/[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 77 by nigel, Sat Feb 24 21:40:45 2007 UTC revision 189 by ph10, Thu Jul 5 11:50:49 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) 1997-2005 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 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  modules, but which are not relevant to the exported API. This includes some  modules, but which are not relevant to the exported API. This includes some
43  functions whose names all begin with "_pcre_". */  functions whose names all begin with "_pcre_". */
44    
45    #ifndef PCRE_INTERNAL_H
46    #define PCRE_INTERNAL_H
47    
48  /* Define DEBUG to get debugging output on stdout. */  /* Define DEBUG to get debugging output on stdout. */
49    
50  /****  #if 0
51  #define DEBUG  #define DEBUG
52  ****/  #endif
53    
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    
# Line 77  setjmp and stdarg are used is when NO_RE Line 83  setjmp and stdarg are used is when NO_RE
83  #include <stdlib.h>  #include <stdlib.h>
84  #include <string.h>  #include <string.h>
85    
86  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
87  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
88    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
89    information there, using __declspec(dllexport) without "extern" we have a
90    definition; with "extern" we have a declaration. The settings here override the
91    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
92    which is all that is needed for applications (they just import the symbols). We
93    use:
94    
95      PCRE_EXP_DECL       for declarations
96      PCRE_EXP_DEFN       for definitions of exported functions
97      PCRE_EXP_DATA_DEFN  for definitions of exported variables
98    
99    The reason for the two DEFN macros is that in non-Windows environments, one
100    does not want to have "extern" before variable definitions because it leads to
101    compiler warnings. So we distinguish between functions and variables. In
102    Windows, the two should always be the same.
103    
104    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
105    which is an application, but needs to import this file in order to "peek" at
106    internals, can #include pcre.h first to get an application's-eye view.
107    
108    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
109    special-purpose environments) might want to stick other stuff in front of
110    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
111    PCRE_EXP_DATA_DEFN only if they are not already set. */
112    
113    #ifndef PCRE_EXP_DECL
114    #  ifdef _WIN32
115    #    ifdef DLL_EXPORT
116    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
117    #      define PCRE_EXP_DEFN       __declspec(dllexport)
118    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
119    #    else
120    #      define PCRE_EXP_DECL       extern
121    #      define PCRE_EXP_DEFN
122    #      define PCRE_EXP_DATA_DEFN
123    #    endif
124    #
125    #  else
126    #    ifdef __cplusplus
127    #      define PCRE_EXP_DECL       extern "C"
128    #    else
129    #      define PCRE_EXP_DECL       extern
130    #    endif
131    #    ifndef PCRE_EXP_DEFN
132    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
133    #    endif
134    #    ifndef PCRE_EXP_DATA_DEFN
135    #      define PCRE_EXP_DATA_DEFN
136    #    endif
137    #  endif
138  #endif  #endif
139    
140  /* 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 112  Unix, where it is defined in sys/types, Line 168  Unix, where it is defined in sys/types,
168    
169  typedef unsigned char uschar;  typedef unsigned char uschar;
170    
171  /* Include the public PCRE header */  /* This is an unsigned int value that no character can ever have. UTF-8
172    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
173    0x0010ffff). */
174    
175    #define NOTACHAR 0xffffffff
176    
177    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
178    "any" and "anycrlf" at present). The following macros are used to package up
179    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
180    modules to indicate in which datablock the parameters exist, and what the
181    start/end of string field names are. */
182    
183    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
184    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
185    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
186    
187    /* This macro checks for a newline at the given position */
188    
189    #define IS_NEWLINE(p) \
190      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
191        ((p) < NLBLOCK->PSEND && \
192         _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
193           utf8)) \
194        : \
195        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
196         (p)[0] == NLBLOCK->nl[0] && \
197         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
198        ) \
199      )
200    
201    /* This macro checks for a newline immediately preceding the given position */
202    
203    #define WAS_NEWLINE(p) \
204      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
205        ((p) > NLBLOCK->PSSTART && \
206         _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
207           &(NLBLOCK->nllen), utf8)) \
208        : \
209        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
