/[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 190 by ph10, Thu Jul 19 10:38:20 2007 UTC revision 573 by ph10, Fri Nov 19 19:29:44 2010 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) 1997-2007 University of Cambridge             Copyright (c) 1997-2010 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 45  functions whose names all begin with "_p Line 45  functions whose names all begin with "_p
45  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
46  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
47    
48  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
49    
50  #if 0  #if 0
51  #define DEBUG  #define PCRE_DEBUG
52    #endif
53    
54    /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
55    script prevents both being selected, but not everybody uses "configure". */
56    
57    #if defined EBCDIC && defined SUPPORT_UTF8
58    #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
59    #endif
60    
61    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
62    "configure" script ensures this, but not everybody uses "configure". */
63    
64    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
65    #define SUPPORT_UTF8 1
66  #endif  #endif
67    
68  /* 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
# Line 60  It turns out that the Mac Debugging.h he Line 74  It turns out that the Mac Debugging.h he
74  be absolutely sure we get our version. */  be absolutely sure we get our version. */
75    
76  #undef DPRINTF  #undef DPRINTF
77  #ifdef DEBUG  #ifdef PCRE_DEBUG
78  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
79  #else  #else
80  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
81  #endif  #endif
82    
83    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
84  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
85  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
86    
87  #include <ctype.h>  #include <ctype.h>
88  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
89  #include <stddef.h>  #include <stddef.h>
90  #include <stdio.h>  #include <stdio.h>
91  #include <stdlib.h>  #include <stdlib.h>
# Line 112  PCRE_EXP_DATA_DEFN only if they are not Line 120  PCRE_EXP_DATA_DEFN only if they are not
120    
121  #ifndef PCRE_EXP_DECL  #ifndef PCRE_EXP_DECL
122  #  ifdef _WIN32  #  ifdef _WIN32
123  #    ifdef DLL_EXPORT  #    ifndef PCRE_STATIC
124  #      define PCRE_EXP_DECL       extern __declspec(dllexport)  #      define PCRE_EXP_DECL       extern __declspec(dllexport)
125  #      define PCRE_EXP_DEFN       __declspec(dllexport)  #      define PCRE_EXP_DEFN       __declspec(dllexport)
126  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
# Line 121  PCRE_EXP_DATA_DEFN only if they are not Line 129  PCRE_EXP_DATA_DEFN only if they are not
129  #      define PCRE_EXP_DEFN  #      define PCRE_EXP_DEFN
130  #      define PCRE_EXP_DATA_DEFN  #      define PCRE_EXP_DATA_DEFN
131  #    endif  #    endif
 #  
132  #  else  #  else
133  #    ifdef __cplusplus  #    ifdef __cplusplus
134  #      define PCRE_EXP_DECL       extern "C"  #      define PCRE_EXP_DECL       extern "C"
# Line 137  PCRE_EXP_DATA_DEFN only if they are not Line 144  PCRE_EXP_DATA_DEFN only if they are not
144  #  endif  #  endif
145  #endif  #endif
146    
147    /* When compiling with the MSVC compiler, it is sometimes necessary to include
148    a "calling convention" before exported function names. (This is secondhand
149    information; I know nothing about MSVC myself). For example, something like
150    
151      void __cdecl function(....)
152    
153    might be needed. In order so make this easy, all the exported functions have
154    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
155    set, we ensure here that it has no effect. */
156    
157    #ifndef PCRE_CALL_CONVENTION
158    #define PCRE_CALL_CONVENTION
159    #endif
160    
161  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
162  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
163  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
# Line 145  preprocessor time in standard C environm Line 166  preprocessor time in standard C environm
166    
167  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
168    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
169      typedef short pcre_int16;
170  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
171    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
172      typedef int pcre_int16;
173  #else  #else
174    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
175  #endif  #endif
176    
177  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
178    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
179      typedef int pcre_int32;
180  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
181    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
182      typedef long int pcre_int32;
183  #else  #else
184    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
185  #endif  #endif
186    
187    /* When checking for integer overflow in pcre_compile(), we need to handle
188    large integers. If a 64-bit integer type is available, we can use that.
189    Otherwise we have to cast to double, which of course requires floating point
190    arithmetic. Handle this by defining a macro for the appropriate type. If
191    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
192    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
193    by "configure". */
194    
195    #if HAVE_STDINT_H
196    #include <stdint.h>
197    #elif HAVE_INTTYPES_H
198    #include <inttypes.h>
199    #endif
200    
201    #if defined INT64_MAX || defined int64_t
202    #define INT64_OR_DOUBLE int64_t
203    #else
204    #define INT64_OR_DOUBLE double
205    #endif
206    
207  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
208  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
209  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char *, because that
# Line 363  capturing parenthesis numbers in back re Line 408  capturing parenthesis numbers in back re
408    
409  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
410  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
411  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
412    not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should
413    never be called in byte mode. To make sure they can never even appear when
414    UTF-8 support is omitted, we don't even define them. */
415    
416  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
417  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 371  byte-mode, and more complicated ones for Line 419  byte-mode, and more complicated ones for
419  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
420  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
421  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
422  #define BACKCHAR(eptr)  /* #define GETCHARLENTEST(c, eptr, len) */
423    /* #define BACKCHAR(eptr) */
424    
425  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
426    
427    /* These macros were originally written in the form of loops that used data
428    from the tables whose names start with _pcre_utf8_table. They were rewritten by
429    a user so as not to use loops, because in some environments this gives a
430    significant performance advantage, and it seems never to do any harm. */
431    
432    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
433    advancing the pointer. */
434    
435    #define GETUTF8(c, eptr) \
436        { \
437        if ((c & 0x20) == 0) \
438          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
439        else if ((c & 0x10) == 0) \
440          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
441        else if ((c & 0x08) == 0) \
442          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
443          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
444        else if ((c & 0x04) == 0) \
445          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
446              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
447              (eptr[4] & 0x3f); \
448        else \
449          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
450              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
451              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
452        }
453    
454  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
455  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
456    
457  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
458    c = *eptr; \    c = *eptr; \
459    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8(c, eptr);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     }  
460    
461  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
462  pointer. */  pointer. */
463    
464  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
465    c = *eptr; \    c = *eptr; \
466    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
467    
468    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
469    the pointer. */
470    
471    #define GETUTF8INC(c, eptr) \
472      { \      { \
473      int gcii; \      if ((c & 0x20) == 0) \
474      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
475      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
476      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
477      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
478          eptr += 2; \
479          } \
480        else if ((c & 0x08) == 0) \
481          { \
482          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
483              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
484          eptr += 3; \
485          } \
486        else if ((c & 0x04) == 0) \
487          { \
488          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
489              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
490              (eptr[3] & 0x3f); \
491          eptr += 4; \
492          } \
493        else \
494        { \        { \
495        gcss -= 6; \        c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
496        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
497              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
498          eptr += 5; \
499        } \        } \
500      }      }
501    
# Line 416  know we are in UTF-8 mode. */ Line 504  know we are in UTF-8 mode. */
504    
505  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
506    c = *eptr++; \    c = *eptr++; \
507    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8INC(c, eptr);
     { \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
       { \  
       gcss -= 6; \  
       c |= (*eptr++ & 0x3f) << gcss; \  
       } \  
     }  
508    
509  /* Get the next character, testing for UTF-8 mode, and advancing the pointer */  /* Get the next character, testing for UTF-8 mode, and advancing the pointer.
