/[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 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 208 by ph10, Mon Aug 6 15:23:29 2007 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) 1997-2006 University of Cambridge             Copyright (c) 1997-2007 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 54  functions whose names all begin with "_p Line 54  functions whose names all begin with "_p
54  /* 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
55  inline, and there are *still* stupid compilers about that don't like indented  inline, and there are *still* stupid compilers about that don't like indented
56  pre-processor statements, or at least there were when I first wrote this. After  pre-processor statements, or at least there were when I first wrote this. After
57  all, it had only been about 10 years then... */  all, it had only been about 10 years then...
58    
59    It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
60    be absolutely sure we get our version. */
61    
62    #undef DPRINTF
63  #ifdef DEBUG  #ifdef DEBUG
64  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
65  #else  #else
66  #define DPRINTF(p) /*nothing*/  #define DPRINTF(p) /* Nothing */
67  #endif  #endif
68    
69    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
70  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
71  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
72    
# Line 79  setjmp and stdarg are used is when NO_RE Line 79  setjmp and stdarg are used is when NO_RE
79  #include <stdlib.h>  #include <stdlib.h>
80  #include <string.h>  #include <string.h>
81    
82  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
83  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
84    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
85    information there, using __declspec(dllexport) without "extern" we have a
86    definition; with "extern" we have a declaration. The settings here override the
87    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
88    which is all that is needed for applications (they just import the symbols). We
89    use:
90    
91      PCRE_EXP_DECL       for declarations
92      PCRE_EXP_DEFN       for definitions of exported functions
93      PCRE_EXP_DATA_DEFN  for definitions of exported variables
94    
95    The reason for the two DEFN macros is that in non-Windows environments, one
96    does not want to have "extern" before variable definitions because it leads to
97    compiler warnings. So we distinguish between functions and variables. In
98    Windows, the two should always be the same.
99    
100    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
101    which is an application, but needs to import this file in order to "peek" at
102    internals, can #include pcre.h first to get an application's-eye view.
103    
104    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
105    special-purpose environments) might want to stick other stuff in front of
106    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
107    PCRE_EXP_DATA_DEFN only if they are not already set. */
108    
109    #ifndef PCRE_EXP_DECL
110    #  ifdef _WIN32
111    #    ifndef PCRE_STATIC
112    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
113    #      define PCRE_EXP_DEFN       __declspec(dllexport)
114    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
115    #    else
116    #      define PCRE_EXP_DECL       extern
117    #      define PCRE_EXP_DEFN
118    #      define PCRE_EXP_DATA_DEFN
119    #    endif
120    #  else
121    #    ifdef __cplusplus
122    #      define PCRE_EXP_DECL       extern "C"
123    #    else
124    #      define PCRE_EXP_DECL       extern
125    #    endif
126    #    ifndef PCRE_EXP_DEFN
127    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
128    #    endif
129    #    ifndef PCRE_EXP_DATA_DEFN
130    #      define PCRE_EXP_DATA_DEFN
131    #    endif
132    #  endif
133  #endif  #endif
134    
135  /* 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 114  Unix, where it is defined in sys/types, Line 163  Unix, where it is defined in sys/types,
163    
164  typedef unsigned char uschar;  typedef unsigned char uschar;
165    
166    /* This is an unsigned int value that no character can ever have. UTF-8
167    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
168    0x0010ffff). */
169    
170    #define NOTACHAR 0xffffffff
171    
172    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
173    "any" and "anycrlf" at present). The following macros are used to package up
174    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
175    modules to indicate in which datablock the parameters exist, and what the
176    start/end of string field names are. */
177    
178    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
179    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
180    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
181    
182    /* This macro checks for a newline at the given position */
183    
184    #define IS_NEWLINE(p) \
185      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
186        ((p) < NLBLOCK->PSEND && \
187         _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
188           utf8)) \
189        : \
190        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
191         (p)[0] == NLBLOCK->nl[0] && \
192         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
193        ) \
194      )
195    
196    /* This macro checks for a newline immediately preceding the given position */
197    
198    #define WAS_NEWLINE(p) \
199      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
200        ((p) > NLBLOCK->PSSTART && \
201         _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
202           &(NLBLOCK->nllen), utf8)) \
203        : \
204        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
205         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
206         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
207        ) \
208      )
209    
210  /* When PCRE is compiled as a C++ library, the subject pointer can be replaced  /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
