/[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 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 369 by ph10, Sun Aug 24 16:53:47 2008 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) 1997-2008 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
134    
135    /* When compiling with the MSVC compiler, it is sometimes necessary to include
136    a "calling convention" before exported function names. (This is secondhand
137    information; I know nothing about MSVC myself). For example, something like
138    
139      void __cdecl function(....)
140    
141    might be needed. In order so make this easy, all the exported functions have
142    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
143    set, we ensure here that it has no effect. */
144    
145    #ifndef PCRE_CALL_CONVENTION
146    #define PCRE_CALL_CONVENTION
147  #endif  #endif
148    
149  /* 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 91  preprocessor time in standard C environm Line 154  preprocessor time in standard C environm
154    
155  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
156    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
157      typedef short pcre_int16;
158  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
159    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
160      typedef int pcre_int16;
161  #else  #else
162    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
163  #endif  #endif
164    
165  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
166    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
167      typedef int pcre_int32;
168  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
169    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
170      typedef long int pcre_int32;
171  #else  #else
172    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
173  #endif  #endif
# Line 114  Unix, where it is defined in sys/types, Line 181  Unix, where it is defined in sys/types,
181    
182  typedef unsigned char uschar;  typedef unsigned char uschar;
183    
184    /* This is an unsigned int value that no character can ever have. UTF-8
185    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
186    0x0010ffff). */
187    
188    #define NOTACHAR 0xffffffff
189    
190    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
191    "any" and "anycrlf" at present). The following macros are used to package up
192    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
193    modules to indicate in which datablock the parameters exist, and what the
194    start/end of string field names are. */
195    
196    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
197    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
198    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
199    
200    /* This macro checks for a newline at the given position */
201    
202    #define IS_NEWLINE(p) \
203      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
204        ((p) < NLBLOCK->PSEND && \
205         _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
206           utf8)) \
207        : \
208        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
209         (p)[0] == NLBLOCK->nl[0] && \
210         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
211        ) \
212      )
213    
214    /* This macro checks for a newline immediately preceding the given position */
215    
216    #define WAS_NEWLINE(p) \
217      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
218        ((p) > NLBLOCK->PSSTART && \
219         _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
220           &(NLBLOCK->nllen), utf8)) \
221        : \
222        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
223         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
224         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
225        ) \
226      )
227    
228    /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
229    with a custom type. This makes it possible, for example, to allow pcre_exec()
230    to process subject strings that are discontinuous by using a smart pointer
231    class. It must always be possible to inspect all of the subject string in
232    pcre_exec() because of the way it backtracks. Two macros are required in the
233    normal case, for sign-unspecified and unsigned char pointers. The former is
234    used for the external interface and appears in pcre.h, which is why its name
235    must begin with PCRE_. */
236    
237    #ifdef CUSTOM_SUBJECT_PTR
238    #define PCRE_SPTR CUSTOM_SUBJECT_PTR
239    #define USPTR CUSTOM_SUBJECT_PTR
240    #else
241    #define PCRE_SPTR const char *
242    #define USPTR const unsigned char *
243    #endif
244    
245    
246    
247  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
248  values. */  values. */
249    
# Line 125  need to have their names changed. PCRE m Line 255  need to have their names changed. PCRE m
255  option on the command line. */  option on the command line. */
256    
257  #ifdef VPCOMPAT  #ifdef VPCOMPAT
258    #define strlen(s)        _strlen(s)
259  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
260    #define memcmp(s,c,n)    _memcmp(s,c,n)
261  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
262  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
263  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 134  option on the command line. */ Line 266  option on the command line. */
266  /* 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(),
267  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
268  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
269  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. */  
270    
271  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
272  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
273  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
274  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
275  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
276  void *  static void *
277  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
278  {  {
279  int i;  size_t i;
280  dest += n;  unsigned char *dest = (unsigned char *)d;
281  src += n;  const unsigned char *src = (const unsigned char *)s;
282  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
283      {
284      dest += n;
285      src += n;
286      for (i = 0; i < n; ++i) *(--dest) = *(--src);
287      return (void *)dest;
288      }
289    else
290      {
291      for (i = 0; i < n; ++i) *dest++ = *src++;
292      return (void *)(dest - n);
293      }
294  }  }
295  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
296  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 235  capturing parenthesis numbers in back re Line 376  capturing parenthesis numbers in back re
376    
377  /* 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
378  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
379  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
380    never be called in byte mode. To make sure it can never even appear when UTF-8
381    support is omitted, we don't even define it. */
382    
383  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
384  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 243  byte-mode, and more complicated ones for Line 386  byte-mode, and more complicated ones for
386  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
387  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
388  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
389  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
390    
391  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
392    
# Line 252  we know we are in UTF-8 mode. */ Line 395  we know we are in UTF-8 mode. */
395    