510    This is called when we don't know if we are in UTF-8 mode. */
511    
512  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
513    c = *eptr++; \    c = *eptr++; \
514    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
515    
516    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
517    advancing the pointer, incrementing the length. */
518    
519    #define GETUTF8LEN(c, eptr, len) \
520      { \      { \
521      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
522      int gcss = 6*gcaa; \        { \
523      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
524      while (gcaa-- > 0) \        len++; \
525          } \
526        else if ((c & 0x10)  == 0) \
527          { \
528          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
529          len += 2; \
530          } \
531        else if ((c & 0x08)  == 0) \
532          {\
533          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
534              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
535          len += 3; \
536          } \
537        else if ((c & 0x04)  == 0) \
538        { \        { \
539        gcss -= 6; \        c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
540        c |= (*eptr++ & 0x3f) << gcss; \            ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
541              (eptr[4] & 0x3f); \
542          len += 4; \
543          } \
544        else \
545          {\
546          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
547              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
548              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
549          len += 5; \
550        } \        } \
551      }      }
552    
# Line 449  if there are extra bytes. This is called Line 555  if there are extra bytes. This is called
555    
556  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
557    c = *eptr; \    c = *eptr; \
558    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
559      { \  
560      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
561      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
562      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
563      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
564      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
565        { \    c = *eptr; \
566        gcss -= 6; \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
567    
568  /* 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
569  it is. Called only in UTF-8 mode. */  it is. This is called only in UTF-8 mode - we don't put a test within the macro
570    because almost all calls are already within a block of UTF-8 only code. */
571    
572  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
573    
574  #endif  #endif  /* SUPPORT_UTF8 */
575    
576    
577  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
# Line 483  Standard C system should have one. */ Line 586  Standard C system should have one. */
586    
587  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
588    
589  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
590  The public options defined in pcre.h start at the least significant end. Make  live at the top end of the options word, but that got almost full, so now they
591  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
592  out, there is a dummy word in the structure that could be used for the private  the restrictions on partial matching have been lifted. It remains for backwards
593  bits. */  compatibility. */
594    
595  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
596  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
597  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
598  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
599  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
600    #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
601    
602  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
603    
604  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
605    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
606    
607  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
608  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 505  time, run time, or study time, respectiv Line 610  time, run time, or study time, respectiv
610  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
611                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
612    
613  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
614    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
615     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
616     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| \
617     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
618       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP)
619    
620  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
621    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
622     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
623       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
624    
625  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
626    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
627     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
628       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
629       PCRE_NO_START_OPTIMIZE)
630    
631  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
632    
# Line 542  variable-length repeat, or a anything ot Line 651  variable-length repeat, or a anything ot
651  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
652  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
653    
654  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
655    environments where these macros are defined elsewhere. Unfortunately, there
656    is no way to do the same for the typedef. */
657    
658  typedef int BOOL;  typedef int BOOL;
659    
660    #ifndef FALSE
661  #define FALSE   0  #define FALSE   0
662  #define TRUE    1  #define TRUE    1
663    #endif
664    
665    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
666    character constants like '*' because the compiler would emit their EBCDIC code,
667    which is different from their ASCII/UTF-8 code. Instead we define macros for
668    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
669    is enabled. When UTF-8 support is not enabled, the definitions use character
670    literals. Both character and string versions of each character are needed, and
671    there are some longer strings as well.
672    
673    This means that, on EBCDIC platforms, the PCRE library can handle either
674    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
675    would need different lookups depending on whether PCRE_UTF8 was set or not.
676    This would make it impossible to use characters in switch/case statements,
677    which would reduce performance. For a theoretical use (which nobody has asked
678    for) in a minority area (EBCDIC platforms), this is not sensible. Any
679    application that did need both could compile two versions of the library, using
680    macros to give the functions distinct names. */
681    
682    #ifndef SUPPORT_UTF8
683    
684    /* UTF-8 support is not enabled; use the platform-dependent character literals
685    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
686    
687    #define CHAR_HT                     '\t'
688    #define CHAR_VT                     '\v'
689    #define CHAR_FF                     '\f'
690    #define CHAR_CR                     '\r'
691    #define CHAR_NL                     '\n'
692    #define CHAR_BS                     '\b'
693    #define CHAR_BEL                    '\a'
694    #ifdef EBCDIC
695    #define CHAR_ESC                    '\047'
696    #define CHAR_DEL                    '\007'
697    #else
698    #define CHAR_ESC                    '\033'
699    #define CHAR_DEL                    '\177'
700    #endif
701    
702    #define CHAR_SPACE                  ' '
703    #define CHAR_EXCLAMATION_MARK       '!'
704    #define CHAR_QUOTATION_MARK         '"'
705    #define CHAR_NUMBER_SIGN            '#'
706    #define CHAR_DOLLAR_SIGN            '$'
707    #define CHAR_PERCENT_SIGN           '%'
708    #define CHAR_AMPERSAND              '&'
709    #define CHAR_APOSTROPHE             '\''
710    #define CHAR_LEFT_PARENTHESIS       '('
711    #define CHAR_RIGHT_PARENTHESIS      ')'
712    #define CHAR_ASTERISK               '*'
713    #define CHAR_PLUS                   '+'
714    #define CHAR_COMMA                  ','
715    #define CHAR_MINUS                  '-'
716    #define CHAR_DOT                    '.'