210         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
211         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
212        ) \
213      )
214    
215    /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
216    with a custom type. This makes it possible, for example, to allow pcre_exec()
217    to process subject strings that are discontinuous by using a smart pointer
218    class. It must always be possible to inspect all of the subject string in
219    pcre_exec() because of the way it backtracks. Two macros are required in the
220    normal case, for sign-unspecified and unsigned char pointers. The former is
221    used for the external interface and appears in pcre.h, which is why its name
222    must begin with PCRE_. */
223    
224    #ifdef CUSTOM_SUBJECT_PTR
225    #define PCRE_SPTR CUSTOM_SUBJECT_PTR
226    #define USPTR CUSTOM_SUBJECT_PTR
227    #else
228    #define PCRE_SPTR const char *
229    #define USPTR const unsigned char *
230    #endif
231    
232    
 #include "pcre.h"  
233    
234  /* Include the (copy of) the public ucp header, changing the external name into  /* Include the public PCRE header and the definitions of UCP character property
235  a private one. This does no harm, even if we aren't compiling UCP support. */  values. */
236    
237  #define ucp_findchar _pcre_ucp_findchar  #include "pcre.h"
238  #include "ucp.h"  #include "ucp.h"
239    
240  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
# Line 127  need to have their names changed. PCRE m Line 242  need to have their names changed. PCRE m
242  option on the command line. */  option on the command line. */
243    
244  #ifdef VPCOMPAT  #ifdef VPCOMPAT
245    #define strlen(s)        _strlen(s)
246  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
247    #define memcmp(s,c,n)    _memcmp(s,c,n)
248  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
249  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
250  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 136  option on the command line. */ Line 253  option on the command line. */
253  /* 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(),
254  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
255  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
256  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. */  
257    
258  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
259  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
260  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
261  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
262  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
263  void *  static void *
264  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
265  {  {
266  int i;  size_t i;
267  dest += n;  unsigned char *dest = (unsigned char *)d;
268  src += n;  const unsigned char *src = (const unsigned char *)s;
269  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
270      {
271      dest += n;
272      src += n;
273      for (i = 0; i < n; ++i) *(--dest) = *(--src);
274      return (void *)dest;
275      }
276    else
277      {
278      for (i = 0; i < n; ++i) *dest++ = *src++;
279      return (void *)(dest - n);
280      }
281  }  }
282  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
283  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 254  we know we are in UTF-8 mode. */ Line 380  we know we are in UTF-8 mode. */
380    
381  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
382    c = *eptr; \    c = *eptr; \
383    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
384      { \      { \
385      int gcii; \      int gcii; \
386      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 272  pointer. */ Line 398  pointer. */
398    
399  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
400    c = *eptr; \    c = *eptr; \
401    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
402      { \      { \
403      int gcii; \      int gcii; \
404      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 290  know we are in UTF-8 mode. */ Line 416  know we are in UTF-8 mode. */
416    
417  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
418    c = *eptr++; \    c = *eptr++; \
419    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
420      { \      { \
421      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
422      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 306  know we are in UTF-8 mode. */ Line 432  know we are in UTF-8 mode. */
432    
433  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
434    c = *eptr++; \    c = *eptr++; \
435    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
436      { \      { \
437      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
438      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 323  if there are extra bytes. This is called Line 449  if there are extra bytes. This is called
449    
450  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
451    c = *eptr; \    c = *eptr; \
452    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
453      { \      { \
454      int gcii; \      int gcii; \
455      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 357  Standard C system should have one. */ Line 483  Standard C system should have one. */
483    
484  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
485    
486  /* 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.
