/[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 758 by ph10, Mon Nov 21 12:05:36 2011 UTC revision 1309 by ph10, Fri Apr 5 15:35:59 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;
# Line 171  preprocessor time in standard C environm Line 200  preprocessor time in standard C environm
200    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
201    typedef int pcre_int16;    typedef int pcre_int16;
202  #else  #else
203    #error Cannot determine a type for 16-bit unsigned integers  # error Cannot determine a type for 16-bit unsigned integers
204  #endif  #endif
205    
206  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
# Line 181  preprocessor time in standard C environm Line 210  preprocessor time in standard C environm
210    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
211    typedef long int pcre_int32;    typedef long int pcre_int32;
212  #else  #else
213    #error Cannot determine a type for 32-bit unsigned integers  # error Cannot determine a type for 32-bit unsigned integers
214  #endif  #endif
215    
216  /* 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 221  stdint.h is available, include it; it ma
221  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
222  by "configure". */  by "configure". */
223    
224  #if HAVE_STDINT_H  #if defined HAVE_STDINT_H
225  #include <stdint.h>  #include <stdint.h>
226  #elif HAVE_INTTYPES_H  #elif defined HAVE_INTTYPES_H
227  #include <inttypes.h>  #include <inttypes.h>
228  #endif  #endif
229    
# Line 206  by "configure". */ Line 235  by "configure". */
235    
236  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
237  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
238  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 *,
239  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
240  to save lots of typing. I tried "uchar", but it causes problems on Digital  called pcre_uchar.
241  Unix, where it is defined in sys/types, so use "uschar" instead. */  
242    The IN_UCHARS macro multiply its argument with the byte size of the current
243    pcre_uchar type. Useful for memcpy and such operations, whose require the
244    byte size of their input/output buffers.
245    
246    The MAX_255 macro checks whether its pcre_uchar input is less than 256.
247    
248    The TABLE_GET macro is designed for accessing elements of tables whose contain
249    exactly 256 items. When the character is able to contain more than 256
250    items, some check is needed before accessing these tables.
251    */
252    
253    #if defined COMPILE_PCRE8
254    
255  typedef unsigned char uschar;  typedef unsigned char pcre_uchar;
256    #define IN_UCHARS(x) (x)
257    #define MAX_255(c) 1
258    #define TABLE_GET(c, table, default) ((table)[c])
259    
260    #elif defined COMPILE_PCRE16
261    
262    #if USHRT_MAX != 65535
263    /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
264    pcre.h(.in) and disable (comment out) this message. */
265    #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
266    #endif
267    
268    typedef pcre_uint16 pcre_uchar;
269    #define UCHAR_SHIFT (1)
270    #define IN_UCHARS(x) ((x) << UCHAR_SHIFT)
271    #define MAX_255(c) ((c) <= 255u)
272    #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
273    
274    #elif defined COMPILE_PCRE32
275    
276    typedef pcre_uint32 pcre_uchar;
277    #define UCHAR_SHIFT (2)
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    #else
283    #error Unsupported compiling mode
284    #endif /* COMPILE_PCRE[8|16|32] */
285    
286  /* 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
287  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 304  start/end of string field names are. */
304  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
305    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
306      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
307       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\       PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
308         utf8)) \         &(NLBLOCK->nllen), utf)) \
309      : \      : \
310      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
311       (p)[0] == NLBLOCK->nl[0] && \       RAWUCHARTEST(p) == NLBLOCK->nl[0] && \
312       (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \       (NLBLOCK->nllen == 1 || RAWUCHARTEST(p+1) == NLBLOCK->nl[1])       \
313      ) \      ) \
314    )    )
315    
# Line 248  start/end of string field names are. */ Line 318  start/end of string field names are. */
318  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
319    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
320      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
321       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \       PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
322         &(NLBLOCK->nllen), utf8)) \         &(NLBLOCK->nllen), utf)) \
323      : \      : \
324      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
325       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       RAWUCHARTEST(p - NLBLOCK->nllen) == NLBLOCK->nl[0] &&              \
326       (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \       (NLBLOCK->nllen == 1 || RAWUCHARTEST(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \
327      ) \      ) \
328    )    )
329    
# Line 267  used for the external interface and appe Line 337  used for the external interface and appe
337  must begin with PCRE_. */  must begin with PCRE_. */
338    
339  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
340  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
341  #else  #else
342  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
343  #endif  #endif
344    
   
   
345  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
346  values. */  values. */
347    
348  #include "pcre.h"  #include "pcre.h"
349  #include "ucp.h"  #include "ucp.h"
350    
351    #ifdef COMPILE_PCRE32
352    /* Assert that the public PCRE_UCHAR32 is a 32-bit type */
353    typedef int __assert_pcre_uchar32_size[sizeof(PCRE_UCHAR32) == 4 ? 1 : -1];
354    #endif
355    
356  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
357  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
358  option on the command line. */  option on the command line. */
# Line 343  The macros are controlled by the value o Line 414  The macros are controlled by the value o
414  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
415  is automated on Unix systems via the "configure" command. */  is automated on Unix systems via the "configure" command. */
416    
417    #if defined COMPILE_PCRE8
418    
419  #if LINK_SIZE == 2  #if LINK_SIZE == 2
420    
421  #define PUT(a,n,d)   \  #define PUT(a,n,d)   \
# Line 379  is automated on Unix systems via the "co Line 452  is automated on Unix systems via the "co
452  #define GET(a,n) \  #define GET(a,n) \
453    (((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])
454    
455  #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */  /* Keep it positive */
456    #define MAX_PATTERN_SIZE (1 << 30)
457    
458    #else
459    #error LINK_SIZE must be either 2, 3, or 4
460    #endif
461    
462    #elif defined COMPILE_PCRE16
463    
464    #if LINK_SIZE == 2
465    
466    /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
467    #undef LINK_SIZE
468    #define LINK_SIZE 1
469    
470    #define PUT(a,n,d)   \
471      (a[n] = (d))
472    
473    #define GET(a,n) \
474      (a[n])
475    
476    #define MAX_PATTERN_SIZE (1 << 16)
477    
478    #elif LINK_SIZE == 3 || LINK_SIZE == 4
479    
480    /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
481    #undef LINK_SIZE
482    #define LINK_SIZE 2
483    
484    #define PUT(a,n,d)   \
485      (a[n] = (d) >> 16), \
486      (a[(n)+1] = (d) & 65535)
487    
488    #define GET(a,n) \
489      (((a)[n] << 16) | (a)[(n)+1])
490    
491    /* Keep it positive */
492    #define MAX_PATTERN_SIZE (1 << 30)
493    
494  #else  #else
495  #error LINK_SIZE must be either 2, 3, or 4  #error LINK_SIZE must be either 2, 3, or 4
496  #endif  #endif
497    
498    #elif defined COMPILE_PCRE32
499    
500    /* Only supported LINK_SIZE is 4 */
501    /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
502    #undef LINK_SIZE
503    #define LINK_SIZE 1
504    
505    #define PUT(a,n,d)   \
506      (a[n] = (d))
507    
508    #define GET(a,n) \
509      (a[n])
510    
511    /* Keep it positive */
