/[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 118 by ph10, Fri Mar 9 16:17:46 2007 UTC revision 716 by ph10, Tue Oct 4 16:38:05 2011 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) 1997-2011 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>
92  #include <string.h>  #include <string.h>
93    
94  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
95  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
96    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
97    information there, using __declspec(dllexport) without "extern" we have a
98    definition; with "extern" we have a declaration. The settings here override the
99    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
100    which is all that is needed for applications (they just import the symbols). We
101    use:
102    
103      PCRE_EXP_DECL       for declarations
104      PCRE_EXP_DEFN       for definitions of exported functions
105      PCRE_EXP_DATA_DEFN  for definitions of exported variables
106    
107    The reason for the two DEFN macros is that in non-Windows environments, one
108    does not want to have "extern" before variable definitions because it leads to
109    compiler warnings. So we distinguish between functions and variables. In
110    Windows, the two should always be the same.
111    
112    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
113    which is an application, but needs to import this file in order to "peek" at
114    internals, can #include pcre.h first to get an application's-eye view.
115    
116    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
117    special-purpose environments) might want to stick other stuff in front of
118    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
119    PCRE_EXP_DATA_DEFN only if they are not already set. */
120    
121    #ifndef PCRE_EXP_DECL
122    #  ifdef _WIN32
123    #    ifndef PCRE_STATIC
124    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
125    #      define PCRE_EXP_DEFN       __declspec(dllexport)
126    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
127    #    else
128    #      define PCRE_EXP_DECL       extern
129    #      define PCRE_EXP_DEFN
130    #      define PCRE_EXP_DATA_DEFN
131    #    endif
132    #  else
133    #    ifdef __cplusplus
134    #      define PCRE_EXP_DECL       extern "C"
135    #    else
136    #      define PCRE_EXP_DECL       extern
137    #    endif
138    #    ifndef PCRE_EXP_DEFN
139    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
140    #    endif
141    #    ifndef PCRE_EXP_DATA_DEFN
142    #      define PCRE_EXP_DATA_DEFN
143    #    endif
144    #  endif
145    #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  #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
# Line 95  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 125  characters only go up to 0x7fffffff (tho Line 220  characters only go up to 0x7fffffff (tho
220  #define NOTACHAR 0xffffffff  #define NOTACHAR 0xffffffff
221    
222  /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,  /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
223  and "all" at present). The following macros are used to package up testing for  "any" and "anycrlf" at present). The following macros are used to package up
224  newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to  testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
225  indicate in which datablock the parameters exist, and what the start/end of  modules to indicate in which datablock the parameters exist, and what the
226  string field names are. */  start/end of string field names are. */
227    
228  #define NLTYPE_FIXED   0     /* Newline is a fixed length string */  #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
229  #define NLTYPE_ANY     1     /* Newline is any Unicode line ending */  #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
230    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
231    
232  /* This macro checks for a newline at the given position */  /* This macro checks for a newline at the given position */
233    
234  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
235    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
236      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
237       _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
238      ) \         utf8)) \
239      : \      : \
240      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
241       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 152  string field names are. */ Line 248  string field names are. */
248  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
249    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
250      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
251       _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
252      ) \         &(NLBLOCK->nllen), utf8)) \
253      : \      : \
254      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
255       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 178  must begin with PCRE_. */ Line 274  must begin with PCRE_. */
274  #define USPTR const unsigned char *  #define USPTR const unsigned char *
275  #endif  #endif
276    
277    
278    
279  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
280  values. */  values. */
281    
# Line 202  define a macro for memmove() if HAVE_MEM Line 300  define a macro for memmove() if HAVE_MEM
300  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
301  neither (there some non-Unix environments where this is the case). */  neither (there some non-Unix environments where this is the case). */
302    
303  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
304  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
305  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
306  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
307  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
308  static void *  static void *
# Line 223  if (dest > src) Line 321  if (dest > src)
321  else  else
322    {    {
323    for (i = 0; i < n; ++i) *dest++ = *src++;    for (i = 0; i < n; ++i) *dest++ = *src++;
324    return (void *)(dest - n);    return (void *)(dest - n);
325    }    }
326  }  }
327  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
328  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 310  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 318  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 363  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) \
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
522        { \        { \
523        gcss -= 6; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
524        c |= (*eptr++ & 0x3f) << gcss; \        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          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
540              ((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 396  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 426  Standard C system should have one. */ Line 582  Standard C system should have one. */
582  #endif  #endif
583    
584    
585  /* These are the public options that can change during matching. */  /* Private flags containing information about the compiled regex. They used to
586    live at the top end of the options word, but that got almost full, so now they
587    are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
588    the restrictions on partial matching have been lifted. It remains for backwards
589    compatibility. */
590    
591    #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
592    #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
593    #define PCRE_REQCHSET      0x0004  /* req_byte is set */
594    #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
595    #define PCRE_JCHANGED      0x0010  /* j option used in regex */
596    #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
597    #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
598    
599  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  /* Flags for the "extra" block produced by pcre_study(). */
600    
601  /* Private options flags start at the most significant end of the four bytes.  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
602  The public options defined in pcre.h start at the least significant end. Make  #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
 sure they don't overlap! The bits are getting a bit scarce now -- when we run  
 out, there is a dummy word in the structure that could be used for the private  
 bits. */  
   
 #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  
 #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  
 #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  
 #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  
 #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  
   
 /* Options for the "extra" block produced by pcre_study(). */  
   
 #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  
603    
604  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
605  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
606    
607  #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| \
608                               PCRE_NEWLINE_ANYCRLF)
609    
610  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
611    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
612     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
613     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| \
614     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
615       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
616    
617  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
618    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
619     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
620       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
621    
622  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
623    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
624     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
625       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
626       PCRE_NO_START_OPTIMIZE)
627    
628  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
629       PCRE_STUDY_JIT_COMPILE
630    
631  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. Also used
632  to detect whether a pattern was compiled on a host of different endianness. */  to detect whether a pattern was compiled on a host of different endianness. */
# Line 488  variable-length repeat, or a anything ot Line 649  variable-length repeat, or a anything ot
649  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
650  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
651    
652  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
653    environments where these macros are defined elsewhere. Unfortunately, there
654    is no way to do the same for the typedef. */
655    
656  typedef int BOOL;  typedef int BOOL;
657    
658    #ifndef FALSE
659  #define FALSE   0  #define FALSE   0
660  #define TRUE    1  #define TRUE    1
661    #endif
662    
663    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
664    character constants like '*' because the compiler would emit their EBCDIC code,
665    which is different from their ASCII/UTF-8 code. Instead we define macros for
666    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
667    is enabled. When UTF-8 support is not enabled, the definitions use character
668    literals. Both character and string versions of each character are needed, and
669    there are some longer strings as well.