487  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
488  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
489  significant end. Make sure they don't overlap! */  out, there is a dummy word in the structure that could be used for the private
490    bits. */
491    
492    #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */
493  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */
494  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */
495  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
496  #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 */  
497    
498  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
499    
# Line 375  significant end. Make sure they don't ov Line 502  significant end. Make sure they don't ov
502  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
503  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
504    
505    #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
506                               PCRE_NEWLINE_ANYCRLF)
507    
508  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
509    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
510     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
511     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| \
512       PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
513    
514  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
515    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
516     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS)
517    
518  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
519    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
520     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
521    
522  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
523    
# Line 418  typedef int BOOL; Line 549  typedef int BOOL;
549  #define FALSE   0  #define FALSE   0
550  #define TRUE    1  #define TRUE    1
551    
552  /* 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). */  
553    
554  #ifndef ESC_e  #ifndef ESC_e
555  #define ESC_e 27  #define ESC_e 27
# Line 431  ESC_n is defined as yet another macro, w Line 560  ESC_n is defined as yet another macro, w
560  #endif  #endif
561    
562  #ifndef ESC_n  #ifndef ESC_n
563  #define ESC_n NEWLINE  #define ESC_n '\n'
564  #endif  #endif
565    
566  #ifndef ESC_r  #ifndef ESC_r
# Line 445  ESC_n is defined as yet another macro, w Line 574  ESC_n is defined as yet another macro, w
574  #define ESC_tee '\t'  #define ESC_tee '\t'
575  #endif  #endif
576    
577  /* These are escaped items that aren't just an encoding of a particular data  /* Codes for different types of Unicode property */
 value such as \n. They must have non-zero values, as check_escape() returns  
 their negation. Also, they must appear in the same order as in the opcode  
 definitions below, up to ESC_z. There's a dummy for OP_ANY because it  
 corresponds to "." rather than an escape sequence. The final one must be  
 ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  
 tests in the code for an escape greater than ESC_b and less than ESC_Z to  
 detect the types that may be repeated. These are the types that consume  
 characters. If any new escapes are put in between that don't consume a  
 character, that code will have to change. */  
578    
579  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  #define PT_ANY        0    /* Any property - matches all chars */
580         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
581         ESC_Q, ESC_REF };  #define PT_GC         2    /* General characteristic (e.g. L) */
582    #define PT_PC         3    /* Particular characteristic (e.g. Lu) */
583    #define PT_SC         4    /* Script (e.g. Han) */
584    
585  /* 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
586  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 469  contain UTF-8 characters with values gre Line 591  contain UTF-8 characters with values gre
591  #define XCL_END       0    /* Marks end of individual items */  #define XCL_END       0    /* Marks end of individual items */
592  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
593  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
594  #define XCL_PROP      3    /* Unicode property (one property code) follows */  #define XCL_PROP      3    /* Unicode property (2-byte property code follows) */
595  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
596    
597    /* These are escaped items that aren't just an encoding of a particular data
598    value such as \n. They must have non-zero values, as check_escape() returns
599    their negation. Also, they must appear in the same order as in the opcode
600    definitions below, up to ESC_z. There's a dummy for OP_ANY because it
601    corresponds to "." rather than an escape sequence. The final one must be
602    ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
603    There are two tests in the code for an escape greater than ESC_b and less than
604    ESC_Z to detect the types that may be repeated. These are the types that
605    consume characters. If any new escapes are put in between that don't consume a
606    character, that code will have to change. */
607    
608    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
609           ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,
610           ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };
611    
612    
613  /* 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
614  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
615  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.
616  Note that whenever this list is updated, the two macro definitions that follow  
617  must also be updated to match. */  To keep stored, compiled patterns compatible, new opcodes should be added
618    immediately before OP_BRA, where (since release 7.0) a gap is left for this
619    purpose.