396  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
397    c = *eptr; \    c = *eptr; \
398    if ((c & 0xc0) == 0xc0) \    if (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 270  pointer. */ Line 413  pointer. */
413    
414  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
415    c = *eptr; \    c = *eptr; \
416    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
417      { \      { \
418      int gcii; \      int gcii; \
419      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 288  know we are in UTF-8 mode. */ Line 431  know we are in UTF-8 mode. */
431    
432  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
433    c = *eptr++; \    c = *eptr++; \
434    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
435      { \      { \
436      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
437      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 304  know we are in UTF-8 mode. */ Line 447  know we are in UTF-8 mode. */
447    
448  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
449    c = *eptr++; \    c = *eptr++; \
450    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
451      { \      { \
452      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
453      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 321  if there are extra bytes. This is called Line 464  if there are extra bytes. This is called
464    
465  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
466    c = *eptr; \    c = *eptr; \
467    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
468      { \      { \
469      int gcii; \      int gcii; \
470      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 336  if there are extra bytes. This is called Line 479  if there are extra bytes. This is called
479      }      }
480    
481  /* 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
482  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
483    because almost all calls are already within a block of UTF-8 only code. */
484    
485  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
486    
487  #endif  #endif
488    
# Line 355  Standard C system should have one. */ Line 499  Standard C system should have one. */
499    
500  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
501    
502  /* Private options flags start at the most significant end of the four bytes,  /* Private flags containing information about the compiled regex. They used to
503  but skip the top bit so we can use ints for convenience without getting tangled  live at the top end of the options word, but that got almost full, so now they
504  with negative values. The public options defined in pcre.h start at the least  are in a 16-bit flags word. */
505  significant end. Make sure they don't overlap! */  
506    #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
507  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
508  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
509  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
510  #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
511  #define PCRE_NOPARTIAL     0x04000000  /* can't use partial with this regex */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
512    
513  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
514    
# Line 373  significant end. Make sure they don't ov Line 517  significant end. Make sure they don't ov
517  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
518  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
519    
520    #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
521                               PCRE_NEWLINE_ANYCRLF)
522    
523  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
524    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
525     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
526     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| \
527       PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
528       PCRE_JAVASCRIPT_COMPAT)
529    
530  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
531    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
532     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
533    
534  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
535    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
536     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
537       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
538    
539  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
540    
# Line 409  variable-length repeat, or a anything ot Line 559  variable-length repeat, or a anything ot
559  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
560  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
561    
562  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
563    environments where these macros are defined elsewhere. */
564    
565    #ifndef FALSE
566  typedef int BOOL;  typedef int BOOL;
567    
568  #define FALSE   0  #define FALSE   0
569  #define TRUE    1  #define TRUE    1
570    #endif
571    
572  /* 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). */  
573    
574  #ifndef ESC_e  #ifndef ESC_e
575  #define ESC_e 27  #define ESC_e 27
# Line 429  ESC_n is defined as yet another macro, w Line 580  ESC_n is defined as yet another macro, w
580  #endif  #endif
581    
582  #ifndef ESC_n  #ifndef ESC_n
583  #define ESC_n NEWLINE  #define ESC_n '\n'
584  #endif  #endif
585    
586  #ifndef ESC_r  #ifndef ESC_r
# Line 443  ESC_n is defined as yet another macro, w Line 594  ESC_n is defined as yet another macro, w
594  #define ESC_tee '\t'  #define ESC_tee '\t'
595  #endif  #endif
596    
597  /* These are escaped items that aren't just an encoding of a particular data  /* Codes for different types of Unicode property */
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  #define PT_ANY        0    /* Any property - matches all chars */
600  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
601  corresponds to "." rather than an escape sequence. The final one must be  #define PT_GC         2    /* General characteristic (e.g. L) */
602  ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */
603  tests in the code for an escape greater than ESC_b and less than ESC_Z to  #define PT_SC         4    /* Script (e.g. Han) */
 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. */  
   
 enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  
        ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,  
        ESC_Q, ESC_REF };  
604    
605  /* 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
606  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 467  contain UTF-8 characters with values gre Line 611  contain UTF-8 characters with values gre
611  #define XCL_END       0    /* Marks end of individual items */  #define XCL_END       0    /* Marks end of individual items */
612  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
613  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
614  #define XCL_PROP      3    /* Unicode property (one property code) follows */  #define XCL_PROP      3    /* Unicode property (2-byte property code follows) */
615  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
616    
617    /* These are escaped items that aren't just an encoding of a particular data
618    value such as \n. They must have non-zero values, as check_escape() returns
619    their negation. Also, they must appear in the same order as in the opcode
620    definitions below, up to ESC_z. There's a dummy for OP_ANY because it
621    corresponds to "." rather than an escape sequence, and another for OP_ALLANY
622    (which is used for [^] in JavaScript compatibility mode).