670    
671    This means that, on EBCDIC platforms, the PCRE library can handle either
672    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
673    would need different lookups depending on whether PCRE_UTF8 was set or not.
674    This would make it impossible to use characters in switch/case statements,
675    which would reduce performance. For a theoretical use (which nobody has asked
676    for) in a minority area (EBCDIC platforms), this is not sensible. Any
677    application that did need both could compile two versions of the library, using
678    macros to give the functions distinct names. */
679    
680    #ifndef SUPPORT_UTF8
681    
682    /* UTF-8 support is not enabled; use the platform-dependent character literals
683    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
684    
685    #define CHAR_HT                     '\t'
686    #define CHAR_VT                     '\v'
687    #define CHAR_FF                     '\f'
688    #define CHAR_CR                     '\r'
689    #define CHAR_NL                     '\n'
690    #define CHAR_BS                     '\b'
691    #define CHAR_BEL                    '\a'
692    #ifdef EBCDIC
693    #define CHAR_ESC                    '\047'
694    #define CHAR_DEL                    '\007'
695    #else
696    #define CHAR_ESC                    '\033'
697    #define CHAR_DEL                    '\177'
698    #endif
699    
700    #define CHAR_SPACE                  ' '
701    #define CHAR_EXCLAMATION_MARK       '!'
702    #define CHAR_QUOTATION_MARK         '"'
703    #define CHAR_NUMBER_SIGN            '#'
704    #define CHAR_DOLLAR_SIGN            '$'
705    #define CHAR_PERCENT_SIGN           '%'
706    #define CHAR_AMPERSAND              '&'
707    #define CHAR_APOSTROPHE             '\''
708    #define CHAR_LEFT_PARENTHESIS       '('
709    #define CHAR_RIGHT_PARENTHESIS      ')'
710    #define CHAR_ASTERISK               '*'
711    #define CHAR_PLUS                   '+'
712    #define CHAR_COMMA                  ','
713    #define CHAR_MINUS                  '-'
714    #define CHAR_DOT                    '.'
715    #define CHAR_SLASH                  '/'
716    #define CHAR_0                      '0'
717    #define CHAR_1                      '1'
718    #define CHAR_2                      '2'
719    #define CHAR_3                      '3'
720    #define CHAR_4                      '4'
721    #define CHAR_5                      '5'
722    #define CHAR_6                      '6'
723    #define CHAR_7                      '7'
724    #define CHAR_8                      '8'
725    #define CHAR_9                      '9'
726    #define CHAR_COLON                  ':'
727    #define CHAR_SEMICOLON              ';'
728    #define CHAR_LESS_THAN_SIGN         '<'
729    #define CHAR_EQUALS_SIGN            '='
730    #define CHAR_GREATER_THAN_SIGN      '>'
731    #define CHAR_QUESTION_MARK          '?'
732    #define CHAR_COMMERCIAL_AT          '@'
733    #define CHAR_A                      'A'
734    #define CHAR_B                      'B'
735    #define CHAR_C                      'C'
736    #define CHAR_D                      'D'
737    #define CHAR_E                      'E'
738    #define CHAR_F                      'F'
739    #define CHAR_G                      'G'
740    #define CHAR_H                      'H'
741    #define CHAR_I                      'I'
742    #define CHAR_J                      'J'
743    #define CHAR_K                      'K'
744    #define CHAR_L                      'L'
745    #define CHAR_M                      'M'
746    #define CHAR_N                      'N'
747    #define CHAR_O                      'O'
748    #define CHAR_P                      'P'
749    #define CHAR_Q                      'Q'
750    #define CHAR_R                      'R'
751    #define CHAR_S                      'S'
752    #define CHAR_T                      'T'
753    #define CHAR_U                      'U'
754    #define CHAR_V                      'V'
755    #define CHAR_W                      'W'
756    #define CHAR_X                      'X'
757    #define CHAR_Y                      'Y'
758    #define CHAR_Z                      'Z'
759    #define CHAR_LEFT_SQUARE_BRACKET    '['
760    #define CHAR_BACKSLASH              '\\'
761    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
762    #define CHAR_CIRCUMFLEX_ACCENT      '^'
763    #define CHAR_UNDERSCORE             '_'
764    #define CHAR_GRAVE_ACCENT           '`'
765    #define CHAR_a                      'a'
766    #define CHAR_b                      'b'
767    #define CHAR_c                      'c'
768    #define CHAR_d                      'd'
769    #define CHAR_e                      'e'
770    #define CHAR_f                      'f'
771    #define CHAR_g                      'g'
772    #define CHAR_h                      'h'
773    #define CHAR_i                      'i'
774    #define CHAR_j                      'j'
775    #define CHAR_k                      'k'
776    #define CHAR_l                      'l'
777    #define CHAR_m                      'm'
778    #define CHAR_n                      'n'
779    #define CHAR_o                      'o'
780    #define CHAR_p                      'p'
781    #define CHAR_q                      'q'
782    #define CHAR_r                      'r'
783    #define CHAR_s                      's'
784    #define CHAR_t                      't'
785    #define CHAR_u                      'u'
786    #define CHAR_v                      'v'
787    #define CHAR_w                      'w'
788    #define CHAR_x                      'x'
789    #define CHAR_y                      'y'
790    #define CHAR_z                      'z'
791    #define CHAR_LEFT_CURLY_BRACKET     '{'
792    #define CHAR_VERTICAL_LINE          '|'
793    #define CHAR_RIGHT_CURLY_BRACKET    '}'
794    #define CHAR_TILDE                  '~'
795    
796    #define STR_HT                      "\t"
797    #define STR_VT                      "\v"
798    #define STR_FF                      "\f"
799    #define STR_CR                      "\r"
800    #define STR_NL                      "\n"
801    #define STR_BS                      "\b"
802    #define STR_BEL                     "\a"
803    #ifdef EBCDIC
804    #define STR_ESC                     "\047"
805    #define STR_DEL                     "\007"
806    #else
807    #define STR_ESC                     "\033"
808    #define STR_DEL                     "\177"
809    #endif
810    
811    #define STR_SPACE                   " "
812    #define STR_EXCLAMATION_MARK        "!"