211  with a custom type. This makes it possible, for example, to allow pcre_exec()  with a custom type. This makes it possible, for example, to allow pcre_exec()
212  to process subject strings that are discontinuous by using a smart pointer  to process subject strings that are discontinuous by using a smart pointer
# Line 131  must begin with PCRE_. */ Line 224  must begin with PCRE_. */
224  #define USPTR const unsigned char *  #define USPTR const unsigned char *
225  #endif  #endif
226    
227    
228    
229  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
230  values. */  values. */
231    
# Line 142  need to have their names changed. PCRE m Line 237  need to have their names changed. PCRE m
237  option on the command line. */  option on the command line. */
238    
239  #ifdef VPCOMPAT  #ifdef VPCOMPAT
240    #define strlen(s)        _strlen(s)
241  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
242    #define memcmp(s,c,n)    _memcmp(s,c,n)
243  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
244  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
245  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 151  option on the command line. */ Line 248  option on the command line. */
248  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
249  define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY  define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
250  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
251  neither (there some non-Unix environments where this is the case). This assumes  neither (there some non-Unix environments where this is the case). */
 that all calls to memmove are moving strings upwards in store, which is the  
 case in PCRE. */  
252    
253  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
254  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
255  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
256  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
257  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
258  void *  static void *
259  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
260  {  {
261  size_t i;  size_t i;
262  dest += n;  unsigned char *dest = (unsigned char *)d;
263  src += n;  const unsigned char *src = (const unsigned char *)s;
264  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
265  return dest;    {
266      dest += n;
267      src += n;
268      for (i = 0; i < n; ++i) *(--dest) = *(--src);
269      return (void *)dest;
270      }
271    else
272      {
273      for (i = 0; i < n; ++i) *dest++ = *src++;
274      return (void *)(dest - n);
275      }
276  }  }
277  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
278  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 253  capturing parenthesis numbers in back re Line 358  capturing parenthesis numbers in back re
358    
359  /* 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
360  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
361  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
362    never be called in byte mode. To make sure it can never even appear when UTF-8
363    support is omitted, we don't even define it. */
364    
365  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
366  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 261  byte-mode, and more complicated ones for Line 368  byte-mode, and more complicated ones for
368  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
369  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
370  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
371  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
372    
373  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
374    
# Line 270  we know we are in UTF-8 mode. */ Line 377  we know we are in UTF-8 mode. */
377    
378  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
379    c = *eptr; \    c = *eptr; \
380    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
381      { \      { \
382      int gcii; \      int gcii; \
383      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 288  pointer. */ Line 395  pointer. */
395    
396  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
397    c = *eptr; \    c = *eptr; \
398    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
399      { \      { \
400      int gcii; \      int gcii; \
401      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 306  know we are in UTF-8 mode. */ Line 413  know we are in UTF-8 mode. */
413    
414  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
415    c = *eptr++; \    c = *eptr++; \
416    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
417      { \      { \
418      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
419      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 322  know we are in UTF-8 mode. */ Line 429  know we are in UTF-8 mode. */
429    
430  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
431    c = *eptr++; \    c = *eptr++; \
432    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
433      { \      { \
434      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
435      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 339  if there are extra bytes. This is called Line 446  if there are extra bytes. This is called
446    
447  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
448    c = *eptr; \    c = *eptr; \
449    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
450      { \      { \
451      int gcii; \      int gcii; \
452      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 354  if there are extra bytes. This is called Line 461  if there are extra bytes. This is called
461      }      }
462    
463  /* 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
464  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
465    because almost all calls are already within a block of UTF-8 only code. */
466    
467  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
468    
469  #endif  #endif
470    
# Line 373  Standard C system should have one. */ Line 481  Standard C system should have one. */
481    
482  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
483    
484  /* Private options flags start at the most significant end of the four bytes,  /* Private options flags start at the most significant end of the four bytes.