717    #define CHAR_SLASH                  '/'
718    #define CHAR_0                      '0'
719    #define CHAR_1                      '1'
720    #define CHAR_2                      '2'
721    #define CHAR_3                      '3'
722    #define CHAR_4                      '4'
723    #define CHAR_5                      '5'
724    #define CHAR_6                      '6'
725    #define CHAR_7                      '7'
726    #define CHAR_8                      '8'
727    #define CHAR_9                      '9'
728    #define CHAR_COLON                  ':'
729    #define CHAR_SEMICOLON              ';'
730    #define CHAR_LESS_THAN_SIGN         '<'
731    #define CHAR_EQUALS_SIGN            '='
732    #define CHAR_GREATER_THAN_SIGN      '>'
733    #define CHAR_QUESTION_MARK          '?'
734    #define CHAR_COMMERCIAL_AT          '@'
735    #define CHAR_A                      'A'
736    #define CHAR_B                      'B'
737    #define CHAR_C                      'C'
738    #define CHAR_D                      'D'
739    #define CHAR_E                      'E'
740    #define CHAR_F                      'F'
741    #define CHAR_G                      'G'
742    #define CHAR_H                      'H'
743    #define CHAR_I                      'I'
744    #define CHAR_J                      'J'
745    #define CHAR_K                      'K'
746    #define CHAR_L                      'L'
747    #define CHAR_M                      'M'
748    #define CHAR_N                      'N'
749    #define CHAR_O                      'O'
750    #define CHAR_P                      'P'
751    #define CHAR_Q                      'Q'
752    #define CHAR_R                      'R'
753    #define CHAR_S                      'S'
754    #define CHAR_T                      'T'
755    #define CHAR_U                      'U'
756    #define CHAR_V                      'V'
757    #define CHAR_W                      'W'
758    #define CHAR_X                      'X'
759    #define CHAR_Y                      'Y'
760    #define CHAR_Z                      'Z'
761    #define CHAR_LEFT_SQUARE_BRACKET    '['
762    #define CHAR_BACKSLASH              '\\'
763    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
764    #define CHAR_CIRCUMFLEX_ACCENT      '^'
765    #define CHAR_UNDERSCORE             '_'
766    #define CHAR_GRAVE_ACCENT           '`'
767    #define CHAR_a                      'a'
768    #define CHAR_b                      'b'
769    #define CHAR_c                      'c'
770    #define CHAR_d                      'd'
771    #define CHAR_e                      'e'
772    #define CHAR_f                      'f'
773    #define CHAR_g                      'g'
774    #define CHAR_h                      'h'
775    #define CHAR_i                      'i'
776    #define CHAR_j                      'j'
777    #define CHAR_k                      'k'
778    #define CHAR_l                      'l'
779    #define CHAR_m                      'm'
780    #define CHAR_n                      'n'
781    #define CHAR_o                      'o'
782    #define CHAR_p                      'p'
783    #define CHAR_q                      'q'
784    #define CHAR_r                      'r'
785    #define CHAR_s                      's'
786    #define CHAR_t                      't'
787    #define CHAR_u                      'u'
788    #define CHAR_v                      'v'
789    #define CHAR_w                      'w'
790    #define CHAR_x                      'x'
791    #define CHAR_y                      'y'
792    #define CHAR_z                      'z'
793    #define CHAR_LEFT_CURLY_BRACKET     '{'
794    #define CHAR_VERTICAL_LINE          '|'
795    #define CHAR_RIGHT_CURLY_BRACKET    '}'
796    #define CHAR_TILDE                  '~'
797    
798    #define STR_HT                      "\t"
799    #define STR_VT                      "\v"
800    #define STR_FF                      "\f"
801    #define STR_CR                      "\r"
802    #define STR_NL                      "\n"
803    #define STR_BS                      "\b"
804    #define STR_BEL                     "\a"
805    #ifdef EBCDIC
806    #define STR_ESC                     "\047"
807    #define STR_DEL                     "\007"
808    #else
809    #define STR_ESC                     "\033"
810    #define STR_DEL                     "\177"
811    #endif
812    
813    #define STR_SPACE                   " "
814    #define STR_EXCLAMATION_MARK        "!"
815    #define STR_QUOTATION_MARK          "\""
816    #define STR_NUMBER_SIGN             "#"
817    #define STR_DOLLAR_SIGN             "$"
818    #define STR_PERCENT_SIGN            "%"
819    #define STR_AMPERSAND               "&"
820    #define STR_APOSTROPHE              "'"
821    #define STR_LEFT_PARENTHESIS        "("
822    #define STR_RIGHT_PARENTHESIS       ")"
823    #define STR_ASTERISK                "*"
824    #define STR_PLUS                    "+"
825    #define STR_COMMA                   ","
826    #define STR_MINUS                   "-"
827    #define STR_DOT                     "."
828    #define STR_SLASH                   "/"
829    #define STR_0                       "0"
830    #define STR_1                       "1"
831    #define STR_2                       "2"
832    #define STR_3                       "3"
833    #define STR_4                       "4"
834    #define STR_5                       "5"
835    #define STR_6                       "6"
836    #define STR_7                       "7"
837    #define STR_8                       "8"
838    #define STR_9                       "9"
839    #define STR_COLON                   ":"
840    #define STR_SEMICOLON               ";"
841    #define STR_LESS_THAN_SIGN          "<"
842    #define STR_EQUALS_SIGN             "="
843    #define STR_GREATER_THAN_SIGN       ">"
844    #define STR_QUESTION_MARK           "?"