813    #define STR_QUOTATION_MARK          "\""
814    #define STR_NUMBER_SIGN             "#"
815    #define STR_DOLLAR_SIGN             "$"
816    #define STR_PERCENT_SIGN            "%"
817    #define STR_AMPERSAND               "&"
818    #define STR_APOSTROPHE              "'"
819    #define STR_LEFT_PARENTHESIS        "("
820    #define STR_RIGHT_PARENTHESIS       ")"
821    #define STR_ASTERISK                "*"
822    #define STR_PLUS                    "+"
823    #define STR_COMMA                   ","
824    #define STR_MINUS                   "-"
825    #define STR_DOT                     "."
826    #define STR_SLASH                   "/"
827    #define STR_0                       "0"
828    #define STR_1                       "1"
829    #define STR_2                       "2"
830    #define STR_3                       "3"
831    #define STR_4                       "4"
832    #define STR_5                       "5"
833    #define STR_6                       "6"
834    #define STR_7                       "7"
835    #define STR_8                       "8"
836    #define STR_9                       "9"
837    #define STR_COLON                   ":"
838    #define STR_SEMICOLON               ";"
839    #define STR_LESS_THAN_SIGN          "<"
840    #define STR_EQUALS_SIGN             "="
841    #define STR_GREATER_THAN_SIGN       ">"
842    #define STR_QUESTION_MARK           "?"
843    #define STR_COMMERCIAL_AT           "@"
844    #define STR_A                       "A"
845    #define STR_B                       "B"
846    #define STR_C                       "C"
847    #define STR_D                       "D"
848    #define STR_E                       "E"
849    #define STR_F                       "F"
850    #define STR_G                       "G"
851    #define STR_H                       "H"
852    #define STR_I                       "I"
853    #define STR_J                       "J"
854    #define STR_K                       "K"
855    #define STR_L                       "L"
856    #define STR_M                       "M"
857    #define STR_N                       "N"
858    #define STR_O                       "O"
859    #define STR_P                       "P"
860    #define STR_Q                       "Q"
861    #define STR_R                       "R"
862    #define STR_S                       "S"
863    #define STR_T                       "T"
864    #define STR_U                       "U"
865    #define STR_V                       "V"
866    #define STR_W                       "W"
867    #define STR_X                       "X"
868    #define STR_Y                       "Y"
869    #define STR_Z                       "Z"
870    #define STR_LEFT_SQUARE_BRACKET     "["
871    #define STR_BACKSLASH               "\\"
872    #define STR_RIGHT_SQUARE_BRACKET    "]"
873    #define STR_CIRCUMFLEX_ACCENT       "^"
874    #define STR_UNDERSCORE              "_"
875    #define STR_GRAVE_ACCENT            "`"
876    #define STR_a                       "a"
877    #define STR_b                       "b"
878    #define STR_c                       "c"
879    #define STR_d                       "d"
880    #define STR_e                       "e"
881    #define STR_f                       "f"
882    #define STR_g                       "g"
883    #define STR_h                       "h"
884    #define STR_i                       "i"
885    #define STR_j                       "j"
886    #define STR_k                       "k"
887    #define STR_l                       "l"
888    #define STR_m                       "m"
889    #define STR_n                       "n"
890    #define STR_o                       "o"
891    #define STR_p                       "p"
892    #define STR_q                       "q"
893    #define STR_r                       "r"
894    #define STR_s                       "s"
895    #define STR_t                       "t"
896    #define STR_u                       "u"
897    #define STR_v                       "v"
898    #define STR_w                       "w"
899    #define STR_x                       "x"
900    #define STR_y                       "y"
901    #define STR_z                       "z"
902    #define STR_LEFT_CURLY_BRACKET      "{"
903    #define STR_VERTICAL_LINE           "|"
904    #define STR_RIGHT_CURLY_BRACKET     "}"
905    #define STR_TILDE                   "~"
906    
907    #define STRING_ACCEPT0              "ACCEPT\0"
908    #define STRING_COMMIT0              "COMMIT\0"
909    #define STRING_F0                   "F\0"
910    #define STRING_FAIL0                "FAIL\0"
911    #define STRING_MARK0                "MARK\0"
912    #define STRING_PRUNE0               "PRUNE\0"
913    #define STRING_SKIP0                "SKIP\0"
914    #define STRING_THEN                 "THEN"
915    
916    #define STRING_alpha0               "alpha\0"
917    #define STRING_lower0               "lower\0"
918    #define STRING_upper0               "upper\0"
919    #define STRING_alnum0               "alnum\0"
920    #define STRING_ascii0               "ascii\0"
921    #define STRING_blank0               "blank\0"
922    #define STRING_cntrl0               "cntrl\0"
923    #define STRING_digit0               "digit\0"
924    #define STRING_graph0               "graph\0"
925    #define STRING_print0               "print\0"
926    #define STRING_punct0               "punct\0"
927    #define STRING_space0               "space\0"
928    #define STRING_word0                "word\0"
929    #define STRING_xdigit               "xdigit"
930    
931    #define STRING_DEFINE               "DEFINE"
932    
933    #define STRING_CR_RIGHTPAR             "CR)"
934    #define STRING_LF_RIGHTPAR             "LF)"
935    #define STRING_CRLF_RIGHTPAR           "CRLF)"
936    #define STRING_ANY_RIGHTPAR            "ANY)"
937    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
938    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
939    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
940    #define STRING_UTF8_RIGHTPAR           "UTF8)"
941    #define STRING_UCP_RIGHTPAR            "UCP)"
942    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
943    
944    #else  /* SUPPORT_UTF8 */
945    
946    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
947    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
948    only. */
949    
950    #define CHAR_HT                     '\011'
951    #define CHAR_VT                     '\013'
952    #define CHAR_FF                     '\014'
953    #define CHAR_CR                     '\015'
954    #define CHAR_NL                     '\012'
955    #define CHAR_BS                     '\010'
956    #define CHAR_BEL                    '\007'
957    #define CHAR_ESC                    '\033'
958    #define CHAR_DEL                    '\177'
959    
960    #define CHAR_SPACE                  '\040'
961    #define CHAR_EXCLAMATION_MARK       '\041'
962    #define CHAR_QUOTATION_MARK         '\042'
963    #define CHAR_NUMBER_SIGN            '\043'
964    #define CHAR_DOLLAR_SIGN            '\044'
965    #define CHAR_PERCENT_SIGN           '\045'
966    #define CHAR_AMPERSAND              '\046'
967    #define CHAR_APOSTROPHE             '\047'
968    #define CHAR_LEFT_PARENTHESIS       '\050'
969    #define CHAR_RIGHT_PARENTHESIS      '\051'
970    #define CHAR_ASTERISK               '\052'
971    #define CHAR_PLUS                   '\053'
972    #define CHAR_COMMA                  '\054'
973    #define CHAR_MINUS                  '\055'
974    #define CHAR_DOT                    '\056'
975    #define CHAR_SLASH                  '\057'
976    #define CHAR_0                      '\060'
977    #define CHAR_1                      '\061'
978    #define CHAR_2                      '\062'
979    #define CHAR_3                      '\063'
980    #define CHAR_4                      '\064'
981    #define CHAR_5                      '\065'
982    #define CHAR_6                      '\066'
983    #define CHAR_7                      '\067'
984    #define CHAR_8                      '\070'
985    #define CHAR_9                      '\071'
986    #define CHAR_COLON                  '\072'
987    #define CHAR_SEMICOLON              '\073'
988    #define CHAR_LESS_THAN_SIGN         '\074'
989    #define CHAR_EQUALS_SIGN            '\075'
990    #define CHAR_GREATER_THAN_SIGN      '\076'
991    #define CHAR_QUESTION_MARK          '\077'
992    #define CHAR_COMMERCIAL_AT          '\100'
993    #define CHAR_A                      '\101'
994    #define CHAR_B                      '\102'
995    #define CHAR_C                      '\103'
996    #define CHAR_D                      '\104'
997    #define CHAR_E                      '\105'
998    #define CHAR_F                      '\106'
999    #define CHAR_G                      '\107'
1000    #define CHAR_H                      '\110'
1001    #define CHAR_I                      '\111'
1002    #define CHAR_J                      '\112'
1003    #define CHAR_K                      '\113'
1004    #define CHAR_L                      '\114'
1005    #define CHAR_M                      '\115'
1006    #define CHAR_N                      '\116'
1007    #define CHAR_O                      '\117'
1008    #define CHAR_P                      '\120'
1009    #define CHAR_Q                      '\121'
1010    #define CHAR_R                      '\122'
1011    #define CHAR_S                      '\123'
1012    #define CHAR_T                      '\124'
1013    #define CHAR_U                      '\125'
1014    #define CHAR_V                      '\126'
1015    #define CHAR_W                      '\127'
1016    #define CHAR_X                      '\130'
1017    #define CHAR_Y                      '\131'