512    #define MAX_PATTERN_SIZE (1 << 30)
513    
514    #else
515    #error Unsupported compiling mode
516    #endif /* COMPILE_PCRE[8|16|32] */
517    
518  /* Convenience macro defined in terms of the others */  /* Convenience macro defined in terms of the others */
519    
# Line 396  is automated on Unix systems via the "co Line 524  is automated on Unix systems via the "co
524  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
525  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
526    
527    #if defined COMPILE_PCRE8
528    
529    #define IMM2_SIZE 2
530    
531  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
532    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
533    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
534    
535    /* For reasons that I do not understand, the expression in this GET2 macro is
536    treated by gcc as a signed expression, even when a is declared as unsigned. It
537    seems that any kind of arithmetic results in a signed value. */
538    
539  #define GET2(a,n) \  #define GET2(a,n) \
540    (((a)[n] << 8) | (a)[(n)+1])    (unsigned int)(((a)[n] << 8) | (a)[(n)+1])
541    
542  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  #elif defined COMPILE_PCRE16
543    
544    #define IMM2_SIZE 1
545    
546  /* When UTF-8 encoding is being used, a character is no longer just a single  #define PUT2(a,n,d)   \
547  byte. The macros for character handling generate simple sequences when used in     a[n] = d
548  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is  
549  not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should  #define GET2(a,n) \
550  never be called in byte mode. To make sure they can never even appear when     a[n]
551  UTF-8 support is omitted, we don't even define them. */  
552    #elif defined COMPILE_PCRE32
553    
554    #define IMM2_SIZE 1
555    
556    #define PUT2(a,n,d)   \
557       a[n] = d
558    
559    #define GET2(a,n) \
560       a[n]
561    
562  #ifndef SUPPORT_UTF8  #else
563    #error Unsupported compiling mode
564    #endif /* COMPILE_PCRE[8|16|32] */
565    
566    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
567    
568    /* The maximum length of a MARK name is currently one data unit; it may be
569    changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */
570    
571    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
572    #define MAX_MARK ((1u << 16) - 1)
573    #else
574    #define MAX_MARK ((1u << 8) - 1)
575    #endif
576    
577    /* When UTF encoding is being used, a character is no longer just a single
578    byte. The macros for character handling generate simple sequences when used in
579    character-mode, and more complicated ones for UTF characters. GETCHARLENTEST
580    and other macros are not used when UTF is not supported, so they are not
581    defined. To make sure they can never even appear when UTF support is omitted,
582    we don't even define them. */
583    
584    #ifndef SUPPORT_UTF
585    
586    /* #define MAX_VALUE_FOR_SINGLE_CHAR */
587    /* #define HAS_EXTRALEN(c) */
588    /* #define GET_EXTRALEN(c) */
589    /* #define NOT_FIRSTCHAR(c) */
590  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
591  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
592  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
593  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
594  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
595    #define RAWUCHAR(eptr) (*(eptr))
596    #define RAWUCHARINC(eptr) (*(eptr)++)
597    #define RAWUCHARTEST(eptr) (*(eptr))
598    #define RAWUCHARINCTEST(eptr) (*(eptr)++)
599  /* #define GETCHARLENTEST(c, eptr, len) */  /* #define GETCHARLENTEST(c, eptr, len) */
600  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
601    /* #define FORWARDCHAR(eptr) */
602    /* #define ACROSSCHAR(condition, eptr, action) */
603    
604  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF */
605    
606  /* These macros were originally written in the form of loops that used data  /* Tests whether the code point needs extra characters to decode. */
607  from the tables whose names start with _pcre_utf8_table. They were rewritten by  
608  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. */  
609    
610  /* 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
611  advancing the pointer. */  advancing the pointer. */
# Line 451  advancing the pointer. */ Line 629  advancing the pointer. */
629            ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \            ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
630      }      }
631    
 /* 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);  
   
632  /* 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
633  the pointer. */  the pointer. */
634    
# Line 499  the pointer. */ Line 663  the pointer. */
663        } \        } \
664      }      }
665    
666    #if defined COMPILE_PCRE8
667    
668    /* These macros were originally written in the form of loops that used data
669    from the tables whose names start with PRIV(utf8_table). They were rewritten by
670    a user so as not to use loops, because in some environments this gives a
671    significant performance advantage, and it seems never to do any harm. */
672    
673    /* Tells the biggest code point which can be encoded as a single character. */
674    
675    #define MAX_VALUE_FOR_SINGLE_CHAR 127
676    
677    /* Tests whether the code point needs extra characters to decode. */
678    
679    #define HAS_EXTRALEN(c) ((c) >= 0xc0)
680    
681    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
682    Otherwise it has an undefined behaviour. */
683    
684    #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
685    
686    /* Returns TRUE, if the given character is not the first character
687    of a UTF sequence. */
688    
689    #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
690    
691    /* Get the next UTF-8 character, not advancing the pointer. This is called when
692    we know we are in UTF-8 mode. */
693    
694    #define GETCHAR(c, eptr) \
695      c = *eptr; \
696      if (c >= 0xc0) GETUTF8(c, eptr);
697    
698    /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
699    pointer. */
700    
701    #define GETCHARTEST(c, eptr) \
702      c = *eptr; \
703      if (utf && c >= 0xc0) GETUTF8(c, eptr);
704    
705  /* 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
706  know we are in UTF-8 mode. */  know we are in UTF-8 mode. */
707    
# Line 511  This is called when we don't know if we Line 714  This is called when we don't know if we
714    
715  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
716    c = *eptr++; \    c = *eptr++; \
717    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);    if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
718    
719  /* 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
720  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 766  do not know if we are in UTF-8 mode. */
766    
767  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
768    c = *eptr; \    c = *eptr; \
769    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
770    
771    /* Returns the next uchar, not advancing the pointer. This is called when
772    we know we are in UTF mode. */
773    
774    #define RAWUCHAR(eptr) \
775      (*(eptr))
776    
777    /* Returns the next uchar, advancing the pointer. This is called when
778    we know we are in UTF mode. */
779    
780    #define RAWUCHARINC(eptr) \
781      (*((eptr)++))
782    
783    /* Returns the next uchar, testing for UTF mode, and not advancing the
784    pointer. */
785    
786    #define RAWUCHARTEST(eptr) \
787      (*(eptr))
788    
789    /* Returns the next uchar, testing for UTF mode, advancing the
790    pointer. */
791    
792    #define RAWUCHARINCTEST(eptr) \
793      (*((eptr)++))
794    
795  /* 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
796  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 798  because almost all calls are already wit
798    
799  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
800    
801  #endif  /* SUPPORT_UTF8 */  /* Same as above, just in the other direction. */
802    #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
803    
804    /* Same as above, but it allows a fully customizable form. */
805    #define ACROSSCHAR(condition, eptr, action) \
806      while((condition) && ((eptr) & 0xc0) == 0x80) action
807    
808    #elif defined COMPILE_PCRE16