845    #define STR_COMMERCIAL_AT           "@"
846    #define STR_A                       "A"
847    #define STR_B                       "B"
848    #define STR_C                       "C"
849    #define STR_D                       "D"
850    #define STR_E                       "E"
851    #define STR_F                       "F"
852    #define STR_G                       "G"
853    #define STR_H                       "H"
854    #define STR_I                       "I"
855    #define STR_J                       "J"
856    #define STR_K                       "K"
857    #define STR_L                       "L"
858    #define STR_M                       "M"
859    #define STR_N                       "N"
860    #define STR_O                       "O"
861    #define STR_P                       "P"
862    #define STR_Q                       "Q"
863    #define STR_R                       "R"
864    #define STR_S                       "S"
865    #define STR_T                       "T"
866    #define STR_U                       "U"
867    #define STR_V                       "V"
868    #define STR_W                       "W"
869    #define STR_X                       "X"
870    #define STR_Y                       "Y"
871    #define STR_Z                       "Z"
872    #define STR_LEFT_SQUARE_BRACKET     "["
873    #define STR_BACKSLASH               "\\"
874    #define STR_RIGHT_SQUARE_BRACKET    "]"
875    #define STR_CIRCUMFLEX_ACCENT       "^"
876    #define STR_UNDERSCORE              "_"
877    #define STR_GRAVE_ACCENT            "`"
878    #define STR_a                       "a"
879    #define STR_b                       "b"
880    #define STR_c                       "c"
881    #define STR_d                       "d"
882    #define STR_e                       "e"
883    #define STR_f                       "f"
884    #define STR_g                       "g"
885    #define STR_h                       "h"
886    #define STR_i                       "i"
887    #define STR_j                       "j"
888    #define STR_k                       "k"
889    #define STR_l                       "l"
890    #define STR_m                       "m"
891    #define STR_n                       "n"
892    #define STR_o                       "o"
893    #define STR_p                       "p"
894    #define STR_q                       "q"
895    #define STR_r                       "r"
896    #define STR_s                       "s"
897    #define STR_t                       "t"
898    #define STR_u                       "u"
899    #define STR_v                       "v"
900    #define STR_w                       "w"
901    #define STR_x                       "x"
902    #define STR_y                       "y"
903    #define STR_z                       "z"
904    #define STR_LEFT_CURLY_BRACKET      "{"
905    #define STR_VERTICAL_LINE           "|"
906    #define STR_RIGHT_CURLY_BRACKET     "}"
907    #define STR_TILDE                   "~"
908    
909    #define STRING_ACCEPT0              "ACCEPT\0"
910    #define STRING_COMMIT0              "COMMIT\0"
911    #define STRING_F0                   "F\0"
912    #define STRING_FAIL0                "FAIL\0"
913    #define STRING_MARK0                "MARK\0"
914    #define STRING_PRUNE0               "PRUNE\0"
915    #define STRING_SKIP0                "SKIP\0"
916    #define STRING_THEN                 "THEN"
917    
918    #define STRING_alpha0               "alpha\0"
919    #define STRING_lower0               "lower\0"
920    #define STRING_upper0               "upper\0"
921    #define STRING_alnum0               "alnum\0"
922    #define STRING_ascii0               "ascii\0"
923    #define STRING_blank0               "blank\0"
924    #define STRING_cntrl0               "cntrl\0"
925    #define STRING_digit0               "digit\0"
926    #define STRING_graph0               "graph\0"
927    #define STRING_print0               "print\0"
928    #define STRING_punct0               "punct\0"
929    #define STRING_space0               "space\0"
930    #define STRING_word0                "word\0"
931    #define STRING_xdigit               "xdigit"
932    
933    #define STRING_DEFINE               "DEFINE"
934    
935    #define STRING_CR_RIGHTPAR          "CR)"
936    #define STRING_LF_RIGHTPAR          "LF)"
937    #define STRING_CRLF_RIGHTPAR        "CRLF)"
938    #define STRING_ANY_RIGHTPAR         "ANY)"
939    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
940    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
941    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
942    #define STRING_UTF8_RIGHTPAR        "UTF8)"
943    #define STRING_UCP_RIGHTPAR         "UCP)"
944    
945    #else  /* SUPPORT_UTF8 */
946    
947    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
948    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
949    only. */
950    
951    #define CHAR_HT                     '\011'
952    #define CHAR_VT                     '\013'
953    #define CHAR_FF                     '\014'
954    #define CHAR_CR                     '\015'
955    #define CHAR_NL                     '\012'
956    #define CHAR_BS                     '\010'
957    #define CHAR_BEL                    '\007'
958    #define CHAR_ESC                    '\033'
959    #define CHAR_DEL                    '\177'
960    
961    #define CHAR_SPACE                  '\040'
962    #define CHAR_EXCLAMATION_MARK       '\041'
963    #define CHAR_QUOTATION_MARK         '\042'
964    #define CHAR_NUMBER_SIGN            '\043'
965    #define CHAR_DOLLAR_SIGN            '\044'
966    #define CHAR_PERCENT_SIGN           '\045'
967    #define CHAR_AMPERSAND              '\046'
968    #define CHAR_APOSTROPHE             '\047'
969    #define CHAR_LEFT_PARENTHESIS       '\050'
970    #define CHAR_RIGHT_PARENTHESIS      '\051'
971    #define CHAR_ASTERISK               '\052'
972    #define CHAR_PLUS                   '\053'
973    #define CHAR_COMMA                  '\054'
974    #define CHAR_MINUS                  '\055'
975    #define CHAR_DOT                    '\056'
976    #define CHAR_SLASH                  '\057'
977    #define CHAR_0                      '\060'
978    #define CHAR_1                      '\061'
979    #define CHAR_2                      '\062'
980    #define CHAR_3                      '\063'
981    #define CHAR_4                      '\064'
982    #define CHAR_5                      '\065'
983    #define CHAR_6                      '\066'
984    #define CHAR_7                      '\067'
985    #define CHAR_8                      '\070'
986    #define CHAR_9                      '\071'
987    #define CHAR_COLON                  '\072'
988    #define CHAR_SEMICOLON              '\073'
989    #define CHAR_LESS_THAN_SIGN         '\074'
990    #define CHAR_EQUALS_SIGN            '\075'
991    #define CHAR_GREATER_THAN_SIGN      '\076'
992    #define CHAR_QUESTION_MARK          '\077'
993    #define CHAR_COMMERCIAL_AT          '\100'
994    #define CHAR_A                      '\101'
995    #define CHAR_B                      '\102'
996    #define CHAR_C                      '\103'
997    #define CHAR_D                      '\104'
998    #define CHAR_E                      '\105'
999    #define CHAR_F                      '\106'
1000    #define CHAR_G                      '\107'
1001    #define CHAR_H                      '\110'
1002    #define CHAR_I                      '\111'
1003    #define CHAR_J                      '\112'
1004    #define CHAR_K                      '\113'
1005    #define CHAR_L                      '\114'
1006    #define CHAR_M                      '\115'
1007    #define CHAR_N                      '\116'
1008    #define CHAR_O                      '\117'
1009    #define CHAR_P                      '\120'
1010    #define CHAR_Q                      '\121'
1011    #define CHAR_R                      '\122'
1012    #define CHAR_S                      '\123'
1013    #define CHAR_T                      '\124'
1014    #define CHAR_U                      '\125'
1015    #define CHAR_V                      '\126'
1016    #define CHAR_W                      '\127'
1017    #define CHAR_X                      '\130'
1018    #define CHAR_Y                      '\131'
1019    #define CHAR_Z                      '\132'
1020    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1021    #define CHAR_BACKSLASH              '\134'
1022    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1023    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1024    #define CHAR_UNDERSCORE             '\137'
1025    #define CHAR_GRAVE_ACCENT           '\140'
1026    #define CHAR_a                      '\141'
1027    #define CHAR_b                      '\142'
1028    #define CHAR_c                      '\143'
1029    #define CHAR_d                      '\144'
1030    #define CHAR_e                      '\145'
1031    #define CHAR_f                      '\146'
1032    #define CHAR_g                      '\147'
1033    #define CHAR_h                      '\150'
1034    #define CHAR_i                      '\151'
1035    #define CHAR_j                      '\152'
1036    #define CHAR_k                      '\153'
1037    #define CHAR_l                      '\154'
1038    #define CHAR_m                      '\155'
1039    #define CHAR_n                      '\156'
1040    #define CHAR_o                      '\157'
1041    #define CHAR_p                      '\160'
1042    #define CHAR_q                      '\161'