485  but skip the top bit so we can use ints for convenience without getting tangled  The public options defined in pcre.h start at the least significant end. Make
486  with negative values. The public options defined in pcre.h start at the least  sure they don't overlap! The bits are getting a bit scarce now -- when we run
487  significant end. Make sure they don't overlap! */  out, there is a dummy word in the structure that could be used for the private
488    bits. */
489    
490    #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */
491  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */
492  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */
493  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
494  #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */
 #define PCRE_NOPARTIAL     0x04000000  /* can't use partial with this regex */  
495    
496  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
497    
# Line 391  significant end. Make sure they don't ov Line 500  significant end. Make sure they don't ov
500  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
501  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
502    
503    #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
504                               PCRE_NEWLINE_ANYCRLF)
505    
506  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
507    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
508     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
509     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| \
510       PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
511    
512  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
513    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
514     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS)
515    
516  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
517    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
518     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
519    
520  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
521    
# Line 434  typedef int BOOL; Line 547  typedef int BOOL;
547  #define FALSE   0  #define FALSE   0
548  #define TRUE    1  #define TRUE    1
549    
550  /* Escape items that are just an encoding of a particular data value. Note that  /* Escape items that are just an encoding of a particular data value. */
 ESC_n is defined as yet another macro, which is set in config.h to either \n  
 (the default) or \r (which some people want). */  
551    
552  #ifndef ESC_e  #ifndef ESC_e
553  #define ESC_e 27  #define ESC_e 27
# Line 447  ESC_n is defined as yet another macro, w Line 558  ESC_n is defined as yet another macro, w
558  #endif  #endif
559    
560  #ifndef ESC_n  #ifndef ESC_n
561  #define ESC_n NEWLINE  #define ESC_n '\n'
562  #endif  #endif
563    
564  #ifndef ESC_r  #ifndef ESC_r
# Line 486  value such as \n. They must have non-zer Line 597  value such as \n. They must have non-zer
597  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
598  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_ANY because it
599  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence. The final one must be
600  ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
601  tests in the code for an escape greater than ESC_b and less than ESC_Z to  There are two tests in the code for an escape greater than ESC_b and less than
602  detect the types that may be repeated. These are the types that consume  ESC_Z to detect the types that may be repeated. These are the types that
603  characters. If any new escapes are put in between that don't consume a  consume characters. If any new escapes are put in between that don't consume a
604  character, that code will have to change. */  character, that code will have to change. */
605    
606  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
607         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,
608         ESC_Q, ESC_REF };         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };
609    
610    
611  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
612  that extract substrings. 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
613  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.
614  Note that whenever this list is updated, the two macro definitions that follow  
615  must also be updated to match. */  To keep stored, compiled patterns compatible, new opcodes should be added
616    immediately before OP_BRA, where (since release 7.0) a gap is left for this
617    purpose.
