712    
2713    extern int               PRIV(strcmp_uc_uc_utf)(const pcre_uchar *,
2714                               const pcre_uchar *);
2715    extern int               PRIV(strcmp_uc_c8_utf)(const pcre_uchar *,
2716                               const char *);
2717    
2718    #define STRCMP_UC_UC_TEST(str1, str2) \
2719      (utf ? PRIV(strcmp_uc_uc_utf)((str1), (str2)) : PRIV(strcmp_uc_uc)((str1), (str2)))
2720    #define STRCMP_UC_C8_TEST(str1, str2) \
2721      (utf ? PRIV(strcmp_uc_c8_utf)((str1), (str2)) : PRIV(strcmp_uc_c8)((str1), (str2)))
2722    
2723    #endif /* COMPILE_PCRE[8|16|32] */
2724    
2725    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2726    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2727                               int *, BOOL);
2728    extern unsigned int      PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2729    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2730    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2731                               int *, BOOL);
2732    extern BOOL              PRIV(xclass)(pcre_uint32, const pcre_uchar *, BOOL);
2733    
2734  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
2735  extern void          _pcre_jit_compile(const real_pcre *, pcre_extra *);  extern void              PRIV(jit_compile)(const REAL_PCRE *,
2736  extern int           _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR,                             PUBL(extra) *, int);
2737                          int, int, int, int, int *, int);  extern int               PRIV(jit_exec)(const PUBL(extra) *,
2738  extern void          _pcre_jit_free(void *);                             const pcre_uchar *, int, int, int, int *, int);
2739    extern void              PRIV(jit_free)(void *);
2740    extern int               PRIV(jit_get_size)(void *);
2741    extern const char*       PRIV(jit_get_target)(void);
2742  #endif  #endif
2743    
2744  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
2745    
2746  typedef struct {  typedef struct {
2747    uschar script;    pcre_uint8 script;     /* ucp_Arabic, etc. */
2748    uschar chartype;    pcre_uint8 chartype;   /* ucp_Cc, etc. (general categories) */
2749    pcre_int32 other_case;    pcre_uint8 gbprop;     /* ucp_gbControl, etc. (grapheme break property) */
2750      pcre_uint8 caseset;    /* offset to multichar other cases or zero */
2751      pcre_int32 other_case; /* offset to other case, or zero if none */
2752  } ucd_record;  } ucd_record;
2753    
2754  extern const ucd_record  _pcre_ucd_records[];  extern const pcre_uint32 PRIV(ucd_caseless_sets)[];
2755  extern const uschar      _pcre_ucd_stage1[];  extern const ucd_record  PRIV(ucd_records)[];
2756  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint8  PRIV(ucd_stage1)[];
2757  extern const int         _pcre_ucp_gentype[];  extern const pcre_uint16 PRIV(ucd_stage2)[];
2758    extern const pcre_uint32 PRIV(ucp_gentype)[];
2759    extern const pcre_uint32 PRIV(ucp_gbtable)[];
2760  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
2761  extern const int         _pcre_ucp_typerange[];  extern const int         PRIV(ucp_typerange)[];
2762  #endif  #endif
2763    
2764    #ifdef SUPPORT_UCP
2765  /* UCD access macros */  /* UCD access macros */
2766    
2767  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
2768  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (PRIV(ucd_records) + \
2769          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          PRIV(ucd_stage2)[PRIV(ucd_stage1)[(int)(ch) / UCD_BLOCK_SIZE] * \
2770          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (int)(ch) % UCD_BLOCK_SIZE])
2771    
2772  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)    GET_UCD(ch)->chartype
2773  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)      GET_UCD(ch)->script
2774  #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]  #define UCD_CATEGORY(ch)    PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2775  #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)  #define UCD_GRAPHBREAK(ch)  GET_UCD(ch)->gbprop
2776    #define UCD_CASESET(ch)     GET_UCD(ch)->caseset
2777    #define UCD_OTHERCASE(ch)   ((pcre_uint32)((int)ch + (int)(GET_UCD(ch)->other_case)))
2778    
2779    #endif /* SUPPORT_UCP */
2780    
2781  #endif  #endif
2782    

Legend:
Removed from v.733  
changed lines
  Added in v.1359

  ViewVC Help
Powered by ViewVC 1.1.5