809    
810    /* Tells the biggest code point which can be encoded as a single character. */
811    
812    #define MAX_VALUE_FOR_SINGLE_CHAR 65535
813    
814    /* Tests whether the code point needs extra characters to decode. */
815    
816    #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
817    
818    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
819    Otherwise it has an undefined behaviour. */
820    
821    #define GET_EXTRALEN(c) 1
822    
823    /* Returns TRUE, if the given character is not the first character
824    of a UTF sequence. */
825    
826    #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
827    
828    /* Base macro to pick up the low surrogate of a UTF-16 character, not
829    advancing the pointer. */
830    
831    #define GETUTF16(c, eptr) \
832       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
833    
834    /* Get the next UTF-16 character, not advancing the pointer. This is called when
835    we know we are in UTF-16 mode. */
836    
837    #define GETCHAR(c, eptr) \
838      c = *eptr; \
839      if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
840    
841    /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
842    pointer. */
843    
844    #define GETCHARTEST(c, eptr) \
845      c = *eptr; \
846      if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
847    
848    /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
849    the pointer. */
850    
851    #define GETUTF16INC(c, eptr) \
852       { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
853    
854    /* Get the next UTF-16 character, advancing the pointer. This is called when we
855    know we are in UTF-16 mode. */
856    
857    #define GETCHARINC(c, eptr) \
858      c = *eptr++; \
859      if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
860    
861    /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
862    This is called when we don't know if we are in UTF-16 mode. */
863    
864    #define GETCHARINCTEST(c, eptr) \
865      c = *eptr++; \
866      if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
867    
868    /* Base macro to pick up the low surrogate of a UTF-16 character, not
869    advancing the pointer, incrementing the length. */
870    
871    #define GETUTF16LEN(c, eptr, len) \
872       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
873    
874    /* Get the next UTF-16 character, not advancing the pointer, incrementing
875    length if there is a low surrogate. This is called when we know we are in
876    UTF-16 mode. */
877    
878    #define GETCHARLEN(c, eptr, len) \
879      c = *eptr; \
880      if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
881    
882    /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
883    pointer, incrementing length if there is a low surrogate. This is called when
884    we do not know if we are in UTF-16 mode. */
885    
886    #define GETCHARLENTEST(c, eptr, len) \
887      c = *eptr; \
888      if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
889    
890    /* Returns the next uchar, not advancing the pointer. This is called when
891    we know we are in UTF mode. */
892    
893    #define RAWUCHAR(eptr) \
894      (*(eptr))
895    
896    /* Returns the next uchar, advancing the pointer. This is called when
897    we know we are in UTF mode. */
898    
899    #define RAWUCHARINC(eptr) \
900      (*((eptr)++))
901    
902    /* Returns the next uchar, testing for UTF mode, and not advancing the
903    pointer. */
904    
905    #define RAWUCHARTEST(eptr) \
906      (*(eptr))
907    
908    /* Returns the next uchar, testing for UTF mode, advancing the
909    pointer. */
910    
911  /* In case there is no definition of offsetof() provided - though any proper  #define RAWUCHARINCTEST(eptr) \
912  Standard C system should have one. */    (*((eptr)++))
913    
914    /* If the pointer is not at the start of a character, move it back until
915    it is. This is called only in UTF-16 mode - we don't put a test within the
916    macro because almost all calls are already within a block of UTF-16 only
917    code. */
918    
919    #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
920    
921    /* Same as above, just in the other direction. */
922    #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
923    
924    /* Same as above, but it allows a fully customizable form. */
925    #define ACROSSCHAR(condition, eptr, action) \
926      if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
927    
928    #elif defined COMPILE_PCRE32
929    
930    /* These are trivial for the 32-bit library, since all UTF-32 characters fit
931    into one pcre_uchar unit. */
932    #define MAX_VALUE_FOR_SINGLE_CHAR (0x10ffffu)
933    #define HAS_EXTRALEN(c) (0)
934    #define GET_EXTRALEN(c) (0)
935    #define NOT_FIRSTCHAR(c) (0)
936    
937    /* Get the next UTF-32 character, not advancing the pointer. This is called when
938    we know we are in UTF-32 mode. */
939    
940    #define GETCHAR(c, eptr) \
941      c = *(eptr);
942    
943    /* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the
944    pointer. */
945    
946    #define GETCHARTEST(c, eptr) \
947      c = *(eptr);
948    
949  #ifndef offsetof  /* Get the next UTF-32 character, advancing the pointer. This is called when we
950  #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))  know we are in UTF-32 mode. */
951    
952    #define GETCHARINC(c, eptr) \
953      c = *((eptr)++);
954    
955    /* Get the next character, testing for UTF-32 mode, and advancing the pointer.
956    This is called when we don't know if we are in UTF-32 mode. */
957    
958    #define GETCHARINCTEST(c, eptr) \
959      c = *((eptr)++);
960    
961    /* Get the next UTF-32 character, not advancing the pointer, not incrementing
962    length (since all UTF-32 is of length 1). This is called when we know we are in
963    UTF-32 mode. */
964    
965    #define GETCHARLEN(c, eptr, len) \
966      GETCHAR(c, eptr)
967    
968    /* Get the next UTF-32character, testing for UTF-32 mode, not advancing the
969    pointer, not incrementing the length (since all UTF-32 is of length 1).
970    This is called when we do not know if we are in UTF-32 mode. */
971    
972    #define GETCHARLENTEST(c, eptr, len) \
973      GETCHARTEST(c, eptr)
974    
975    /* Returns the next uchar, not advancing the pointer. This is called when
976    we know we are in UTF mode. */
977    
978    #define RAWUCHAR(eptr) \
979      (*(eptr))
980    
981    /* Returns the next uchar, advancing the pointer. This is called when
982    we know we are in UTF mode. */
983    
984    #define RAWUCHARINC(eptr) \
985      (*((eptr)++))
986    
987    /* Returns the next uchar, testing for UTF mode, and not advancing the
988    pointer. */
989    
990    #define RAWUCHARTEST(eptr) \
991      (*(eptr))
992    
993    /* Returns the next uchar, testing for UTF mode, advancing the
994    pointer. */
995    
996    #define RAWUCHARINCTEST(eptr) \
997      (*((eptr)++))
998    
999    /* If the pointer is not at the start of a character, move it back until
1000    it is. This is called only in UTF-32 mode - we don't put a test within the
1001    macro because almost all calls are already within a block of UTF-32 only
1002    code.
1003    These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */
1004    
1005    #define BACKCHAR(eptr) do { } while (0)
1006    
1007    /* Same as above, just in the other direction. */
1008    #define FORWARDCHAR(eptr) do { } while (0)
1009    
1010    /* Same as above, but it allows a fully customizable form. */
1011    #define ACROSSCHAR(condition, eptr, action) do { } while (0)
1012    
1013    #else
1014    #error Unsupported compiling mode
1015    #endif /* COMPILE_PCRE[8|16|32] */
1016    
1017    #endif  /* SUPPORT_UTF */
1018    
1019    /* Tests for Unicode horizontal and vertical whitespace characters must check a
1020    number of different values. Using a switch statement for this generates the
1021    fastest code (no loop, no memory access), and there are several places in the
1022    interpreter code where this happens. In order to ensure that all the case lists
1023    remain in step, we use macros so that there is only one place where the lists
1024    are defined.
1025    
1026    These values are also required as lists in pcre_compile.c when processing \h,
1027    \H, \v and \V in a character class. The lists are defined in pcre_tables.c, but