1043    #define CHAR_r                      '\162'
1044    #define CHAR_s                      '\163'
1045    #define CHAR_t                      '\164'
1046    #define CHAR_u                      '\165'
1047    #define CHAR_v                      '\166'
1048    #define CHAR_w                      '\167'
1049    #define CHAR_x                      '\170'
1050    #define CHAR_y                      '\171'
1051    #define CHAR_z                      '\172'
1052    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1053    #define CHAR_VERTICAL_LINE          '\174'
1054    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1055    #define CHAR_TILDE                  '\176'
1056    
1057    #define STR_HT                      "\011"
1058    #define STR_VT                      "\013"
1059    #define STR_FF                      "\014"
1060    #define STR_CR                      "\015"
1061    #define STR_NL                      "\012"
1062    #define STR_BS                      "\010"
1063    #define STR_BEL                     "\007"
1064    #define STR_ESC                     "\033"
1065    #define STR_DEL                     "\177"
1066    
1067    #define STR_SPACE                   "\040"
1068    #define STR_EXCLAMATION_MARK        "\041"
1069    #define STR_QUOTATION_MARK          "\042"
1070    #define STR_NUMBER_SIGN             "\043"
1071    #define STR_DOLLAR_SIGN             "\044"
1072    #define STR_PERCENT_SIGN            "\045"
1073    #define STR_AMPERSAND               "\046"
1074    #define STR_APOSTROPHE              "\047"
1075    #define STR_LEFT_PARENTHESIS        "\050"
1076    #define STR_RIGHT_PARENTHESIS       "\051"
1077    #define STR_ASTERISK                "\052"
1078    #define STR_PLUS                    "\053"
1079    #define STR_COMMA                   "\054"
1080    #define STR_MINUS                   "\055"
1081    #define STR_DOT                     "\056"
1082    #define STR_SLASH                   "\057"
1083    #define STR_0                       "\060"
1084    #define STR_1                       "\061"
1085    #define STR_2                       "\062"
1086    #define STR_3                       "\063"
1087    #define STR_4                       "\064"
1088    #define STR_5                       "\065"
1089    #define STR_6                       "\066"
1090    #define STR_7                       "\067"
1091    #define STR_8                       "\070"
1092    #define STR_9                       "\071"
1093    #define STR_COLON                   "\072"
1094    #define STR_SEMICOLON               "\073"
1095    #define STR_LESS_THAN_SIGN          "\074"
1096    #define STR_EQUALS_SIGN             "\075"
1097    #define STR_GREATER_THAN_SIGN       "\076"
1098    #define STR_QUESTION_MARK           "\077"
1099    #define STR_COMMERCIAL_AT           "\100"
1100    #define STR_A                       "\101"
1101    #define STR_B                       "\102"
1102    #define STR_C                       "\103"
1103    #define STR_D                       "\104"
1104    #define STR_E                       "\105"
1105    #define STR_F                       "\106"
1106    #define STR_G                       "\107"
1107    #define STR_H                       "\110"
1108    #define STR_I                       "\111"
1109    #define STR_J                       "\112"
1110    #define STR_K                       "\113"
1111    #define STR_L                       "\114"
1112    #define STR_M                       "\115"
1113    #define STR_N                       "\116"
1114    #define STR_O                       "\117"
1115    #define STR_P                       "\120"
1116    #define STR_Q                       "\121"
1117    #define STR_R                       "\122"
1118    #define STR_S                       "\123"
1119    #define STR_T                       "\124"
1120    #define STR_U                       "\125"
1121    #define STR_V                       "\126"
1122    #define STR_W                       "\127"
1123    #define STR_X                       "\130"
1124    #define STR_Y                       "\131"
1125    #define STR_Z                       "\132"
1126    #define STR_LEFT_SQUARE_BRACKET     "\133"
1127    #define STR_BACKSLASH               "\134"
1128    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1129    #define STR_CIRCUMFLEX_ACCENT       "\136"
1130    #define STR_UNDERSCORE              "\137"
1131    #define STR_GRAVE_ACCENT            "\140"
1132    #define STR_a                       "\141"
1133    #define STR_b                       "\142"
1134    #define STR_c                       "\143"
1135    #define STR_d                       "\144"
1136    #define STR_e                       "\145"
1137    #define STR_f                       "\146"
1138    #define STR_g                       "\147"
1139    #define STR_h                       "\150"
1140    #define STR_i                       "\151"
1141    #define STR_j                       "\152"
1142    #define STR_k                       "\153"
1143    #define STR_l                       "\154"
1144    #define STR_m                       "\155"
1145    #define STR_n                       "\156"
1146    #define STR_o                       "\157"
1147    #define STR_p                       "\160"
1148    #define STR_q                       "\161"
1149    #define STR_r                       "\162"
1150    #define STR_s                       "\163"
1151    #define STR_t                       "\164"
1152    #define STR_u                       "\165"
1153    #define STR_v                       "\166"
1154    #define STR_w                       "\167"
1155    #define STR_x                       "\170"
1156    #define STR_y                       "\171"
1157    #define STR_z                       "\172"
1158    #define STR_LEFT_CURLY_BRACKET      "\173"
1159    #define STR_VERTICAL_LINE           "\174"
1160    #define STR_RIGHT_CURLY_BRACKET     "\175"
1161    #define STR_TILDE                   "\176"
1162    
1163    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1164    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1165    #define STRING_F0                   STR_F "\0"
1166    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1167    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1168    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1169    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1170    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1171    
1172    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1173    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1174    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1175    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1176    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1177    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1178    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1179    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1180    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1181    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1182    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1183    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1184    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1185    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1186    
1187    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1188    
1189    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1190    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1191    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1192    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1193    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1194    #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
1195    #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
1196    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1197    #define STRING_UCP_RIGHTPAR         STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1198    
1199    #endif  /* SUPPORT_UTF8 */
1200    
1201  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1202    
1203  #ifndef ESC_e  #ifndef ESC_e
1204  #define ESC_e 27  #define ESC_e CHAR_ESC
1205  #endif  #endif
1206    
1207  #ifndef ESC_f  #ifndef ESC_f
1208  #define ESC_f '\f'  #define ESC_f CHAR_FF
1209  #endif  #endif
1210    
1211  #ifndef ESC_n  #ifndef ESC_n
1212  #define ESC_n '\n'  #define ESC_n CHAR_NL
1213  #endif  #endif
1214    
1215  #ifndef ESC_r  #ifndef ESC_r
1216  #define ESC_r '\r'  #define ESC_r CHAR_CR
1217  #endif  #endif
1218    
1219  /* We can't officially use ESC_t because it is a POSIX reserved identifier  /* We can't officially use ESC_t because it is a POSIX reserved identifier
1220  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1221    
1222  #ifndef ESC_tee  #ifndef ESC_tee
1223  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1224  #endif  #endif
1225    
1226  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1227    
1228  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1229  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1230  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1231  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1232  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1233    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1234    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1235    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1236    #define PT_WORD       8    /* Word - L plus N plus underscore */