1018    #define CHAR_Z                      '\132'
1019    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1020    #define CHAR_BACKSLASH              '\134'
1021    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1022    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1023    #define CHAR_UNDERSCORE             '\137'
1024    #define CHAR_GRAVE_ACCENT           '\140'
1025    #define CHAR_a                      '\141'
1026    #define CHAR_b                      '\142'
1027    #define CHAR_c                      '\143'
1028    #define CHAR_d                      '\144'
1029    #define CHAR_e                      '\145'
1030    #define CHAR_f                      '\146'
1031    #define CHAR_g                      '\147'
1032    #define CHAR_h                      '\150'
1033    #define CHAR_i                      '\151'
1034    #define CHAR_j                      '\152'
1035    #define CHAR_k                      '\153'
1036    #define CHAR_l                      '\154'
1037    #define CHAR_m                      '\155'
1038    #define CHAR_n                      '\156'
1039    #define CHAR_o                      '\157'
1040    #define CHAR_p                      '\160'
1041    #define CHAR_q                      '\161'
1042    #define CHAR_r                      '\162'
1043    #define CHAR_s                      '\163'
1044    #define CHAR_t                      '\164'
1045    #define CHAR_u                      '\165'
1046    #define CHAR_v                      '\166'
1047    #define CHAR_w                      '\167'
1048    #define CHAR_x                      '\170'
1049    #define CHAR_y                      '\171'
1050    #define CHAR_z                      '\172'
1051    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1052    #define CHAR_VERTICAL_LINE          '\174'
1053    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1054    #define CHAR_TILDE                  '\176'
1055    
1056    #define STR_HT                      "\011"
1057    #define STR_VT                      "\013"
1058    #define STR_FF                      "\014"
1059    #define STR_CR                      "\015"
1060    #define STR_NL                      "\012"
1061    #define STR_BS                      "\010"
1062    #define STR_BEL                     "\007"
1063    #define STR_ESC                     "\033"
1064    #define STR_DEL                     "\177"
1065    
1066    #define STR_SPACE                   "\040"
1067    #define STR_EXCLAMATION_MARK        "\041"
1068    #define STR_QUOTATION_MARK          "\042"
1069    #define STR_NUMBER_SIGN             "\043"
1070    #define STR_DOLLAR_SIGN             "\044"
1071    #define STR_PERCENT_SIGN            "\045"
1072    #define STR_AMPERSAND               "\046"
1073    #define STR_APOSTROPHE              "\047"
1074    #define STR_LEFT_PARENTHESIS        "\050"
1075    #define STR_RIGHT_PARENTHESIS       "\051"
1076    #define STR_ASTERISK                "\052"
1077    #define STR_PLUS                    "\053"
1078    #define STR_COMMA                   "\054"
1079    #define STR_MINUS                   "\055"
1080    #define STR_DOT                     "\056"
1081    #define STR_SLASH                   "\057"
1082    #define STR_0                       "\060"
1083    #define STR_1                       "\061"
1084    #define STR_2                       "\062"
1085    #define STR_3                       "\063"
1086    #define STR_4                       "\064"
1087    #define STR_5                       "\065"
1088    #define STR_6                       "\066"
1089    #define STR_7                       "\067"
1090    #define STR_8                       "\070"
1091    #define STR_9                       "\071"
1092    #define STR_COLON                   "\072"
1093    #define STR_SEMICOLON               "\073"
1094    #define STR_LESS_THAN_SIGN          "\074"
1095    #define STR_EQUALS_SIGN             "\075"
1096    #define STR_GREATER_THAN_SIGN       "\076"
1097    #define STR_QUESTION_MARK           "\077"
1098    #define STR_COMMERCIAL_AT           "\100"
1099    #define STR_A                       "\101"
1100    #define STR_B                       "\102"
1101    #define STR_C                       "\103"
1102    #define STR_D                       "\104"
1103    #define STR_E                       "\105"
1104    #define STR_F                       "\106"
1105    #define STR_G                       "\107"
1106    #define STR_H                       "\110"
1107    #define STR_I                       "\111"
1108    #define STR_J                       "\112"
1109    #define STR_K                       "\113"
1110    #define STR_L                       "\114"
1111    #define STR_M                       "\115"
1112    #define STR_N                       "\116"
1113    #define STR_O                       "\117"
1114    #define STR_P                       "\120"
1115    #define STR_Q                       "\121"
1116    #define STR_R                       "\122"
1117    #define STR_S                       "\123"
1118    #define STR_T                       "\124"
1119    #define STR_U                       "\125"
1120    #define STR_V                       "\126"
1121    #define STR_W                       "\127"
1122    #define STR_X                       "\130"
1123    #define STR_Y                       "\131"
1124    #define STR_Z                       "\132"
1125    #define STR_LEFT_SQUARE_BRACKET     "\133"
1126    #define STR_BACKSLASH               "\134"
1127    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1128    #define STR_CIRCUMFLEX_ACCENT       "\136"
1129    #define STR_UNDERSCORE              "\137"
1130    #define STR_GRAVE_ACCENT            "\140"
1131    #define STR_a                       "\141"
1132    #define STR_b                       "\142"
1133    #define STR_c                       "\143"
1134    #define STR_d                       "\144"
1135    #define STR_e                       "\145"
1136    #define STR_f                       "\146"
1137    #define STR_g                       "\147"
1138    #define STR_h                       "\150"
1139    #define STR_i                       "\151"
1140    #define STR_j                       "\152"
1141    #define STR_k                       "\153"
1142    #define STR_l                       "\154"
1143    #define STR_m                       "\155"
1144    #define STR_n                       "\156"
1145    #define STR_o                       "\157"
1146    #define STR_p                       "\160"
1147    #define STR_q                       "\161"
1148    #define STR_r                       "\162"
1149    #define STR_s                       "\163"
1150    #define STR_t                       "\164"
1151    #define STR_u                       "\165"
1152    #define STR_v                       "\166"
1153    #define STR_w                       "\167"
1154    #define STR_x                       "\170"
1155    #define STR_y                       "\171"
1156    #define STR_z                       "\172"
1157    #define STR_LEFT_CURLY_BRACKET      "\173"
1158    #define STR_VERTICAL_LINE           "\174"
1159    #define STR_RIGHT_CURLY_BRACKET     "\175"
1160    #define STR_TILDE                   "\176"
1161    
1162    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1163    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1164    #define STRING_F0                   STR_F "\0"
1165    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1166    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1167    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1168    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1169    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1170    
1171    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1172    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1173    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1174    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1175    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1176    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1177    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1178    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1179    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1180    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1181    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1182    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1183    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1184    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1185    
1186    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1187    
1188    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1189    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1190    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1191    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1192    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1193    #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
1194    #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
1195    #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1196    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1197    #define STRING_NO_START_OPT_RIGHTPAR   STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS
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 543  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  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  looked up from a table.
1262         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,  
1263         ESC_E, ESC_Q, ESC_k, ESC_REF };  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,