618    
619    *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
620    that follow must also be updated to match. There is also a table called
621    "coptable" in pcre_dfa_exec.c that must be updated. */
622    
623  enum {  enum {
624    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 509  enum { Line 627  enum {
627    
628    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
629    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
630    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
631    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
632    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
633    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
634    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
635    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
636    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
637    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
638    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
639    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character */
640    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */
641    OP_PROP,           /* 14 \p (Unicode property) */    OP_NOTPROP,        /* 14 \P (not Unicode property) */
642    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_PROP,           /* 15 \p (Unicode property) */
643    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_ANYNL,          /* 16 \R (any newline sequence) */
644    OP_EOD,            /* 17 End of data: \z */    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */
645      OP_HSPACE,         /* 18 \h (horizontal whitespace) */
646    OP_OPT,            /* 18 Set runtime options */    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */
647    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_VSPACE,         /* 20 \v (vertical whitespace) */
648    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */
649    OP_CHAR,           /* 21 Match one character, casefully */    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */
650    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EOD,            /* 23 End of data: \z */
651    OP_NOT,            /* 23 Match anything but the following char */  
652      OP_OPT,            /* 24 Set runtime options */
653    OP_STAR,           /* 24 The maximizing and minimizing versions of */    OP_CIRC,           /* 25 Start of line - varies with multiline switch */
654    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_DOLL,           /* 26 End of line - varies with multiline switch */
655    OP_PLUS,           /* 26 the minimizing one second. */    OP_CHAR,           /* 27 Match one character, casefully */
656    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_CHARNC,         /* 28 Match one character, caselessly */
657    OP_QUERY,          /* 28 */    OP_NOT,            /* 29 Match one character, not the following one */
658    OP_MINQUERY,       /* 29 */  
659    OP_UPTO,           /* 30 From 0 to n matches */    OP_STAR,           /* 30 The maximizing and minimizing versions of */
660    OP_MINUPTO,        /* 31 */    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */
661    OP_EXACT,          /* 32 Exactly n matches */    OP_PLUS,           /* 32 the minimizing one second. */
662      OP_MINPLUS,        /* 33 This first set applies to single characters.*/
663    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_QUERY,          /* 34 */
664    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    OP_MINQUERY,       /* 35 */
665    OP_NOTPLUS,        /* 35 the minimizing one second. */  
666    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */    OP_UPTO,           /* 36 From 0 to n matches */
667    OP_NOTQUERY,       /* 37 */    OP_MINUPTO,        /* 37 */
668    OP_NOTMINQUERY,    /* 38 */    OP_EXACT,          /* 38 Exactly n matches */
669    OP_NOTUPTO,        /* 39 From 0 to n matches */  
670    OP_NOTMINUPTO,     /* 40 */    OP_POSSTAR,        /* 39 Possessified star */
671    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_POSPLUS,        /* 40 Possessified plus */
672      OP_POSQUERY,       /* 41 Posesssified query */
673    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_POSUPTO,        /* 42 Possessified upto */
674    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */  
675    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */
676    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */
677    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */
678    OP_TYPEMINQUERY,   /* 47 */    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */
679    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */
680    OP_TYPEMINUPTO,    /* 49 */    OP_NOTMINQUERY,    /* 48 */
681    OP_TYPEEXACT,      /* 50 Exactly n matches */  
682      OP_NOTUPTO,        /* 49 From 0 to n matches */
683    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTMINUPTO,     /* 50 */
684    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTEXACT,       /* 51 Exactly n matches */
685    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */  
686    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */    OP_NOTPOSSTAR,     /* 52 Possessified versions */
687    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_NOTPOSPLUS,     /* 53 */
688    OP_CRMINQUERY,     /* 56 */    OP_NOTPOSQUERY,    /* 54 */
689    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_NOTPOSUPTO,     /* 55 */
690    OP_CRMINRANGE,     /* 58 */  
691      OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */
692      OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */
693      OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */
694      OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */
695      OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */
696      OP_TYPEMINQUERY,   /* 61 */
697    
698      OP_TYPEUPTO,       /* 62 From 0 to n matches */
699      OP_TYPEMINUPTO,    /* 63 */
700      OP_TYPEEXACT,      /* 64 Exactly n matches */
701    
702      OP_TYPEPOSSTAR,    /* 65 Possessified versions */
703      OP_TYPEPOSPLUS,    /* 66 */
704      OP_TYPEPOSQUERY,   /* 67 */
705      OP_TYPEPOSUPTO,    /* 68 */
706    
707      OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */
708      OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */
709      OP_CRPLUS,         /* 71 the minimizing one second. These codes must */
710      OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */
711      OP_CRQUERY,        /* 73 These are for character classes and back refs */
712      OP_CRMINQUERY,     /* 74 */
713      OP_CRRANGE,        /* 75 These are different to the three sets above. */
714      OP_CRMINRANGE,     /* 76 */
715    
716    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 77 Match a character class, chars < 256 only */
717    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative
718                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
719                             character > 255 is encountered. */                             character > 255 is encountered. */
720    
721    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the
722                             class. This does both positive and negative. */                             class. This does both positive and negative. */
723    
724    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 80 Match a back reference */
725    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */
726    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 82 Call out to external function if provided */