1237    
1238  /* 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
1239  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 597  contain UTF-8 characters with values gre Line 1250  contain UTF-8 characters with values gre
1250  /* 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
1251  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
1252  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
1253  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1254  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1255  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  used for [^] in JavaScript compatibility mode. In non-DOTALL mode, "." behaves
1256  There are two tests in the code for an escape greater than ESC_b and less than  like \N.
1257  ESC_Z to detect the types that may be repeated. These are the types that  
1258  consume characters. If any new escapes are put in between that don't consume a  The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1259  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1260    They must be contiguous, and remain in order so that the replacements can be
1261    looked up from a table.
1262    
1263    The final escape must be ESC_REF as subsequent values are used for
1264    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1265    greater than ESC_b and less than ESC_Z to detect the types that may be
1266    repeated. These are the types that consume characters. If any new escapes are
1267    put in between that don't consume a character, that code will have to change.
1268    */
1269    
1270  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1271         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1272         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1273           ESC_E, ESC_Q, ESC_g, ESC_k,
1274           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1275           ESC_REF };
1276    
1277  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  
1278  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.
1279    
 To keep stored, compiled patterns compatible, new opcodes should be added  
 immediately before OP_BRA, where (since release 7.0) a gap is left for this  
 purpose.  
   
1280  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1281  that follow must also be updated to match. There is also a table called  that follow must also be updated to match. There are also tables called
1282  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1283    
1284  enum {  enum {
1285    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 638  enum { Line 1297  enum {
1297    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1298    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1299    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1300    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character except newline */
1301    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character */
1302    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1303    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1304    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
1305    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1306    OP_HSPACE,         /* 18 \h (horizontal whitespace) */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1307    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1308    OP_VSPACE,         /* 20 \v (vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1309    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1310    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1311    OP_EOD,            /* 23 End of data: \z */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1312      OP_EOD,            /* 24 End of data: \z */
1313    OP_OPT,            /* 24 Set runtime options */  
1314    OP_CIRC,           /* 25 Start of line - varies with multiline switch */    OP_OPT,            /* 25 Set runtime options */
1315    OP_DOLL,           /* 26 End of line - varies with multiline switch */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1316    OP_CHAR,           /* 27 Match one character, casefully */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1317    OP_CHARNC,         /* 28 Match one character, caselessly */    OP_CHAR,           /* 28 Match one character, casefully */
1318    OP_NOT,            /* 29 Match one character, not the following one */    OP_CHARNC,         /* 29 Match one character, caselessly */
1319      OP_NOT,            /* 30 Match one character, not the following one */
1320    OP_STAR,           /* 30 The maximizing and minimizing versions of */  
1321    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
1322    OP_PLUS,           /* 32 the minimizing one second. */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1323    OP_MINPLUS,        /* 33 This first set applies to single characters.*/    OP_PLUS,           /* 33 the minimizing one second. */
1324    OP_QUERY,          /* 34 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1325    OP_MINQUERY,       /* 35 */    OP_QUERY,          /* 35 */
1326      OP_MINQUERY,       /* 36 */
1327    OP_UPTO,           /* 36 From 0 to n matches */  
1328    OP_MINUPTO,        /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
1329    OP_EXACT,          /* 38 Exactly n matches */    OP_MINUPTO,        /* 38 */
1330      OP_EXACT,          /* 39 Exactly n matches */
1331    OP_POSSTAR,        /* 39 Possessified star */  
1332    OP_POSPLUS,        /* 40 Possessified plus */    OP_POSSTAR,        /* 40 Possessified star */
1333    OP_POSQUERY,       /* 41 Posesssified query */    OP_POSPLUS,        /* 41 Possessified plus */
1334    OP_POSUPTO,        /* 42 Possessified upto */    OP_POSQUERY,       /* 42 Posesssified query */
1335      OP_POSUPTO,        /* 43 Possessified upto */
1336    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
1337    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1338    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1339    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1340    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1341    OP_NOTMINQUERY,    /* 48 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1342      OP_NOTMINQUERY,    /* 49 */
1343    OP_NOTUPTO,        /* 49 From 0 to n matches */  
1344    OP_NOTMINUPTO,     /* 50 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1345    OP_NOTEXACT,       /* 51 Exactly n matches */    OP_NOTMINUPTO,     /* 51 */
1346      OP_NOTEXACT,       /* 52 Exactly n matches */
1347    OP_NOTPOSSTAR,     /* 52 Possessified versions */  
1348    OP_NOTPOSPLUS,     /* 53 */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1349    OP_NOTPOSQUERY,    /* 54 */    OP_NOTPOSPLUS,     /* 54 */
1350    OP_NOTPOSUPTO,     /* 55 */    OP_NOTPOSQUERY,    /* 55 */
1351      OP_NOTPOSUPTO,     /* 56 */
1352    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
1353    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1354    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1355    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1356    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1357    OP_TYPEMINQUERY,   /* 61 */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1358      OP_TYPEMINQUERY,   /* 62 */
1359    OP_TYPEUPTO,       /* 62 From 0 to n matches */  
1360    OP_TYPEMINUPTO,    /* 63 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
1361    OP_TYPEEXACT,      /* 64 Exactly n matches */    OP_TYPEMINUPTO,    /* 64 */
1362      OP_TYPEEXACT,      /* 65 Exactly n matches */
1363    OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
1364    OP_TYPEPOSPLUS,    /* 66 */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1365    OP_TYPEPOSQUERY,   /* 67 */    OP_TYPEPOSPLUS,    /* 67 */
1366    OP_TYPEPOSUPTO,    /* 68 */    OP_TYPEPOSQUERY,   /* 68 */
1367      OP_TYPEPOSUPTO,    /* 69 */
1368    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
1369    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1370    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1371    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1372    OP_CRQUERY,        /* 73 These are for character classes and back refs */    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1373    OP_CRMINQUERY,     /* 74 */    OP_CRQUERY,        /* 74 These are for character classes and back refs */
1374    OP_CRRANGE,        /* 75 These are different to the three sets above. */    OP_CRMINQUERY,     /* 75 */
1375    OP_CRMINRANGE,     /* 76 */    OP_CRRANGE,        /* 76 These are different to the three sets above. */
1376      OP_CRMINRANGE,     /* 77 */
1377    
1378    OP_CLASS,          /* 77 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1379    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1380                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1381                             character > 255 is encountered. */                             character > 255 is encountered. */
1382    
1383    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1384                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1385    
1386    OP_REF,            /* 80 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1387    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1388    OP_CALLOUT,        /* 82 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1389    