1028    macros that define the values are here so that all the definitions are
1029    together. The lists must be in ascending character order, terminated by
1030    NOTACHAR (which is 0xffffffff).
1031    
1032    Any changes should ensure that the various macros are kept in step with each
1033    other. NOTE: The values also appear in pcre_jit_compile.c. */
1034    
1035    /* ------ ASCII/Unicode environments ------ */
1036    
1037    #ifndef EBCDIC
1038    
1039    #define HSPACE_LIST \
1040      CHAR_HT, CHAR_SPACE, 0xa0, \
1041      0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \
1042      0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \
1043      NOTACHAR
1044    
1045    #define HSPACE_MULTIBYTE_CASES \
1046      case 0x1680:  /* OGHAM SPACE MARK */ \
1047      case 0x180e:  /* MONGOLIAN VOWEL SEPARATOR */ \
1048      case 0x2000:  /* EN QUAD */ \
1049      case 0x2001:  /* EM QUAD */ \
1050      case 0x2002:  /* EN SPACE */ \
1051      case 0x2003:  /* EM SPACE */ \
1052      case 0x2004:  /* THREE-PER-EM SPACE */ \
1053      case 0x2005:  /* FOUR-PER-EM SPACE */ \
1054      case 0x2006:  /* SIX-PER-EM SPACE */ \
1055      case 0x2007:  /* FIGURE SPACE */ \
1056      case 0x2008:  /* PUNCTUATION SPACE */ \
1057      case 0x2009:  /* THIN SPACE */ \
1058      case 0x200A:  /* HAIR SPACE */ \
1059      case 0x202f:  /* NARROW NO-BREAK SPACE */ \
1060      case 0x205f:  /* MEDIUM MATHEMATICAL SPACE */ \
1061      case 0x3000   /* IDEOGRAPHIC SPACE */
1062    
1063    #define HSPACE_BYTE_CASES \
1064      case CHAR_HT: \
1065      case CHAR_SPACE: \
1066      case 0xa0     /* NBSP */
1067    
1068    #define HSPACE_CASES \
1069      HSPACE_BYTE_CASES: \
1070      HSPACE_MULTIBYTE_CASES
1071    
1072    #define VSPACE_LIST \
1073      CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR
1074    
1075    #define VSPACE_MULTIBYTE_CASES \
1076      case 0x2028:    /* LINE SEPARATOR */ \
1077      case 0x2029     /* PARAGRAPH SEPARATOR */
1078    
1079    #define VSPACE_BYTE_CASES \
1080      case CHAR_LF: \
1081      case CHAR_VT: \
1082      case CHAR_FF: \
1083      case CHAR_CR: \
1084      case CHAR_NEL
1085    
1086    #define VSPACE_CASES \
1087      VSPACE_BYTE_CASES: \
1088      VSPACE_MULTIBYTE_CASES
1089    
1090    /* ------ EBCDIC environments ------ */
1091    
1092    #else
1093    #define HSPACE_LIST CHAR_HT, CHAR_SPACE
1094    
1095    #define HSPACE_BYTE_CASES \
1096      case CHAR_HT: \
1097      case CHAR_SPACE
1098    
1099    #define HSPACE_CASES HSPACE_BYTE_CASES
1100    
1101    #ifdef EBCDIC_NL25
1102    #define VSPACE_LIST \
1103      CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR
1104    #else
1105    #define VSPACE_LIST \
1106      CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR
1107  #endif  #endif
1108    
1109    #define VSPACE_BYTE_CASES \
1110      case CHAR_LF: \
1111      case CHAR_VT: \
1112      case CHAR_FF: \
1113      case CHAR_CR: \
1114      case CHAR_NEL
1115    
1116    #define VSPACE_CASES VSPACE_BYTE_CASES
1117    #endif  /* EBCDIC */
1118    
1119    /* ------ End of whitespace macros ------ */
1120    
1121    
1122    
1123  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
1124  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 now they
# Line 588  are in a 16-bit flags word. From release Line 1126  are in a 16-bit flags word. From release
1126  the restrictions on partial matching have been lifted. It remains for backwards  the restrictions on partial matching have been lifted. It remains for backwards
1127  compatibility. */  compatibility. */
1128    
1129  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_MODE8         0x0001  /* compiled in 8 bit mode */
1130  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_MODE16        0x0002  /* compiled in 16 bit mode */
1131  #define PCRE_REQCHSET      0x0004  /* req_byte is set */  #define PCRE_MODE32        0x0004  /* compiled in 32 bit mode */
1132  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #define PCRE_FIRSTSET      0x0010  /* first_char is set */
1133  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_FCH_CASELESS  0x0020  /* caseless first char */
1134  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #define PCRE_REQCHSET      0x0040  /* req_byte is set */
1135  #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */  #define PCRE_RCH_CASELESS  0x0080  /* caseless requested char */
1136    #define PCRE_STARTLINE     0x0100  /* start after \n for multiline */
1137    #define PCRE_NOPARTIAL     0x0200  /* can't use partial with this regex */
1138    #define PCRE_JCHANGED      0x0400  /* j option used in regex */
1139    #define PCRE_HASCRORLF     0x0800  /* explicit \r or \n in pattern */
1140    #define PCRE_HASTHEN       0x1000  /* pattern contains (*THEN) */
1141    
1142    #if defined COMPILE_PCRE8
1143    #define PCRE_MODE          PCRE_MODE8
1144    #elif defined COMPILE_PCRE16
1145    #define PCRE_MODE          PCRE_MODE16
1146    #elif defined COMPILE_PCRE32
1147    #define PCRE_MODE          PCRE_MODE32
1148    #endif
1149    #define PCRE_MODE_MASK     (PCRE_MODE8 | PCRE_MODE16 | PCRE_MODE32)
1150    
1151  /* Flags for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
1152    
# Line 612  time, run time, or study time, respectiv Line 1164  time, run time, or study time, respectiv
1164     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
1165     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| \
1166     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1167     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
1168    
1169  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
1170    (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 1178  time, run time, or study time, respectiv
1178     PCRE_NO_START_OPTIMIZE)     PCRE_NO_START_OPTIMIZE)
1179    
1180  #define PUBLIC_STUDY_OPTIONS \  #define PUBLIC_STUDY_OPTIONS \
1181     PCRE_STUDY_JIT_COMPILE     (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
1182        PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)
1183    
1184  /* Magic number to provide a small check against being handed junk. Also used  #define PUBLIC_JIT_EXEC_OPTIONS \
1185  to detect whether a pattern was compiled on a host of different endianness. */     (PCRE_NO_UTF8_CHECK|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|\
1186        PCRE_NOTEMPTY_ATSTART|PCRE_PARTIAL_SOFT|PCRE_PARTIAL_HARD)
1187    
1188    /* Magic number to provide a small check against being handed junk. */
1189    
1190  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
1191    
1192  /* Negative values for the firstchar and reqchar variables */  /* This variable is used to detect a loaded regular expression
1193    in different endianness. */
1194    
1195  #define REQ_UNSET (-2)  #define REVERSED_MAGIC_NUMBER  0x45524350UL   /* 'ERCP' */
 #define REQ_NONE  (-1)  
1196    
1197  /* 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
1198  req_byte match. */  req_byte match. */
1199    
1200  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
1201    
 /* 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 */  
   
1202  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
1203  environments where these macros are defined elsewhere. Unfortunately, there  environments where these macros are defined elsewhere. Unfortunately, there
1204  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 1227  for) in a minority area (EBCDIC platform
1227  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
1228  macros to give the functions distinct names. */  macros to give the functions distinct names. */
1229    
1230  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF
1231    
1232  /* UTF-8 support is not enabled; use the platform-dependent character literals  /* UTF-8 support is not enabled; use the platform-dependent character literals
1233  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
1234    mode. Newline characters are problematic in EBCDIC. Though it has CR and LF
1235    characters, a common practice has been to use its NL (0x15) character as the
1236    line terminator in C-like processing environments. However, sometimes the LF
1237    (0x25) character is used instead, according to this Unicode document:
1238    
1239    http://unicode.org/standard/reports/tr13/tr13-5.html
1240    
1241    PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25
1242    instead. Whichever is *not* chosen is defined as NEL.
1243    
1244    In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the
1245    same code point. */
1246    
1247    #ifdef EBCDIC
1248    
1249    #ifndef EBCDIC_NL25
1250    #define CHAR_NL                     '\x15'
1251    #define CHAR_NEL                    '\x25'
1252    #define STR_NL                      "\x15"
1253    #define STR_NEL                     "\x25"
1254    #else
1255    #define CHAR_NL                     '\x25'
1256    #define CHAR_NEL                    '\x15'
1257    #define STR_NL                      "\x25"
1258    #define STR_NEL                     "\x15"
1259    #endif
1260    
1261    #define CHAR_LF                     CHAR_NL
1262    #define STR_LF                      STR_NL
1263    
1264    #define CHAR_ESC                    '\047'
1265    #define CHAR_DEL                    '\007'
1266    #define STR_ESC                     "\047"
1267    #define STR_DEL                     "\007"
1268    
1269    #else  /* Not EBCDIC */
1270    
1271    /* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for
1272    compatibility. NEL is the Unicode newline character; make sure it is
1273    a positive value. */
1274    
1275    #define CHAR_LF                     '\n'
1276    #define CHAR_NL                     CHAR_LF
1277    #define CHAR_NEL                    ((unsigned char)'\x85')
1278    #define CHAR_ESC                    '\033'
1279    #define CHAR_DEL                    '\177'
1280    
1281    #define STR_LF                      "\n"
1282    #define STR_NL                      STR_LF
1283    #define STR_NEL                     "\x85"
1284    #define STR_ESC                     "\033"
1285    #define STR_DEL                     "\177"
1286    
1287    #endif  /* EBCDIC */
1288    
1289    /* The remaining definitions work in both environments. */
1290    
1291    #define CHAR_NULL                   '\0'
1292  #define CHAR_HT                     '\t'  #define CHAR_HT                     '\t'
1293  #define CHAR_VT                     '\v'  #define CHAR_VT                     '\v'
1294  #define CHAR_FF                     '\f'  #define CHAR_FF                     '\f'
1295  #define CHAR_CR                     '\r'  #define CHAR_CR                     '\r'
 #define CHAR_NL                     '\n'  
1296  #define CHAR_BS                     '\b'  #define CHAR_BS                     '\b'
1297  #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  
1298    
1299  #define CHAR_SPACE                  ' '  #define CHAR_SPACE                  ' '
1300  #define CHAR_EXCLAMATION_MARK       '!'  #define CHAR_EXCLAMATION_MARK       '!'