1271           ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1272           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 575  enum { Line 1288  enum {
1288    
1289    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1290    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1291    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1292    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1293    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1294    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1295    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1296    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1297    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1298    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1299    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1300    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */  
1301    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANY,            /* 12 Match any character except newline */
1302    OP_PROP,           /* 14 \p (Unicode property) */    OP_ALLANY,         /* 13 Match any character */
1303    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1304    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1305    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_PROP,           /* 16 \p (Unicode property) */
1306    OP_EOD,            /* 18 End of data: \z */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1307      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1308    OP_OPT,            /* 19 Set runtime options */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1309    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1310    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1311    OP_CHAR,           /* 22 Match one character, casefully */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1312    OP_CHARNC,         /* 23 Match one character, caselessly */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1313    OP_NOT,            /* 24 Match one character, not the following one */    OP_EOD,            /* 24 End of data: \z */
1314    
1315    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_CIRC,           /* 25 Start of line - not multiline */
1316    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_CIRCM,          /* 26 Start of line - multiline */
1317    OP_PLUS,           /* 27 the minimizing one second. */    OP_DOLL,           /* 27 End of line - not multiline */
1318    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_DOLLM,          /* 28 End of line - multiline */
1319    OP_QUERY,          /* 29 */    OP_CHAR,           /* 29 Match one character, casefully */
1320    OP_MINQUERY,       /* 30 */    OP_CHARI,          /* 30 Match one character, caselessly */
1321      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1322    OP_UPTO,           /* 31 From 0 to n matches */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1323    OP_MINUPTO,        /* 32 */  
1324    OP_EXACT,          /* 33 Exactly n matches */    /* The following sets of 13 opcodes must always be kept in step because
1325      the offset from the first one is used to generate the others. */
1326    OP_POSSTAR,        /* 34 Possessified star */  
1327    OP_POSPLUS,        /* 35 Possessified plus */    /**** Single characters, caseful, must precede the caseless ones ****/
1328    OP_POSQUERY,       /* 36 Posesssified query */  
1329    OP_POSUPTO,        /* 37 Possessified upto */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1330      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1331    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */    OP_PLUS,           /* 35 the minimizing one second. */
1332    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_MINPLUS,        /* 36 */
1333    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_QUERY,          /* 37 */
1334    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_MINQUERY,       /* 38 */
1335    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */  
1336    OP_NOTMINQUERY,    /* 43 */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1337      OP_MINUPTO,        /* 40 */
1338    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_EXACT,          /* 41 Exactly n matches */
1339    OP_NOTMINUPTO,     /* 45 */  
1340    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1341      OP_POSPLUS,        /* 43 Possessified plus, caseful */
1342    OP_NOTPOSSTAR,     /* 47 Possessified versions */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1343    OP_NOTPOSPLUS,     /* 48 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1344    OP_NOTPOSQUERY,    /* 49 */  
1345    OP_NOTPOSUPTO,     /* 50 */    /**** Single characters, caseless, must follow the caseful ones */
1346    
1347    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */    OP_STARI,          /* 46 */
1348    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_MINSTARI,       /* 47 */
1349    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_PLUSI,          /* 48 */
1350    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_MINPLUSI,       /* 49 */
1351    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */    OP_QUERYI,         /* 50 */
1352    OP_TYPEMINQUERY,   /* 56 */    OP_MINQUERYI,      /* 51 */
1353    
1354    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1355    OP_TYPEMINUPTO,    /* 58 */    OP_MINUPTOI,       /* 53 */
1356    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_EXACTI,         /* 54 */
1357    
1358    OP_TYPEPOSSTAR,    /* 60 Possessified versions */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1359    OP_TYPEPOSPLUS,    /* 61 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1360    OP_TYPEPOSQUERY,   /* 62 */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1361    OP_TYPEPOSUPTO,    /* 63 */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1362    
1363    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */    /**** The negated ones must follow the non-negated ones, and match them ****/
1364    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    /**** Negated single character, caseful; must precede the caseless ones ****/
1365    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */  
1366    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1367    OP_CRQUERY,        /* 68 These are for character classes and back refs */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1368    OP_CRMINQUERY,     /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1369    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1370    OP_CRMINRANGE,     /* 71 */    OP_NOTQUERY,       /* 63 */
1371      OP_NOTMINQUERY,    /* 64 */
1372    OP_CLASS,          /* 72 Match a character class, chars < 256 only */  
1373    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1374                             class - the difference is relevant only when a UTF-8    OP_NOTMINUPTO,     /* 66 */
1375                             character > 255 is encountered. */    OP_NOTEXACT,       /* 67 Exactly n matches */
1376    
1377    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1378                             class. This does both positive and negative. */    OP_NOTPOSPLUS,     /* 69 */
1379      OP_NOTPOSQUERY,    /* 70 */
1380    OP_REF,            /* 75 Match a back reference */    OP_NOTPOSUPTO,     /* 71 */
1381    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */  
1382    OP_CALLOUT,        /* 77 Call out to external function if provided */    /**** Negated single character, caseless; must follow the caseful ones ****/
1383    
1384    OP_ALT,            /* 78 Start of alternation */    OP_NOTSTARI,       /* 72 */
1385    OP_KET,            /* 79 End of group that doesn't have an unbounded repeat */    OP_NOTMINSTARI,    /* 73 */
1386    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_NOTPLUSI,       /* 74 */
1387    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_NOTMINPLUSI,    /* 75 */
1388      OP_NOTQUERYI,      /* 76 */
1389    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTMINQUERYI,   /* 77 */
1390    
1391    OP_ASSERT,         /* 82 Positive lookahead */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1392    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_NOTMINUPTOI,    /* 79 */
1393    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1394    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */  
1395    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1396      OP_NOTPOSPLUSI,    /* 82 */
1397    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_NOTPOSQUERYI,   /* 83 */
1398    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    OP_NOTPOSUPTOI,    /* 84 */
1399    
1400    OP_ONCE,           /* 87 Atomic group */    /**** Character types ****/
1401    OP_BRA,            /* 88 Start of non-capturing bracket */  
1402    OP_CBRA,           /* 89 Start of capturing bracket */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1403    OP_COND,           /* 90 Conditional group */    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1404      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1405      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1406      OP_TYPEQUERY,      /* 89 */
1407      OP_TYPEMINQUERY,   /* 90 */
1408    
1409      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1410      OP_TYPEMINUPTO,    /* 92 */
1411      OP_TYPEEXACT,      /* 93 Exactly n matches */
1412    
1413      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1414      OP_TYPEPOSPLUS,    /* 95 */
1415      OP_TYPEPOSQUERY,   /* 96 */
1416      OP_TYPEPOSUPTO,    /* 97 */
1417    