1390    OP_ALT,            /* 83 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1391    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
1392    OP_KETRMAX,        /* 85 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1393    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1394    
1395    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1396    
1397    OP_ASSERT,         /* 87 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1398    OP_ASSERT_NOT,     /* 88 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1399    OP_ASSERTBACK,     /* 89 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1400    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1401    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1402    
1403    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
1404    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
1405    
1406    OP_ONCE,           /* 92 Atomic group */    OP_ONCE,           /* 93 Atomic group */
1407    OP_BRA,            /* 93 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
1408    OP_CBRA,           /* 94 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
1409    OP_COND,           /* 95 Conditional group */    OP_COND,           /* 96 Conditional group */
1410    
1411    /* These three must follow the previous three, in the same order. There's a    /* These three must follow the previous three, in the same order. There's a
1412    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1413    
1414    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1415    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1416    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1417    
1418    OP_CREF,           /* 99 Used to hold a capture number as condition */    /* The next two pairs must (respectively) be kept together. */
1419    OP_RREF,           /* 100 Used to hold a recursion number as condition */  
1420    OP_DEF,            /* 101 The DEFINE condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1421      OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1422      OP_RREF,           /* 102 Used to hold a recursion number as condition */
1423      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1424      OP_DEF,            /* 104 The DEFINE condition */
1425    
1426      OP_BRAZERO,        /* 105 These two must remain together and in this */
1427      OP_BRAMINZERO,     /* 106 order. */
1428    
1429      /* These are backtracking control verbs */
1430    
1431      OP_MARK,           /* 107 always has an argument */
1432      OP_PRUNE,          /* 108 */
1433      OP_PRUNE_ARG,      /* 109 same, but with argument */
1434      OP_SKIP,           /* 110 */
1435      OP_SKIP_ARG,       /* 111 same, but with argument */
1436      OP_THEN,           /* 112 */
1437      OP_THEN_ARG,       /* 113 same, but with argument */
1438      OP_COMMIT,         /* 114 */
1439    
1440      /* These are forced failure and success verbs */
1441    
1442      OP_FAIL,           /* 115 */
1443      OP_ACCEPT,         /* 116 */
1444      OP_CLOSE,          /* 117 Used before OP_ACCEPT to close open captures */
1445    
1446      /* This is used to skip a subpattern with a {0} quantifier */
1447    
1448    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_SKIPZERO,       /* 118 */
1449    OP_BRAMINZERO      /* 103 order. */  
1450      /* This is not an opcode, but is used to check that tables indexed by opcode
1451      are the correct length, in order to catch updating errors - there have been
1452      some in the past. */
1453    
1454      OP_TABLE_LENGTH
1455  };  };
1456    
1457    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1458    definitions that follow must also be updated to match. There are also tables
1459    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1460    
1461    
1462  /* This macro defines textual names for all the opcodes. These are used only  /* This macro defines textual names for all the opcodes. These are used only
1463  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
1464    
1465  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1466    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1467    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1468    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1469    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1470    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
# Line 783  for debugging. The macro is referenced o Line 1478  for debugging. The macro is referenced o
1478    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
1479    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1480    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1481    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1482    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1483      "Brazero", "Braminzero",                                        \
1484      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1485      "*THEN", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",                \
1486      "Close", "Skip zero"
1487    
1488    
1489  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 800  in UTF-8 mode. The code that uses this t Line 1499  in UTF-8 mode. The code that uses this t
1499    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1500    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1501    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1502    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1503    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1504    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1505      1,                             /* \X                                     */ \
1506    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1507    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1508    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 844  in UTF-8 mode. The code that uses this t Line 1544  in UTF-8 mode. The code that uses this t
1544    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1545    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1546    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1547    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1548    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1549    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1550    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1551      3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1552      1, 3,                          /* SKIP, SKIP_ARG                         */ \
1553      1+LINK_SIZE, 3+LINK_SIZE,      /* THEN, THEN_ARG                         */ \
1554      1, 1, 1, 3, 1                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */
1555    
1556    
1557  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1558    condition. */
1559    
1560  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1561    
1562  /* Error code numbers. They are given names so that they can more easily be  /* Compile time error code numbers. They are given names so that they can more
1563  tracked. */  easily be tracked. When a new number is added, the table called eint in
1564    pcreposix.c must be updated. */
1565    
1566  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1567         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1568         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1569         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1570         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1571         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1572           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERRCOUNT };
1573    
1574  /* 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
1575  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 878  Because people can now save and re-use c Line 1585  Because people can now save and re-use c
1585  structure should be made at the end, and something earlier (e.g. a new  structure should be made at the end, and something earlier (e.g. a new
1586  flag in the options or one of the dummy fields) should indicate that the new  flag in the options or one of the dummy fields) should indicate that the new
1587  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1588  NOTE NOTE NOTE:  NOTE NOTE NOTE
1589  */  */
1590    
1591  typedef struct real_pcre {  typedef struct real_pcre {
1592    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1593    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1594    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1595    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1596      pcre_uint16 dummy1;             /* For future use */
1597    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1598    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1599    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 905  remark (see NOTE above) about extending Line 1612  remark (see NOTE above) about extending
1612    
1613  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1614    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1615    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1616    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1617      pcre_uint32 minlength;          /* Minimum subject length */
1618  } pcre_study_data;  } pcre_study_data;
1619    
1620    /* Structure for building a chain of open capturing subpatterns during