727    
728    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 83 Start of alternation */
729    OP_KET,            /* 66 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */
730    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 85 These two must remain together and in this */
731    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */
732    
733    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
734    
735    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 87 Positive lookahead */
736    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 88 Negative lookahead */
737    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 89 Positive lookbehind */
738    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */
739    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */
740    
741    /* ONCE and COND must come after the assertions, with ONCE first, as there's    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
742    a test for >= ONCE for a subpattern that isn't an assertion. */    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
743    
744    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */    OP_ONCE,           /* 92 Atomic group */
745    OP_COND,           /* 75 Conditional group */    OP_BRA,            /* 93 Start of non-capturing bracket */
746    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */    OP_CBRA,           /* 94 Start of capturing bracket */
747      OP_COND,           /* 95 Conditional group */
748    OP_BRAZERO,        /* 77 These two must remain together and in this */  
749    OP_BRAMINZERO,     /* 78 order. */    /* These three must follow the previous three, in the same order. There's a
750      check for >= SBRA to distinguish the two sets. */
751    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
752                             than can fit into an opcode. */    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */
753      OP_SCBRA,          /* 97 Start of capturing bracket, check empty */
754    OP_BRA             /* 80 This and greater values are used for brackets that    OP_SCOND,          /* 98 Conditional group, check empty */
755                             extract substrings up to EXTRACT_BASIC_MAX. After  
756                             that, use is made of OP_BRANUMBER. */    OP_CREF,           /* 99 Used to hold a capture number as condition */
757  };    OP_RREF,           /* 100 Used to hold a recursion number as condition */
758      OP_DEF,            /* 101 The DEFINE condition */
 /* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and  
 study.c that all opcodes are less than 128 in value. This makes handling UTF-8  
 character sequences easier. */  
   
 /* The highest extraction number before we have to start using additional  
 bytes. (Originally PCRE didn't have support for extraction counts highter than  
 this number.) The value is limited by the number of opcodes left after OP_BRA,  
 i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional  
 opcodes. */  
759    
760  #define EXTRACT_BASIC_MAX  100    OP_BRAZERO,        /* 102 These two must remain together and in this */
761      OP_BRAMINZERO      /* 103 order. */
762    };
763    
764    
765  /* 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
766  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
767    
768  #define OP_NAME_LIST \  #define OP_NAME_LIST \
769    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
770    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
771    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
772    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
773    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
774    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
775      "*+","++", "?+", "{",                                           \
776    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
777      "*+","++", "?+", "{",                                           \
778    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
779      "*+","++", "?+", "{",                                           \
780    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
781    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
782    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
783    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
784    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
785      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
786    
787    
788  /* 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 657  in UTF-8 mode. The code that uses this t Line 796  in UTF-8 mode. The code that uses this t
796    
797  #define OP_LENGTHS \  #define OP_LENGTHS \
798    1,                             /* End                                    */ \    1,                             /* End                                    */ \
799    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                     */ \
800      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
801    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
802    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
803      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
804    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
805    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
806    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 667  in UTF-8 mode. The code that uses this t Line 808  in UTF-8 mode. The code that uses this t
808    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
809    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
810    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
811      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
812    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
813    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
814    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
815      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
816    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
817    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
818    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
819      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
820    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
821    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
822    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 691  in UTF-8 mode. The code that uses this t Line 835  in UTF-8 mode. The code that uses this t
835    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
836    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
837    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
838    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
839      1+LINK_SIZE,                   /* BRA                                    */ \
840      3+LINK_SIZE,                   /* CBRA                                   */ \
841    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
842      1+LINK_SIZE,                   /* SBRA                                   */ \
843      3+LINK_SIZE,                   /* SCBRA                                  */ \
844      1+LINK_SIZE,                   /* SCOND                                  */ \
845    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
846      3,                             /* RREF                                   */ \
847      1,                             /* DEF                                    */ \
848    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
   3,                             /* BRANUMBER                              */ \  