1418      /* These are used for character classes and back references; only the
1419      first six are the same as the sets above. */
1420    
1421      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1422      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1423      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1424      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1425      OP_CRQUERY,        /* 102 */
1426      OP_CRMINQUERY,     /* 103 */
1427    
1428      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1429      OP_CRMINRANGE,     /* 105 */
1430    
1431      /* End of quantifier opcodes */
1432    
1433      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1434      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1435                                  class - the difference is relevant only when a
1436                                  UTF-8 character > 255 is encountered. */
1437      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1438                                  class. This does both positive and negative. */
1439      OP_REF,            /* 109 Match a back reference, casefully */
1440      OP_REFI,           /* 110 Match a back reference, caselessly */
1441      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1442      OP_CALLOUT,        /* 112 Call out to external function if provided */
1443    
1444      OP_ALT,            /* 113 Start of alternation */
1445      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1446      OP_KETRMAX,        /* 115 These two must remain together and in this */
1447      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1448      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1449    
1450      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1451      asserts must remain in order. */
1452    
1453      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1454      OP_ASSERT,         /* 119 Positive lookahead */
1455      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1456      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1457      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1458    
1459      /* ONCE, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately after the
1460      assertions, with ONCE first, as there's a test for >= ONCE for a subpattern
1461      that isn't an assertion. The POS versions must immediately follow the non-POS
1462      versions in each case. */
1463    
1464      OP_ONCE,           /* 123 Atomic group */
1465      OP_BRA,            /* 124 Start of non-capturing bracket */
1466      OP_BRAPOS,         /* 125 Ditto, with unlimited, possessive repeat */
1467      OP_CBRA,           /* 126 Start of capturing bracket */
1468      OP_CBRAPOS,        /* 127 Ditto, with unlimited, possessive repeat */
1469      OP_COND,           /* 128 Conditional group */
1470    
1471    /* These three must follow the previous three, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
1472    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1473    
1474    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 129 Start of non-capturing bracket, check empty  */
1475    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 130 Ditto, with unlimited, possessive repeat */
1476    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCBRA,          /* 131 Start of capturing bracket, check empty */
1477      OP_SCBRAPOS,       /* 132 Ditto, with unlimited, possessive repeat */
1478    OP_CREF,           /* 94 Used to hold a capture number as condition */    OP_SCOND,          /* 133 Conditional group, check empty */
1479    OP_RREF,           /* 95 Used to hold a recursion number as condition */  
1480    OP_DEF,            /* 96 The DEFINE condition */    /* The next two pairs must (respectively) be kept together. */
1481    
1482      OP_CREF,           /* 134 Used to hold a capture number as condition */
1483      OP_NCREF,          /* 135 Same, but generated by a name reference*/
1484      OP_RREF,           /* 136 Used to hold a recursion number as condition */
1485      OP_NRREF,          /* 137 Same, but generated by a name reference*/
1486      OP_DEF,            /* 138 The DEFINE condition */
1487    
1488      OP_BRAZERO,        /* 139 These two must remain together and in this */
1489      OP_BRAMINZERO,     /* 140 order. */
1490      OP_BRAPOSZERO,     /* 141 */
1491    
1492      /* These are backtracking control verbs */
1493    
1494      OP_MARK,           /* 142 always has an argument */
1495      OP_PRUNE,          /* 143 */
1496      OP_PRUNE_ARG,      /* 144 same, but with argument */
1497      OP_SKIP,           /* 145 */
1498      OP_SKIP_ARG,       /* 146 same, but with argument */
1499      OP_THEN,           /* 147 */
1500      OP_THEN_ARG,       /* 148 same, but with argument */
1501      OP_COMMIT,         /* 149 */
1502    
1503      /* These are forced failure and success verbs */
1504    
1505      OP_FAIL,           /* 150 */
1506      OP_ACCEPT,         /* 151 */
1507      OP_ASSERT_ACCEPT,  /* 152 Used inside assertions */
1508      OP_CLOSE,          /* 153 Used before OP_ACCEPT to close open captures */
1509    
1510      /* This is used to skip a subpattern with a {0} quantifier */
1511    
1512      OP_SKIPZERO,       /* 154 */
1513    
1514      /* This is not an opcode, but is used to check that tables indexed by opcode
1515      are the correct length, in order to catch updating errors - there have been
1516      some in the past. */
1517    
1518    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_TABLE_LENGTH
   OP_BRAMINZERO      /* 98 order. */  
1519  };  };
1520    
1521    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1522    definitions that follow must also be updated to match. There are also tables
1523    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1524    
1525    
1526  /* 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
1527  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging, and some of them are only partial names. The macro is referenced
1528    only in pcre_printint.c, which fills out the full names in many cases (and in
1529    some cases doesn't actually use these names at all). */
1530    
1531  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1532    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1533    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1534    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1535    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1536    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1537    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1538      "{", "{", "{",                                                  \
1539    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1540    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1541      "{", "{", "{",                                                  \
1542      "*+","++", "?+", "{",                                           \
1543      "*", "*?", "+", "+?", "?", "??",                                \
1544      "{", "{", "{",                                                  \
1545      "*+","++", "?+", "{",                                           \
1546      "*", "*?", "+", "+?", "?", "??",                                \
1547      "{", "{", "{",                                                  \
1548    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1549    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1550    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1551    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1552    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1553    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1554    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1555    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1556    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Once",                                                         \
1557      "Bra", "BraPos", "CBra", "CBraPos",                             \
1558      "Cond",                                                         \
1559      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1560      "SCond",                                                        \
1561      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1562      "Brazero", "Braminzero", "Braposzero",                          \
1563      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1564      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1565      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1566      "Close", "Skip zero"
1567    
1568    
1569  /* 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 739  in UTF-8 mode. The code that uses this t Line 1577  in UTF-8 mode. The code that uses this t
1577    
1578  #define OP_LENGTHS \  #define OP_LENGTHS \
1579    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1580    1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1581    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1582    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1583    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    3, 3,                          /* \P, \p                                 */ \
1584      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1585      1,                             /* \X                                     */ \
1586      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1587    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1588    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1589    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1590    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1591    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1592    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1593      4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \
1594    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1595      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1596      4, 4, 4,                       /* upto I, minupto I, exact I             */ \
1597      2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \
1598    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1599    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1600    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1601    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \
1602      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1603      4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \
1604      2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \
1605    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1606    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1607    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
# Line 765  in UTF-8 mode. The code that uses this t Line 1613  in UTF-8 mode. The code that uses this t
1613   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1614    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1615    3,                             /* REF                                    */ \    3,                             /* REF                                    */ \
1616      3,                             /* REFI                                   */ \
1617    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1618    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1619    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1620    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1621    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1622    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1623      1+LINK_SIZE,                   /* KetRpos                                */ \
1624      1+LINK_SIZE,                   /* Reverse                                */ \
1625    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1626    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1627    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1628    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1629    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1630    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1631      1+LINK_SIZE,                   /* BRAPOS                                 */ \
1632    3+LINK_SIZE,                   /* CBRA                                   */ \    3+LINK_SIZE,                   /* CBRA                                   */ \
1633      3+LINK_SIZE,                   /* CBRAPOS                                */ \
1634    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1635    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1636      1+LINK_SIZE,                   /* SBRAPOS                                */ \
1637    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1638      3+LINK_SIZE,                   /* SCBRAPOS                               */ \
1639    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1640    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1641    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1642    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1643    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1644      3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1645      1, 3,                          /* SKIP, SKIP_ARG                         */ \
1646      1, 3,                          /* THEN, THEN_ARG                         */ \
1647      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1648      3, 1                           /* CLOSE, SKIPZERO  */
1649    
1650  /* 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"
1651    condition. */
1652    
1653  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1654    
1655  /* 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
1656  tracked. */  easily be tracked. When a new number is added, the table called eint in
1657    pcreposix.c must be updated. */
1658    
1659  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1660         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1661         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1662         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1663         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1664         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1665           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1666           ERRCOUNT };
1667    
1668  /* 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
1669  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 817  Because people can now save and re-use c Line 1679  Because people can now save and re-use c
1679  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
1680  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
1681  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1682  NOTE NOTE NOTE:  NOTE NOTE NOTE
1683  */  */
1684    
1685  typedef struct real_pcre {  typedef struct real_pcre {
1686    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1687    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1688    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1689    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1690      pcre_uint16 dummy1;             /* For future use */
1691    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1692    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1693    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 844  remark (see NOTE above) about extending Line 1706  remark (see NOTE above) about extending
1706    
1707  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1708    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1709    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1710    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1711      pcre_uint32 minlength;          /* Minimum subject length */
1712  } pcre_study_data;  } pcre_study_data;
1713    
1714    /* Structure for building a chain of open capturing subpatterns during
1715    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1716    encountered. This is also used to identify subpatterns that contain recursive
1717    back references to themselves, so that they can be made atomic. */
1718    
1719    typedef struct open_capitem {
1720      struct open_capitem *next;    /* Chain link */
1721      pcre_uint16 number;           /* Capture number */
1722      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1723    } open_capitem;
1724    
1725  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1726  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1727    
# Line 860  typedef struct compile_data { Line 1734  typedef struct compile_data {
1734    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1735    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1736    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1737      open_capitem *open_caps;      /* Chain of open capture items */
1738    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1739    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1740    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1741    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1742    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
1743      int  final_bracount;          /* Saved value after first pass */
1744    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1745    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1746      int  assert_depth;            /* Depth of nested assertions */
1747    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1748      int  external_flags;          /* External flag bits to be set */
1749    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1750    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
1751      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1752    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1753    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1754    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
1755  } compile_data;  } compile_data;
1756    
1757  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1758  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1759    
1760  typedef struct branch_chain {  typedef struct branch_chain {
1761    struct branch_chain *outer;    struct branch_chain *outer;
1762    uschar *current;    uschar *current_branch;
1763  } branch_chain;  } branch_chain;
1764    
1765  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
1766  call within the pattern. */  call within the pattern; used by pcre_exec(). */
1767    
1768  typedef struct recursion_info {  typedef struct recursion_info {
1769    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1770    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
1771    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
1772    USPTR save_start;             /* Old value of md->start_match */    int saved_max;                  /* Number of saved offsets */
1773    int *offset_save;             /* Pointer to start of saved offsets */    USPTR subject_position;         /* Position at start of recursion */
   int saved_max;                /* Number of saved offsets */  
1774  } recursion_info;  } recursion_info;
1775    
1776  /* When compiling in a mode that doesn't use recursive calls to match(),  /* A similar structure for pcre_dfa_exec(). */
 a structure is used to remember local variables on the heap. It is defined in  