620    
621    *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
622    that follow must also be updated to match. There is also a table called
623    "coptable" in pcre_dfa_exec.c that must be updated. */
624    
625  enum {  enum {
626    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 486  enum { Line 629  enum {
629    
630    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
631    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
632    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
633    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
634    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
635    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
636    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
637    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
638    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
639    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
640    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
641    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character */
642    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */
643    OP_PROP,           /* 14 \p (Unicode property) */    OP_NOTPROP,        /* 14 \P (not Unicode property) */
644    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_PROP,           /* 15 \p (Unicode property) */
645    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_ANYNL,          /* 16 \R (any newline sequence) */
646    OP_EOD,            /* 17 End of data: \z */    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */
647      OP_HSPACE,         /* 18 \h (horizontal whitespace) */
648    OP_OPT,            /* 18 Set runtime options */    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */
649    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_VSPACE,         /* 20 \v (vertical whitespace) */
650    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */
651    OP_CHAR,           /* 21 Match one character, casefully */    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */
652    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EOD,            /* 23 End of data: \z */
653    OP_NOT,            /* 23 Match anything but the following char */  
654      OP_OPT,            /* 24 Set runtime options */
655    OP_STAR,           /* 24 The maximizing and minimizing versions of */    OP_CIRC,           /* 25 Start of line - varies with multiline switch */
656    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_DOLL,           /* 26 End of line - varies with multiline switch */
657    OP_PLUS,           /* 26 the minimizing one second. */    OP_CHAR,           /* 27 Match one character, casefully */
658    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_CHARNC,         /* 28 Match one character, caselessly */
659    OP_QUERY,          /* 28 */    OP_NOT,            /* 29 Match one character, not the following one */
660    OP_MINQUERY,       /* 29 */  
661    OP_UPTO,           /* 30 From 0 to n matches */    OP_STAR,           /* 30 The maximizing and minimizing versions of */
662    OP_MINUPTO,        /* 31 */    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */
663    OP_EXACT,          /* 32 Exactly n matches */    OP_PLUS,           /* 32 the minimizing one second. */
664      OP_MINPLUS,        /* 33 This first set applies to single characters.*/
665    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_QUERY,          /* 34 */
666    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    OP_MINQUERY,       /* 35 */
667    OP_NOTPLUS,        /* 35 the minimizing one second. */  
668    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */    OP_UPTO,           /* 36 From 0 to n matches */
669    OP_NOTQUERY,       /* 37 */    OP_MINUPTO,        /* 37 */
670    OP_NOTMINQUERY,    /* 38 */    OP_EXACT,          /* 38 Exactly n matches */
671    OP_NOTUPTO,        /* 39 From 0 to n matches */  
672    OP_NOTMINUPTO,     /* 40 */    OP_POSSTAR,        /* 39 Possessified star */
673    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_POSPLUS,        /* 40 Possessified plus */
674      OP_POSQUERY,       /* 41 Posesssified query */
675    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_POSUPTO,        /* 42 Possessified upto */
676    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */  
677    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */
678    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */
679    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */
680    OP_TYPEMINQUERY,   /* 47 */    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */
681    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */
682    OP_TYPEMINUPTO,    /* 49 */    OP_NOTMINQUERY,    /* 48 */
683    OP_TYPEEXACT,      /* 50 Exactly n matches */  
684      OP_NOTUPTO,        /* 49 From 0 to n matches */
685    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTMINUPTO,     /* 50 */
686    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTEXACT,       /* 51 Exactly n matches */
687    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */  
688    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */    OP_NOTPOSSTAR,     /* 52 Possessified versions */
689    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_NOTPOSPLUS,     /* 53 */
690    OP_CRMINQUERY,     /* 56 */    OP_NOTPOSQUERY,    /* 54 */
691    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_NOTPOSUPTO,     /* 55 */
692    OP_CRMINRANGE,     /* 58 */  
693      OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */
694      OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */
695      OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */
696      OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */
697      OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */
698      OP_TYPEMINQUERY,   /* 61 */
699    
700      OP_TYPEUPTO,       /* 62 From 0 to n matches */
701      OP_TYPEMINUPTO,    /* 63 */
702      OP_TYPEEXACT,      /* 64 Exactly n matches */
703    
704      OP_TYPEPOSSTAR,    /* 65 Possessified versions */
705      OP_TYPEPOSPLUS,    /* 66 */
706      OP_TYPEPOSQUERY,   /* 67 */
707      OP_TYPEPOSUPTO,    /* 68 */
708    
709      OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */
710      OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */
711      OP_CRPLUS,         /* 71 the minimizing one second. These codes must */
712      OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */
713      OP_CRQUERY,        /* 73 These are for character classes and back refs */
714      OP_CRMINQUERY,     /* 74 */
715      OP_CRRANGE,        /* 75 These are different to the three sets above. */
716      OP_CRMINRANGE,     /* 76 */
717    
718    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 77 Match a character class, chars < 256 only */
719    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative
720                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
721                             character > 255 is encountered. */                             character > 255 is encountered. */
722    
723    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the
724                             class. This does both positive and negative. */                             class. This does both positive and negative. */
725    
726    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 80 Match a back reference */
727    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */
728    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 82 Call out to external function if provided */
729    