   1+LINK_SIZE                    /* BRA                                    */ \  
849    
850    
851  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
852    
853  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
854    
855  /* Error code numbers. They are given names so that they can more easily be  /* Error code numbers. They are given names so that they can more easily be
856  tracked. */  tracked. */
# Line 710  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 859  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
859         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
860         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
861         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
862         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
863           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };
864    
865  /* 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
866  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 765  typedef struct compile_data { Line 915  typedef struct compile_data {
915    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
916    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
917    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
918      const uschar *start_workspace;/* The start of working space */
919    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
920    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
921      const uschar *end_pattern;    /* The end of the pattern */
922      uschar *hwm;                  /* High watermark of workspace */
923    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
924    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
925    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
926      int  bracount;                /* Count of capturing parens */
927    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
928    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
929      int  external_options;        /* External (initial) options */
930    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
931    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL nopartial;               /* Set TRUE if partial won't work */
932      int  nltype;                  /* Newline type */
933      int  nllen;                   /* Newline string length */
934      uschar nl[4];                 /* Newline string when fixed length */
935  } compile_data;  } compile_data;
936    
937  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 791  typedef struct recursion_info { Line 949  typedef struct recursion_info {
949    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
950    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
951    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
952    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
953    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
954    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
955  } recursion_info;  } recursion_info;
956    
957  /* When compiling in a mode that doesn't use recursive calls to match(),  /* Structure for building a chain of data for holding the values of the subject
958  a structure is used to remember local variables on the heap. It is defined in  pointer at the start of each subpattern, so as to detect when an empty string
959  pcre.c, close to the match() function, so that it is easy to keep it in step  has been matched by a subpattern - to break infinite loops. */
960  with any changes of local variable. However, the pointer to the current frame  
961  must be saved in some "static" place over a longjmp(). We declare the  typedef struct eptrblock {
962  structure here so that we can put a pointer in the match_data structure.    struct eptrblock *epb_prev;
963  NOTE: This isn't used for a "normal" compilation of pcre. */    USPTR epb_saved_eptr;
964    } eptrblock;
965    
 struct heapframe;  
966    
967  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
968  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
# Line 816  typedef struct match_data { Line 974  typedef struct match_data {
974    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
975    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
976    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
977      int    nltype;                /* Newline type */
978      int    nllen;                 /* Newline string length */
979      uschar nl[4];                 /* Newline string when fixed */
980    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
981    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
982    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
# Line 829  typedef struct match_data { Line 990  typedef struct match_data {
990    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
991    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
992    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
993    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
994    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
995    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
996    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
997    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
998      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
999      int    eptrn;                 /* Next free eptrblock */
1000    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1001    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
   struct heapframe *thisframe;  /* Used only when compiling for no recursion */  
1002  } match_data;  } match_data;
1003    
1004  /* A similar structure is used for the same purpose by the DFA matching  /* A similar structure is used for the same purpose by the DFA matching
# Line 849  typedef struct dfa_match_data { Line 1011  typedef struct dfa_match_data {
1011    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1012    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1013    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1014      int    nltype;                /* Newline type */
1015      int    nllen;                 /* Newline string length */
1016      uschar nl[4];                 /* Newline string when fixed */
1017    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1018  } dfa_match_data;  } dfa_match_data;
1019    
# Line 919  extern const uschar _pcre_OP_lengths[]; Line 1084  extern const uschar _pcre_OP_lengths[];
1084  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
1085  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1086    
1087  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1088  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        int *, BOOL);
1089                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1090  extern int         _pcre_ucp_findprop(const int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1091  extern int         _pcre_ucp_othercase(const int);                        const pcre_study_data *, pcre_study_data *);
1092  extern int         _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);
1093  extern BOOL        _pcre_xclass(int, const uschar *);  extern unsigned int _pcre_ucp_othercase(const unsigned int);
1094    extern int          _pcre_valid_utf8(const uschar *, int);
1095    extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1096                          int *, BOOL);
1097    extern BOOL         _pcre_xclass(int, const uschar *);
1098    
1099  #endif  #endif
1100    

Legend:
Removed from v.87  
changed lines
  Added in v.208

  ViewVC Help
Powered by ViewVC 1.1.5