# Line 797  so that PCRE works on both ASCII and EBC Line 1396  so that PCRE works on both ASCII and EBC
1396  #define STR_VT                      "\v"  #define STR_VT                      "\v"
1397  #define STR_FF                      "\f"  #define STR_FF                      "\f"
1398  #define STR_CR                      "\r"  #define STR_CR                      "\r"
 #define STR_NL                      "\n"  
1399  #define STR_BS                      "\b"  #define STR_BS                      "\b"
1400  #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  
1401    
1402  #define STR_SPACE                   " "  #define STR_SPACE                   " "
1403  #define STR_EXCLAMATION_MARK        "!"  #define STR_EXCLAMATION_MARK        "!"
# Line 938  so that PCRE works on both ASCII and EBC Line 1529  so that PCRE works on both ASCII and EBC
1529  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1530  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1531  #define STRING_UTF8_RIGHTPAR           "UTF8)"  #define STRING_UTF8_RIGHTPAR           "UTF8)"
1532    #define STRING_UTF16_RIGHTPAR          "UTF16)"
1533    #define STRING_UTF32_RIGHTPAR          "UTF32)"
1534    #define STRING_UTF_RIGHTPAR            "UTF)"
1535  #define STRING_UCP_RIGHTPAR            "UCP)"  #define STRING_UCP_RIGHTPAR            "UCP)"
1536  #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"  #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1537    
1538  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF */
1539    
1540  /* 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
1541  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 1545  only. */
1545  #define CHAR_VT                     '\013'  #define CHAR_VT                     '\013'
1546  #define CHAR_FF                     '\014'  #define CHAR_FF                     '\014'
1547  #define CHAR_CR                     '\015'  #define CHAR_CR                     '\015'
1548  #define CHAR_NL                     '\012'  #define CHAR_LF                     '\012'
1549    #define CHAR_NL                     CHAR_LF
1550    #define CHAR_NEL                    ((unsigned char)'\x85')
1551  #define CHAR_BS                     '\010'  #define CHAR_BS                     '\010'
1552  #define CHAR_BEL                    '\007'  #define CHAR_BEL                    '\007'
1553  #define CHAR_ESC                    '\033'  #define CHAR_ESC                    '\033'
1554  #define CHAR_DEL                    '\177'  #define CHAR_DEL                    '\177'
1555    
1556    #define CHAR_NULL                   '\0'
1557  #define CHAR_SPACE                  '\040'  #define CHAR_SPACE                  '\040'
1558  #define CHAR_EXCLAMATION_MARK       '\041'  #define CHAR_EXCLAMATION_MARK       '\041'
1559  #define CHAR_QUOTATION_MARK         '\042'  #define CHAR_QUOTATION_MARK         '\042'
# Line 1193  only. */ Line 1790  only. */
1790  #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
1791  #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
1792  #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
1793    #define STRING_UTF16_RIGHTPAR          STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1794    #define STRING_UTF32_RIGHTPAR          STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS
1795    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
1796  #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
1797  #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
1798    
1799  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
1800    
1801  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1802    
# Line 1209  only. */ Line 1809  only. */
1809  #endif  #endif
1810    
1811  #ifndef ESC_n  #ifndef ESC_n
1812  #define ESC_n CHAR_NL  #define ESC_n CHAR_LF
1813  #endif  #endif
1814    
1815  #ifndef ESC_r  #ifndef ESC_r
# Line 1234  only. */ Line 1834  only. */
1834  #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */  #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1835  #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 */
1836  #define PT_WORD       8    /* Word - L plus N plus underscore */  #define PT_WORD       8    /* Word - L plus N plus underscore */
1837    #define PT_CLIST      9    /* Pseudo-property: match character list */
1838    #define PT_UCNC      10    /* Universal Character nameable character */
1839    
1840  /* 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
1841  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1842    
1843  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1844  #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 1851  contain UTF-8 characters with values gre
1851    
1852  /* 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
1853  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
1854  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
1855  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
1856  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
1857  used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1858  non-DOTALL mode, "." behaves like \N.  non-DOTALL mode, "." behaves like \N.
1859    
1860  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.
1861  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.
1862  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
1863  looked up from a table.  looked up from a table.
1864    
1865  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
1866  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
1867  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
1868  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
1869  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 1873  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1873         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,
1874         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1875         ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_E, ESC_Q, ESC_g, ESC_k,
1876         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 };  
1877    
1878  /* 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
1879  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 1899  enum {
1899    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1900    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1901    
1902    OP_ANY,            /* 12 Match any character except newline */    OP_ANY,            /* 12 Match any character except newline (\N) */
1903    OP_ALLANY,         /* 13 Match any character */    OP_ALLANY,         /* 13 Match any character */
1904    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 */
1905    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
# Line 1309  enum { Line 1910  enum {
1910    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1911    OP_VSPACE,         /* 21 \v (vertical whitespace) */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1912    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1913    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) */
1914    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data (\z) */
1915    
1916    OP_CIRC,           /* 25 Start of line - not multiline */    OP_CIRC,           /* 25 Start of line - not multiline */
1917    OP_CIRCM,          /* 26 Start of line - multiline */    OP_CIRCM,          /* 26 Start of line - multiline */
# Line 1433  enum { Line 2034  enum {
2034    OP_CLASS,          /* 106 Match a character class, chars < 256 only */    OP_CLASS,          /* 106 Match a character class, chars < 256 only */
2035    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
2036                                class - the difference is relevant only when a                                class - the difference is relevant only when a
2037                                UTF-8 character > 255 is encountered. */                                character > 255 is encountered. */
2038    OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
2039                                class. This does both positive and negative. */                                class. This does both positive and negative. */
2040    OP_REF,            /* 109 Match a back reference, casefully */    OP_REF,            /* 109 Match a back reference, casefully */
2041    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 2192  in UTF-8 mode. The code that uses this t
2192    2,                             /* noti                                   */ \    2,                             /* noti                                   */ \
2193    /* Positive single-char repeats                             ** These are */ \    /* Positive single-char repeats                             ** These are */ \
2194    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
2195    4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
2196    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE,                   /* exact                                  */ \
2197      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
2198    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     */ \
2199    4, 4, 4,                       /* upto I, minupto I, exact I             */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
2200    2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \    2+IMM2_SIZE,                   /* exact I                                */ \
2201      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
2202    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
2203    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
2204    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
2205    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
2206      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
2207    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          */ \
2208    4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
2209    2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \    2+IMM2_SIZE,                   /* NOT exact I                            */ \
2210      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
2211    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
2212    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