1621    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1622    encountered. This is also used to identify subpatterns that contain recursive
1623    back references to themselves, so that they can be made atomic. */
1624    
1625    typedef struct open_capitem {
1626      struct open_capitem *next;    /* Chain link */
1627      pcre_uint16 number;           /* Capture number */
1628      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1629    } open_capitem;
1630    
1631  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1632  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1633    
# Line 921  typedef struct compile_data { Line 1640  typedef struct compile_data {
1640    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1641    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1642    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1643      open_capitem *open_caps;      /* Chain of open capture items */
1644    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1645    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1646    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1647    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1648    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
1649      int  final_bracount;          /* Saved value after first pass */
1650    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1651    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1652    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1653      int  external_flags;          /* External flag bits to be set */
1654    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1655    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
1656      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1657    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1658    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1659    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 941  branches, for testing for left recursion Line 1664  branches, for testing for left recursion
1664    
1665  typedef struct branch_chain {  typedef struct branch_chain {
1666    struct branch_chain *outer;    struct branch_chain *outer;
1667    uschar *current;    uschar *current_branch;
1668  } branch_chain;  } branch_chain;
1669    
1670  /* 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 951  typedef struct recursion_info { Line 1674  typedef struct recursion_info {
1674    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1675    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1676    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
   USPTR save_start;             /* Old value of mstart */  
1677    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1678    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1679      int save_offset_top;          /* Current value of offset_top */
1680  } recursion_info;  } recursion_info;
1681    
1682  /* 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 978  typedef struct match_data { Line 1701  typedef struct match_data {
1701    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1702    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1703    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1704      int    name_count;            /* Number of names in name table */
1705      int    name_entry_size;       /* Size of entry in names table */
1706      uschar *name_table;           /* Table of names */
1707    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1708    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1709    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 985  typedef struct match_data { Line 1711  typedef struct match_data {
1711    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1712    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1713    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1714      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1715      BOOL   use_ucp;               /* PCRE_UCP flag */
1716    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1717    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1718    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1719    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1720      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1721    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1722    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1723    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1724    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1725    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1726      USPTR  start_used_ptr;        /* Earliest consulted character */
1727      int    partial;               /* PARTIAL options */
1728    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1729    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1730    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1001  typedef struct match_data { Line 1732  typedef struct match_data {
1732    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1733    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1734    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1735      const uschar *mark;           /* Mark pointer to pass back */
1736  } match_data;  } match_data;
1737    
1738  /* 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
# Line 1010  typedef struct dfa_match_data { Line 1742  typedef struct dfa_match_data {
1742    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1743    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1744    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1745      const uschar *start_used_ptr; /* Earliest consulted character */
1746    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1747      int   start_offset;           /* The start offset value */
1748    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1749    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1750    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1025  typedef struct dfa_match_data { Line 1759  typedef struct dfa_match_data {
1759  #define ctype_letter  0x02  #define ctype_letter  0x02
1760  #define ctype_digit   0x04  #define ctype_digit   0x04
1761  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1762  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1763  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1764    
1765  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
# Line 1053  total length. */ Line 1787  total length. */
1787  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1788    
1789  /* 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
1790  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1791    relocations in shared libraries, it now has an offset into a single string
1792    instead. */
1793    
1794  typedef struct {  typedef struct {
1795    const char *name;    pcre_uint16 name_offset;
1796    pcre_uint16 type;    pcre_uint16 type;
1797    pcre_uint16 value;    pcre_uint16 value;
1798  } ucp_type_table;  } ucp_type_table;
# Line 1074  extern const uschar _pcre_utf8_table4[]; Line 1810  extern const uschar _pcre_utf8_table4[];
1810    
1811  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1812    
1813    extern const char   _pcre_utt_names[];
1814  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1815  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1816    
# Line 1086  extern const uschar _pcre_OP_lengths[]; Line 1823  extern const uschar _pcre_OP_lengths[];
1823  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
1824  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1825    
1826  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1827                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1828  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1829  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1830                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1831  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  extern int           _pcre_valid_utf8(USPTR, int);
1832  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1833  extern int          _pcre_valid_utf8(const uschar *, int);  extern BOOL          _pcre_xclass(int, const uschar *);
1834  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  
1835                        int *, BOOL);  
1836  extern BOOL         _pcre_xclass(int, const uschar *);  /* Unicode character database (UCD) */
1837    
1838    typedef struct {
1839      uschar script;
1840      uschar chartype;
1841      pcre_int32 other_case;
1842    } ucd_record;
1843    
1844    extern const ucd_record  _pcre_ucd_records[];
1845    extern const uschar      _pcre_ucd_stage1[];
1846    extern const pcre_uint16 _pcre_ucd_stage2[];
1847    extern const int         _pcre_ucp_gentype[];
1848    
1849    
1850    /* UCD access macros */
1851    
1852    #define UCD_BLOCK_SIZE 128
1853    #define GET_UCD(ch) (_pcre_ucd_records + \
1854            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1855            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1856    
1857    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1858    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1859    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1860    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1861    
1862  #endif  #endif
1863    

Legend:
Removed from v.190  
changed lines
  Added in v.573

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12