730    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 83 Start of alternation */
731    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 */
732    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 85 These two must remain together and in this */
733    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */
734    
735    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
736    
737    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 87 Positive lookahead */
738    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 88 Negative lookahead */
739    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 89 Positive lookbehind */
740    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */
741    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */
742    
743    /* 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,
744    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. */
745    
746    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */    OP_ONCE,           /* 92 Atomic group */
747    OP_COND,           /* 75 Conditional group */    OP_BRA,            /* 83 Start of non-capturing bracket */
748    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */    OP_CBRA,           /* 94 Start of capturing bracket */
749      OP_COND,           /* 95 Conditional group */
750    OP_BRAZERO,        /* 77 These two must remain together and in this */  
751    OP_BRAMINZERO,     /* 78 order. */    /* These three must follow the previous three, in the same order. There's a
752      check for >= SBRA to distinguish the two sets. */
753    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
754                             than can fit into an opcode. */    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */
755      OP_SCBRA,          /* 97 Start of capturing bracket, check empty */
756    OP_BRA             /* 80 This and greater values are used for brackets that    OP_SCOND,          /* 98 Conditional group, check empty */
757                             extract substrings up to EXTRACT_BASIC_MAX. After  
758                             that, use is made of OP_BRANUMBER. */    OP_CREF,           /* 99 Used to hold a capture number as condition */
759  };    OP_RREF,           /* 100 Used to hold a recursion number as condition */
760      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. */  
761    
762  #define EXTRACT_BASIC_MAX  100    OP_BRAZERO,        /* 102 These two must remain together and in this */
763      OP_BRAMINZERO      /* 103 order. */
764    };
765    
766    
767  /* 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
768  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
769    
770  #define OP_NAME_LIST \  #define OP_NAME_LIST \
771    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
772    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
773    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
774    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
775    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
776    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
777      "*+","++", "?+", "{",                                           \
778    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
779      "*+","++", "?+", "{",                                           \
780    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
781      "*+","++", "?+", "{",                                           \
782    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
783    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
784    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
785    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
786    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \
787      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
788    
789    
790  /* 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 634  in UTF-8 mode. The code that uses this t Line 798  in UTF-8 mode. The code that uses this t
798    
799  #define OP_LENGTHS \  #define OP_LENGTHS \
800    1,                             /* End                                    */ \    1,                             /* End                                    */ \
801    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                     */ \
802      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
803    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
804    2, 2, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
805      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
806    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
807    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
808    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 644  in UTF-8 mode. The code that uses this t Line 810  in UTF-8 mode. The code that uses this t
810    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
811    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
812    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
813      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
814    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
815    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
816    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
817      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
818    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
819    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
820    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
821      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
822    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
823    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
824    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 668  in UTF-8 mode. The code that uses this t Line 837  in UTF-8 mode. The code that uses this t
837    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
838    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
839    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
840    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
841      1+LINK_SIZE,                   /* BRA                                    */ \
842      3+LINK_SIZE,                   /* CBRA                                   */ \
843    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
844      1+LINK_SIZE,                   /* SBRA                                   */ \
845      3+LINK_SIZE,                   /* SCBRA                                  */ \
846      1+LINK_SIZE,                   /* SCOND                                  */ \
847    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
848      3,                             /* RREF                                   */ \
849      1,                             /* DEF                                    */ \
850    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
   3,                             /* BRANUMBER                              */ \  
   1+LINK_SIZE                    /* BRA                                    */ \  
851    
852    
853  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
854    
855  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
856    
857  /* 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
858  tracked. */  tracked. */
# Line 687  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 861  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
861         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
862         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
863         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
864         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