2213    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
2214    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
2215      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
2216    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
2217    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
2218    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
2219   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
2220   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
2221    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
2222    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
2223    3,                             /* REFI                                   */ \    1+IMM2_SIZE,                   /* REFI                                   */ \
2224    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
2225    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
2226    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
# Line 1631  in UTF-8 mode. The code that uses this t Line 2237  in UTF-8 mode. The code that uses this t
2237    1+LINK_SIZE,                   /* ONCE_NC                                */ \    1+LINK_SIZE,                   /* ONCE_NC                                */ \
2238    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
2239    1+LINK_SIZE,                   /* BRAPOS                                 */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
2240    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
2241    3+LINK_SIZE,                   /* CBRAPOS                                */ \    1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
2242    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
2243    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
2244    1+LINK_SIZE,                   /* SBRAPOS                                */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
2245    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
2246    3+LINK_SIZE,                   /* SCBRAPOS                               */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
2247    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
2248    3, 3,                          /* CREF, NCREF                            */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
2249    3, 3,                          /* RREF, NRREF                            */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
2250    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
2251    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
2252    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
2253    1, 3,                          /* SKIP, SKIP_ARG                         */ \    1, 3,                          /* SKIP, SKIP_ARG                         */ \
2254    1, 3,                          /* THEN, THEN_ARG                         */ \    1, 3,                          /* THEN, THEN_ARG                         */ \
2255    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
2256    3, 1                           /* CLOSE, SKIPZERO  */    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
2257    
2258  /* 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"
2259  condition. */  condition. */
# Line 1665  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 2271  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
2271         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2272         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2273         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2274         ERR70, ERR71, ERRCOUNT };         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERRCOUNT };
2275    
2276    /* JIT compiling modes. The function list is indexed by them. */
2277    enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2278           JIT_NUMBER_OF_COMPILE_MODES };
2279    
2280  /* 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
2281  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
# Line 1684  fields are present. Currently PCRE alway Line 2294  fields are present. Currently PCRE alway
2294  NOTE NOTE NOTE  NOTE NOTE NOTE
2295  */  */
2296    
2297  typedef struct real_pcre {  #if defined COMPILE_PCRE8
2298    #define REAL_PCRE real_pcre
2299    #elif defined COMPILE_PCRE16
2300    #define REAL_PCRE real_pcre16
2301    #elif defined COMPILE_PCRE32
2302    #define REAL_PCRE real_pcre32
2303    #endif
2304    
2305    /* It is necessary to fork the struct for 32 bit, since it needs to use
2306     * pcre_uchar for first_char and req_char. Can't put an ifdef inside the
2307     * typedef since pcretest needs access to  the struct of the 8-, 16-
2308     * and 32-bit variants. */
2309    
2310    typedef struct real_pcre8_or_16 {
2311    pcre_uint32 magic_number;    pcre_uint32 magic_number;
2312    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2313    pcre_uint32 options;            /* Public options */    pcre_uint32 options;            /* Public options */
2314    pcre_uint16 flags;              /* Private flags */    pcre_uint16 flags;              /* Private flags */
2315    pcre_uint16 dummy1;             /* For future use */    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
2316    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;        /* Highest numbered group */
2317    pcre_uint16 top_backref;    pcre_uint16 top_backref;        /* Highest numbered back reference */
2318    pcre_uint16 first_byte;    pcre_uint16 first_char;         /* Starting character */
2319    pcre_uint16 req_byte;    pcre_uint16 req_char;           /* This character must be seen */
2320    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
2321    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
2322    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
2323    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
2324      const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
2325      const pcre_uint8 *nullpad;      /* NULL padding */
2326    } real_pcre8_or_16;
2327    
2328    const unsigned char *tables;    /* Pointer to tables or NULL for std */  typedef struct real_pcre8_or_16 real_pcre;
2329    const unsigned char *nullpad;   /* NULL padding */  typedef struct real_pcre8_or_16 real_pcre16;
2330  } real_pcre;  
2331    typedef struct real_pcre32 {
2332      pcre_uint32 magic_number;
2333      pcre_uint32 size;               /* Total that was malloced */
2334      pcre_uint32 options;            /* Public options */
2335      pcre_uint16 flags;              /* Private flags */
2336      pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
2337      pcre_uint16 top_bracket;        /* Highest numbered group */
2338      pcre_uint16 top_backref;        /* Highest numbered back reference */
2339      pcre_uint32 first_char;         /* Starting character */
2340      pcre_uint32 req_char;           /* This character must be seen */
2341      pcre_uint16 name_table_offset;  /* Offset to name table that follows */
2342      pcre_uint16 name_entry_size;    /* Size of any name items */
2343      pcre_uint16 name_count;         /* Number of name items */
2344      pcre_uint16 ref_count;          /* Reference count */
2345      pcre_uint16 dummy1;             /* for later expansion */
2346      pcre_uint16 dummy2;             /* for later expansion */
2347      const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
2348      void *nullpad;                  /* for later expansion */
2349    } real_pcre32;
2350    
2351    /* Assert that the size of REAL_PCRE is divisible by 8 */
2352    typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1];
2353    
2354    /* Needed in pcretest to access some fields in the real_pcre* structures
2355     * directly. They're unified for 8/16/32 bits since the structs only differ
2356     * after these fields; if that ever changes, need to fork those defines into
2357     * 8/16 and 32 bit versions. */
2358    #define REAL_PCRE_MAGIC(re)     (((REAL_PCRE*)re)->magic_number)
2359    #define REAL_PCRE_SIZE(re)      (((REAL_PCRE*)re)->size)
2360    #define REAL_PCRE_OPTIONS(re)   (((REAL_PCRE*)re)->options)
2361    #define REAL_PCRE_FLAGS(re)     (((REAL_PCRE*)re)->flags)
2362    
2363  /* 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
2364  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 2366  remark (see NOTE above) about extending
2366  typedef struct pcre_study_data {  typedef struct pcre_study_data {
2367    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2368    pcre_uint32 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
2369    uschar start_bits[32];          /* Starting char bits */    pcre_uint8 start_bits[32];      /* Starting char bits */
2370    pcre_uint32 minlength;          /* Minimum subject length */    pcre_uint32 minlength;          /* Minimum subject length */
2371  } pcre_study_data;  } pcre_study_data;
2372    
# Line 1728  typedef struct open_capitem { Line 2385  typedef struct open_capitem {
2385  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2386    
2387  typedef struct compile_data {  typedef struct compile_data {
2388    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
2389    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
2390    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
2391    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
2392    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
2393    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
2394    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2395    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
2396    open_capitem *open_caps;      /* Chain of open capture items */    open_capitem *open_caps;          /* Chain of open capture items */
2397    uschar *hwm;                  /* High watermark of workspace */    pcre_uchar *hwm;                  /* High watermark of workspace */
2398    uschar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
2399    int  names_found;             /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
2400    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
2401    int  bracount;                /* Count of capturing parens as we compile */    int  workspace_size;              /* Size of workspace */