 pcre_exec.c, close to the match() function, so that it is easy to keep it in  
 step with any changes of local variable. However, the pointer to the current  
 frame must be saved in some "static" place over a longjmp(). We declare the  
 structure here so that we can put a pointer in the match_data structure. NOTE:  
 This isn't used for a "normal" compilation of pcre. */  
1777    
1778  struct heapframe;  typedef struct dfa_recursion_info {
1779      struct dfa_recursion_info *prevrec;
1780      int group_num;
1781      USPTR subject_position;
1782    } dfa_recursion_info;
1783    
1784  /* 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
1785  pointer at the start of each subpattern, so as to detect when an empty string  pointer at the start of each subpattern, so as to detect when an empty string
1786  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
1787    pcre_exec(). */
1788    
1789  typedef struct eptrblock {  typedef struct eptrblock {
1790    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
# Line 927  typedef struct match_data { Line 1804  typedef struct match_data {
1804    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1805    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1806    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1807      int    name_count;            /* Number of names in name table */
1808      int    name_entry_size;       /* Size of entry in names table */
1809      uschar *name_table;           /* Table of names */
1810    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1811    const uschar *lcc;            /* Points to lower casing table */    const  uschar *lcc;           /* Points to lower casing table */
1812    const uschar *ctypes;         /* Points to table of type maps */    const  uschar *ctypes;        /* Points to table of type maps */
1813    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1814    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1815    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1816    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1817      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1818      BOOL   use_ucp;               /* PCRE_UCP flag */
1819    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1820    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1821    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1822    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1823    const uschar *start_code;     /* For use when recursing */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1824      BOOL   hasthen;               /* Pattern contains (*THEN) */
1825      const  uschar *start_code;    /* For use when recursing */
1826    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1827    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1828    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1829    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1830      USPTR  start_used_ptr;        /* Earliest consulted character */
1831      int    partial;               /* PARTIAL options */
1832    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1833    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1834    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1835      int    match_function_type;   /* Set for certain special calls of MATCH() */
1836    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1837    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1838    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1839    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1840    struct heapframe *thisframe;  /* Used only when compiling for no recursion */    const  uschar *mark;          /* Mark pointer to pass back */
1841      const  uschar *once_target;   /* Where to back up to for atomic groups */
1842  } match_data;  } match_data;
1843    
1844  /* 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
1845  functions. */  functions. */
1846    
1847  typedef struct dfa_match_data {  typedef struct dfa_match_data {
1848    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;      /* Start of the compiled pattern */
1849    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;   /* Start of the subject string */
1850    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;     /* End of subject string */
1851    const uschar *tables;         /* Character tables */    const uschar *start_used_ptr;  /* Earliest consulted character */
1852    int   moptions;               /* Match options */    const uschar *tables;          /* Character tables */
1853    int   poptions;               /* Pattern options */    int   start_offset;            /* The start offset value */
1854    int    nltype;                /* Newline type */    int   moptions;                /* Match options */
1855    int    nllen;                 /* Newline string length */    int   poptions;                /* Pattern options */
1856    uschar nl[4];                 /* Newline string when fixed */    int    nltype;                 /* Newline type */
1857    void  *callout_data;          /* To pass back to callouts */    int    nllen;                  /* Newline string length */
1858      uschar nl[4];                  /* Newline string when fixed */
1859      void  *callout_data;           /* To pass back to callouts */
1860      dfa_recursion_info *recursive; /* Linked list of recursion data */
1861  } dfa_match_data;  } dfa_match_data;
1862    
1863  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 975  typedef struct dfa_match_data { Line 1866  typedef struct dfa_match_data {
1866  #define ctype_letter  0x02  #define ctype_letter  0x02
1867  #define ctype_digit   0x04  #define ctype_digit   0x04
1868  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1869  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1870  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1871    
1872  /* 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 1003  total length. */ Line 1894  total length. */
1894  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1895    
1896  /* 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
1897  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1898    relocations in shared libraries, it now has an offset into a single string
1899    instead. */
1900    
1901  typedef struct {  typedef struct {
1902    const char *name;    pcre_uint16 name_offset;
1903    pcre_uint16 type;    pcre_uint16 type;
1904    pcre_uint16 value;    pcre_uint16 value;
1905  } ucp_type_table;  } ucp_type_table;
# Line 1022  extern const int _pcre_utf8_table2[]; Line 1915  extern const int _pcre_utf8_table2[];
1915  extern const int    _pcre_utf8_table3[];  extern const int    _pcre_utf8_table3[];
1916  extern const uschar _pcre_utf8_table4[];  extern const uschar _pcre_utf8_table4[];
1917    
1918    #ifdef SUPPORT_JIT
1919    extern const uschar _pcre_utf8_char_sizes[];
1920    #endif
1921    
1922  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1923    
1924    extern const char   _pcre_utt_names[];
1925  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1926  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1927    
# Line 1036  extern const uschar _pcre_OP_lengths[]; Line 1934  extern const uschar _pcre_OP_lengths[];
1934  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
1935  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1936    
1937  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1938                        BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1939  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1940  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1941                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1942  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  extern int           _pcre_valid_utf8(USPTR, int, int *);
1943  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1944  extern int          _pcre_valid_utf8(const uschar *, int);  extern BOOL          _pcre_xclass(int, const uschar *);
1945  extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,  
1946                        BOOL);  #ifdef SUPPORT_JIT
1947  extern BOOL         _pcre_xclass(int, const uschar *);  extern void          _pcre_jit_compile(const real_pcre *, pcre_extra *);
1948    extern int           _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR,
1949                            int, int, int, int, int *, int);
1950    extern void          _pcre_jit_free(void *);
1951    #endif
1952    
1953    /* Unicode character database (UCD) */
1954    
1955    typedef struct {
1956      uschar script;
1957      uschar chartype;
1958      pcre_int32 other_case;
1959    } ucd_record;
1960    
1961    extern const ucd_record  _pcre_ucd_records[];
1962    extern const uschar      _pcre_ucd_stage1[];
1963    extern const pcre_uint16 _pcre_ucd_stage2[];
1964    extern const int         _pcre_ucp_gentype[];
1965    #ifdef SUPPORT_JIT
1966    extern const int         _pcre_ucp_typerange[];
1967    #endif
1968    
1969    /* UCD access macros */
1970    
1971    #define UCD_BLOCK_SIZE 128
1972    #define GET_UCD(ch) (_pcre_ucd_records + \
1973            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1974            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
1975    
1976    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1977    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1978    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1979    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1980    
1981  #endif  #endif
1982    

Legend:
Removed from v.118  
changed lines
  Added in v.716

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12