865           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };
866    
867  /* 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
868  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 742  typedef struct compile_data { Line 917  typedef struct compile_data {
917    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
918    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
919    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
920      const uschar *start_workspace;/* The start of working space */
921    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
922    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
923      const uschar *end_pattern;    /* The end of the pattern */
924      uschar *hwm;                  /* High watermark of workspace */
925    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
926    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
927    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
928      int  bracount;                /* Count of capturing parens */
929    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
930    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
931      int  external_options;        /* External (initial) options */
932    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
933    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL nopartial;               /* Set TRUE if partial won't work */
934      int  nltype;                  /* Newline type */
935      int  nllen;                   /* Newline string length */
936      uschar nl[4];                 /* Newline string when fixed length */
937  } compile_data;  } compile_data;
938    
939  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 768  typedef struct recursion_info { Line 951  typedef struct recursion_info {
951    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
952    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
953    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 */
954    const uschar *save_start;     /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
955    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
956    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
957  } recursion_info;  } recursion_info;
958    
959  /* 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
960  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
961  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. */
962  with any changes of local variable. However, the pointer to the current frame  
963  must be saved in some "static" place over a longjmp(). We declare the  typedef struct eptrblock {
964  structure here so that we can put a pointer in the match_data structure.    struct eptrblock *epb_prev;
965  NOTE: This isn't used for a "normal" compilation of pcre. */    USPTR epb_saved_eptr;
966    } eptrblock;
967    
 struct heapframe;  
968    
969  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
970  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
971    
972  typedef struct match_data {  typedef struct match_data {
973    unsigned long int match_call_count; /* As it says */    unsigned long int match_call_count;      /* As it says */
974    unsigned long int match_limit;/* As it says */    unsigned long int match_limit;           /* As it says */
975      unsigned long int match_limit_recursion; /* As it says */
976    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
977    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
978    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
979      int    nltype;                /* Newline type */
980      int    nllen;                 /* Newline string length */
981      uschar nl[4];                 /* Newline string when fixed */
982    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
983    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
984    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
# Line 803  typedef struct match_data { Line 990  typedef struct match_data {
990    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
991    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
992    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
993    const uschar *start_subject;  /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
994    const uschar *end_subject;    /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
995    const uschar *start_match;    /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
996    const uschar *end_match_ptr;  /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
997    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
998    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
999    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1000      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1001      int    eptrn;                 /* Next free eptrblock */
1002    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1003    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 */  
1004  } match_data;  } match_data;
1005    
1006  /* 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 825  typedef struct dfa_match_data { Line 1013  typedef struct dfa_match_data {
1013    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1014    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1015    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1016      int    nltype;                /* Newline type */
1017      int    nllen;                 /* Newline string length */
1018      uschar nl[4];                 /* Newline string when fixed */
1019    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1020  } dfa_match_data;  } dfa_match_data;
1021    
# Line 861  total length. */ Line 1052  total length. */
1052  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
1053  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1054    
1055  /* Layout of the UCP type table that translates property names into codes for  /* Layout of the UCP type table that translates property names into types and
1056  ucp_findchar(). */  codes. */
1057    
1058  typedef struct {  typedef struct {
1059    const char *name;    const char *name;
1060    int value;    pcre_uint16 type;
1061      pcre_uint16 value;
1062  } ucp_type_table;  } ucp_type_table;
1063    
1064    
# Line 894  extern const uschar _pcre_OP_lengths[]; Line 1086  extern const uschar _pcre_OP_lengths[];
1086  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
1087  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1088    
1089  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1090  extern void        _pcre_printint(pcre *, FILE *);                        int *, BOOL);
1091  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,  extern int          _pcre_ord2utf8(int, uschar *);
1092                       const pcre_study_data *, pcre_study_data *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1093  extern int         _pcre_ucp_findchar(const int, int *, int *);                        const pcre_study_data *, pcre_study_data *);
1094  extern int         _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);
1095  extern BOOL        _pcre_xclass(int, const uschar *);  extern unsigned int _pcre_ucp_othercase(const unsigned int);
1096    extern int          _pcre_valid_utf8(const uschar *, int);
1097    extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1098                          int *, BOOL);
1099    extern BOOL         _pcre_xclass(int, const uschar *);
1100    
1101    #endif
1102    
1103  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.77  
changed lines
  Added in v.189

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12