2402    int  final_bracount;          /* Saved value after first pass */    unsigned int  bracount;           /* Count of capturing parens as we compile */
2403    int  top_backref;             /* Maximum back reference */    int  final_bracount;              /* Saved value after first pass */
2404    unsigned int backref_map;     /* Bitmap of low back refs */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2405    int  assert_depth;            /* Depth of nested assertions */    int  top_backref;                 /* Maximum back reference */
2406    int  external_options;        /* External (initial) options */    unsigned int backref_map;         /* Bitmap of low back refs */
2407    int  external_flags;          /* External flag bits to be set */    int  assert_depth;                /* Depth of nested assertions */
2408    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  external_options;            /* External (initial) options */
2409    BOOL had_accept;              /* (*ACCEPT) encountered */    int  external_flags;              /* External flag bits to be set */
2410    BOOL check_lookbehind;        /* Lookbehinds need later checking */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
2411    int  nltype;                  /* Newline type */    BOOL had_accept;                  /* (*ACCEPT) encountered */
2412    int  nllen;                   /* Newline string length */    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
2413    uschar nl[4];                 /* Newline string when fixed length */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
2414      int  nltype;                      /* Newline type */
2415      int  nllen;                       /* Newline string length */
2416      pcre_uchar nl[4];                 /* Newline string when fixed length */
2417  } compile_data;  } compile_data;
2418    
2419  /* 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 2421  branches, for testing for left recursion
2421    
2422  typedef struct branch_chain {  typedef struct branch_chain {
2423    struct branch_chain *outer;    struct branch_chain *outer;
2424    uschar *current_branch;    pcre_uchar *current_branch;
2425  } branch_chain;  } branch_chain;
2426    
2427  /* 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 2429  call within the pattern; used by pcre_ex
2429    
2430  typedef struct recursion_info {  typedef struct recursion_info {
2431    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2432    int group_num;                  /* Number of group that was called */    unsigned int group_num;         /* Number of group that was called */
2433    int *offset_save;               /* Pointer to start of saved offsets */    int *offset_save;               /* Pointer to start of saved offsets */
2434    int saved_max;                  /* Number of saved offsets */    int saved_max;                  /* Number of saved offsets */
2435    USPTR subject_position;         /* Position at start of recursion */    int saved_capture_last;         /* Last capture number */
2436      PCRE_PUCHAR subject_position;   /* Position at start of recursion */
2437  } recursion_info;  } recursion_info;
2438    
2439  /* A similar structure for pcre_dfa_exec(). */  /* A similar structure for pcre_dfa_exec(). */
# Line 1780  typedef struct recursion_info { Line 2441  typedef struct recursion_info {
2441  typedef struct dfa_recursion_info {  typedef struct dfa_recursion_info {
2442    struct dfa_recursion_info *prevrec;    struct dfa_recursion_info *prevrec;
2443    int group_num;    int group_num;
2444    USPTR subject_position;    PCRE_PUCHAR subject_position;
2445  } dfa_recursion_info;  } dfa_recursion_info;
2446    
2447  /* 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 2451  pcre_exec(). */
2451    
2452  typedef struct eptrblock {  typedef struct eptrblock {
2453    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
2454    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
2455  } eptrblock;  } eptrblock;
2456    
2457    
# Line 1801  typedef struct match_data { Line 2462  typedef struct match_data {
2462    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
2463    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
2464    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
2465    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
2466    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
2467    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
2468    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
2469    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
2470    int    name_count;            /* Number of names in name table */    int    name_count;              /* Number of names in name table */
2471    int    name_entry_size;       /* Size of entry in names table */    int    name_entry_size;         /* Size of entry in names table */
2472    uschar *name_table;           /* Table of names */    unsigned int skip_arg_count;    /* For counting SKIP_ARGs */
2473    uschar nl[4];                 /* Newline string when fixed */    unsigned int ignore_skip_arg;   /* For re-run when SKIP arg name not found */
2474    const  uschar *lcc;           /* Points to lower casing table */    pcre_uchar *name_table;         /* Table of names */
2475    const  uschar *ctypes;        /* Points to table of type maps */    pcre_uchar nl[4];               /* Newline string when fixed */
2476    BOOL   offset_overflow;       /* Set if too many extractions */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2477    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2478    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2479    BOOL   utf8;                  /* UTF8 flag */    BOOL   notbol;                  /* NOTBOL flag */
2480    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   noteol;                  /* NOTEOL flag */
2481    BOOL   use_ucp;               /* PCRE_UCP flag */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2482    BOOL   endonly;               /* Dollar not before final \n */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2483    BOOL   notempty;              /* Empty string match not wanted */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2484    BOOL   notempty_atstart;      /* Empty string match at start not wanted */    BOOL   endonly;                 /* Dollar not before final \n */
2485    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   notempty;                /* Empty string match not wanted */
2486    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2487    BOOL   hasthen;               /* Pattern contains (*THEN) */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2488    const  uschar *start_code;    /* For use when recursing */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2489    USPTR  start_subject;         /* Start of the subject string */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
2490    USPTR  end_subject;           /* End of the subject string */    const  pcre_uchar *start_code;  /* For use when recursing */
2491    USPTR  start_match_ptr;       /* Start of matched string */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2492    USPTR  end_match_ptr;         /* Subject position at end match */    PCRE_PUCHAR end_subject;        /* End of the subject string */
2493    USPTR  start_used_ptr;        /* Earliest consulted character */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2494    int    partial;               /* PARTIAL options */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2495    int    end_offset_top;        /* Highwater mark at end of match */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2496    int    capture_last;          /* Most recent capture number */    int    partial;                 /* PARTIAL options */
2497    int    start_offset;          /* The start offset value */    int    end_offset_top;          /* Highwater mark at end of match */
2498    int    match_function_type;   /* Set for certain special calls of MATCH() */    pcre_int32 capture_last;        /* Most recent capture number + overflow flag */
2499    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    int    start_offset;            /* The start offset value */
2500    int    eptrn;                 /* Next free eptrblock */    int    match_function_type;     /* Set for certain special calls of MATCH() */
2501    recursion_info *recursive;    /* Linked list of recursion data */    eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2502    void  *callout_data;          /* To pass back to callouts */    int    eptrn;                   /* Next free eptrblock */
2503    const  uschar *mark;          /* Mark pointer to pass back */    recursion_info *recursive;      /* Linked list of recursion data */
2504    const  uschar *once_target;   /* Where to back up to for atomic groups */    void  *callout_data;            /* To pass back to callouts */
2505      const  pcre_uchar *mark;        /* Mark pointer to pass back on success */
2506      const  pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2507      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2508    #ifdef NO_RECURSE
2509      void  *match_frames_base;       /* For remembering malloc'd frames */
2510    #endif
2511  } match_data;  } match_data;
2512    
2513  /* 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
2514  functions. */  functions. */
2515    
2516  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2517    const uschar *start_code;      /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2518    const uschar *start_subject;   /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2519    const uschar *end_subject;     /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2520    const uschar *start_used_ptr;  /* Earliest consulted character */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2521    const uschar *tables;          /* Character tables */    const pcre_uint8 *tables;         /* Character tables */
2522    int   start_offset;            /* The start offset value */    int   start_offset;               /* The start offset value */
2523    int   moptions;                /* Match options */    int   moptions;                   /* Match options */
2524    int   poptions;                /* Pattern options */    int   poptions;                   /* Pattern options */
2525    int    nltype;                 /* Newline type */    int   nltype;                     /* Newline type */
2526    int    nllen;                  /* Newline string length */    int   nllen;                      /* Newline string length */
2527    uschar nl[4];                  /* Newline string when fixed */    pcre_uchar nl[4];                 /* Newline string when fixed */
2528    void  *callout_data;           /* To pass back to callouts */    void *callout_data;               /* To pass back to callouts */
2529    dfa_recursion_info *recursive; /* Linked list of recursion data */    dfa_recursion_info *recursive;    /* Linked list of recursion data */
2530  } dfa_match_data;  } dfa_match_data;
2531    
2532  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1895  total length. */ Line 2562  total length. */
2562  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2563  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2564    
2565    /* Internal function and data prefixes. */
2566    
2567    #if defined COMPILE_PCRE8
2568    #ifndef PUBL
2569    #define PUBL(name) pcre_##name
2570    #endif
2571    #ifndef PRIV
2572    #define PRIV(name) _pcre_##name
2573    #endif
2574    #elif defined COMPILE_PCRE16
2575    #ifndef PUBL
2576    #define PUBL(name) pcre16_##name
2577    #endif
2578    #ifndef PRIV
2579    #define PRIV(name) _pcre16_##name
2580    #endif
2581    #elif defined COMPILE_PCRE32
2582    #ifndef PUBL
2583    #define PUBL(name) pcre32_##name
2584    #endif
2585    #ifndef PRIV
2586    #define PRIV(name) _pcre32_##name
2587    #endif
2588    #else
2589    #error Unsupported compiling mode
2590    #endif /* COMPILE_PCRE[8|16|32] */
2591    
2592  /* 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
2593  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
2594  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 2606  of the exported public functions. They h
2606  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
2607  pcre_tables.c module. */  pcre_tables.c module. */
2608    
2609  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
2610  extern const int    _pcre_utf8_table2[];  extern const int            PRIV(utf8_table1)[];
2611  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table1_size);
2612  extern const uschar _pcre_utf8_table4[];  extern const int            PRIV(utf8_table2)[];
2613    extern const int            PRIV(utf8_table3)[];
2614  #ifdef SUPPORT_JIT  extern const pcre_uint8     PRIV(utf8_table4)[];
2615  extern const uschar _pcre_utf8_char_sizes[];  #endif /* COMPILE_PCRE8 */
2616  #endif  
2617    extern const char           PRIV(utt_names)[];
2618    extern const ucp_type_table PRIV(utt)[];
2619    extern const int            PRIV(utt_size);
2620    
2621  extern const int    _pcre_utf8_table1_size;  extern const pcre_uint8     PRIV(OP_lengths)[];
2622    extern const pcre_uint8     PRIV(default_tables)[];
2623    
2624  extern const char   _pcre_utt_names[];  extern const pcre_uint32    PRIV(hspace_list)[];
2625  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[];  
2626    
2627    
2628  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2629  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
2630  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2631    
2632  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);  /* String comparison functions. */
2633  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);  #if defined COMPILE_PCRE8
2634  extern int           _pcre_ord2utf8(int, uschar *);  
2635  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
2636                         const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
2637  extern int           _pcre_valid_utf8(USPTR, int, int *);  #define STRCMP_UC_C8(str1, str2) \
2638  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);    strcmp((char *)(str1), (str2))
2639  extern BOOL          _pcre_xclass(int, const uschar *);  #define STRNCMP_UC_UC(str1, str2, num) \
2640      strncmp((char *)(str1), (char *)(str2), (num))
2641    #define STRNCMP_UC_C8(str1, str2, num) \
2642      strncmp((char *)(str1), (str2), (num))
2643    #define STRLEN_UC(str) strlen((const char *)str)
2644    
2645    #elif defined COMPILE_PCRE16 || defined COMPILE_PCRE32
2646    
2647    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2648                               const pcre_uchar *);
2649    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2650                               const char *);
2651    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2652                               const pcre_uchar *, unsigned int num);
2653    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2654                               const char *, unsigned int num);
2655    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2656    
2657    #define STRCMP_UC_UC(str1, str2) \
2658      PRIV(strcmp_uc_uc)((str1), (str2))
2659    #define STRCMP_UC_C8(str1, str2) \
2660      PRIV(strcmp_uc_c8)((str1), (str2))
2661    #define STRNCMP_UC_UC(str1, str2, num) \
2662      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2663    #define STRNCMP_UC_C8(str1, str2, num) \
2664      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2665    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2666    
2667    #endif /* COMPILE_PCRE[8|16|32] */
2668    
2669    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
2670    
2671    #define STRCMP_UC_UC_TEST(str1, str2) STRCMP_UC_UC(str1, str2)
2672    #define STRCMP_UC_C8_TEST(str1, str2) STRCMP_UC_C8(str1, str2)
2673    
2674    #elif defined COMPILE_PCRE32
2675    
2676    extern int               PRIV(strcmp_uc_uc_utf)(const pcre_uchar *,
2677                               const pcre_uchar *);
2678    extern int               PRIV(strcmp_uc_c8_utf)(const pcre_uchar *,
2679                               const char *);
2680    
2681    #define STRCMP_UC_UC_TEST(str1, str2) \
2682      (utf ? PRIV(strcmp_uc_uc_utf)((str1), (str2)) : PRIV(strcmp_uc_uc)((str1), (str2)))
2683    #define STRCMP_UC_C8_TEST(str1, str2) \
2684      (utf ? PRIV(strcmp_uc_c8_utf)((str1), (str2)) : PRIV(strcmp_uc_c8)((str1), (str2)))
2685    
2686    #endif /* COMPILE_PCRE[8|16|32] */
2687    
2688    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2689    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2690                               int *, BOOL);
2691    extern unsigned int      PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2692    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2693    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2694                               int *, BOOL);
2695    extern BOOL              PRIV(xclass)(pcre_uint32, const pcre_uchar *, BOOL);
2696    
2697  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
2698  extern void          _pcre_jit_compile(const real_pcre *, pcre_extra *);  extern void              PRIV(jit_compile)(const REAL_PCRE *,
2699  extern int           _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR,                             PUBL(extra) *, int);
2700                          int, int, int, int, int *, int);  extern int               PRIV(jit_exec)(const PUBL(extra) *,
2701  extern void          _pcre_jit_free(void *);                             const pcre_uchar *, int, int, int, int *, int);
2702    extern void              PRIV(jit_free)(void *);
2703    extern int               PRIV(jit_get_size)(void *);
2704    extern const char*       PRIV(jit_get_target)(void);
2705  #endif  #endif
2706    
2707  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
2708    
2709  typedef struct {  typedef struct {
2710    uschar script;    pcre_uint8 script;     /* ucp_Arabic, etc. */
2711    uschar chartype;    pcre_uint8 chartype;   /* ucp_Cc, etc. (general categories) */
2712    pcre_int32 other_case;    pcre_uint8 gbprop;     /* ucp_gbControl, etc. (grapheme break property) */
2713      pcre_uint8 caseset;    /* offset to multichar other cases or zero */
2714      pcre_int32 other_case; /* offset to other case, or zero if none */
2715  } ucd_record;  } ucd_record;
2716    
2717  extern const ucd_record  _pcre_ucd_records[];  extern const pcre_uint32 PRIV(ucd_caseless_sets)[];
2718  extern const uschar      _pcre_ucd_stage1[];  extern const ucd_record  PRIV(ucd_records)[];
2719  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint8  PRIV(ucd_stage1)[];
2720  extern const int         _pcre_ucp_gentype[];  extern const pcre_uint16 PRIV(ucd_stage2)[];
2721    extern const pcre_uint32 PRIV(ucp_gentype)[];
2722    extern const pcre_uint32 PRIV(ucp_gbtable)[];
2723  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
2724  extern const int         _pcre_ucp_typerange[];  extern const int         PRIV(ucp_typerange)[];
2725  #endif  #endif
2726    
2727    #ifdef SUPPORT_UCP
2728  /* UCD access macros */  /* UCD access macros */
2729    
2730  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
2731  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (PRIV(ucd_records) + \
2732          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          PRIV(ucd_stage2)[PRIV(ucd_stage1)[(int)(ch) / UCD_BLOCK_SIZE] * \
2733          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (int)(ch) % UCD_BLOCK_SIZE])
2734    
2735  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)    GET_UCD(ch)->chartype
2736  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)      GET_UCD(ch)->script
2737  #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]  #define UCD_CATEGORY(ch)    PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2738  #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)  #define UCD_GRAPHBREAK(ch)  GET_UCD(ch)->gbprop
2739    #define UCD_CASESET(ch)     GET_UCD(ch)->caseset
2740    #define UCD_OTHERCASE(ch)   ((pcre_uint32)((int)ch + (int)(GET_UCD(ch)->other_case)))
2741    
2742    #endif /* SUPPORT_UCP */
2743    
2744  #endif  #endif
2745    

Legend:
Removed from v.758  
changed lines
  Added in v.1309

  ViewVC Help
Powered by ViewVC 1.1.5