623    
624    The final escape must be ESC_REF as subsequent values are used for
625    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
626    greater than ESC_b and less than ESC_Z to detect the types that may be
627    repeated. These are the types that consume characters. If any new escapes are
628    put in between that don't consume a character, that code will have to change.
629    */
630    
631    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
632           ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
633           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
634           ESC_REF };
635    
636  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  
637  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
638  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.
639  Note that whenever this list is updated, the two macro definitions that follow  
640  must also be updated to match. */  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
641    that follow must also be updated to match. There is also a table called
642    "coptable" in pcre_dfa_exec.c that must be updated. */
643    
644  enum {  enum {
645    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 484  enum { Line 648  enum {
648    
649    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
650    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
651    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
652    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
653    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
654    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
655    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
656    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
657    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
658    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
659    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
660    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
661    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
662    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
663    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
664    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_PROP,           /* 16 \p (Unicode property) */
665    OP_EOD,            /* 17 End of data: \z */    OP_ANYNL,          /* 17 \R (any newline sequence) */
666      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
667    OP_OPT,            /* 18 Set runtime options */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
668    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
669    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
670    OP_CHAR,           /* 21 Match one character, casefully */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
671    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
672    OP_NOT,            /* 23 Match anything but the following char */    OP_EOD,            /* 24 End of data: \z */
673    
674    OP_STAR,           /* 24 The maximizing and minimizing versions of */    OP_OPT,            /* 25 Set runtime options */
675    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
676    OP_PLUS,           /* 26 the minimizing one second. */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
677    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_CHAR,           /* 28 Match one character, casefully */
678    OP_QUERY,          /* 28 */    OP_CHARNC,         /* 29 Match one character, caselessly */
679    OP_MINQUERY,       /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
680    OP_UPTO,           /* 30 From 0 to n matches */  
681    OP_MINUPTO,        /* 31 */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
682    OP_EXACT,          /* 32 Exactly n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
683      OP_PLUS,           /* 33 the minimizing one second. */
684    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
685    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    OP_QUERY,          /* 35 */
686    OP_NOTPLUS,        /* 35 the minimizing one second. */    OP_MINQUERY,       /* 36 */
687    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */  
688    OP_NOTQUERY,       /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
689    OP_NOTMINQUERY,    /* 38 */    OP_MINUPTO,        /* 38 */
690    OP_NOTUPTO,        /* 39 From 0 to n matches */    OP_EXACT,          /* 39 Exactly n matches */
691    OP_NOTMINUPTO,     /* 40 */  
692    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_POSSTAR,        /* 40 Possessified star */
693      OP_POSPLUS,        /* 41 Possessified plus */
694    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_POSQUERY,       /* 42 Posesssified query */
695    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */    OP_POSUPTO,        /* 43 Possessified upto */
696    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */  
697    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
698    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
699    OP_TYPEMINQUERY,   /* 47 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
700    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
701    OP_TYPEMINUPTO,    /* 49 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
702    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_NOTMINQUERY,    /* 49 */
703    
704    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTUPTO,        /* 50 From 0 to n matches */
705    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTMINUPTO,     /* 51 */
706    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */    OP_NOTEXACT,       /* 52 Exactly n matches */
707    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */  
708    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
709    OP_CRMINQUERY,     /* 56 */    OP_NOTPOSPLUS,     /* 54 */
710    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_NOTPOSQUERY,    /* 55 */
711    OP_CRMINRANGE,     /* 58 */    OP_NOTPOSUPTO,     /* 56 */
712    
713      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
714      OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
715      OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
716      OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
717      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
718      OP_TYPEMINQUERY,   /* 62 */
719    
720      OP_TYPEUPTO,       /* 63 From 0 to n matches */
721      OP_TYPEMINUPTO,    /* 64 */
722      OP_TYPEEXACT,      /* 65 Exactly n matches */
723    
724      OP_TYPEPOSSTAR,    /* 66 Possessified versions */
725      OP_TYPEPOSPLUS,    /* 67 */
726      OP_TYPEPOSQUERY,   /* 68 */
727      OP_TYPEPOSUPTO,    /* 69 */
728    
729      OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
730      OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
731      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
732      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
733      OP_CRQUERY,        /* 74 These are for character classes and back refs */
734      OP_CRMINQUERY,     /* 75 */
735      OP_CRRANGE,        /* 76 These are different to the three sets above. */
736      OP_CRMINRANGE,     /* 77 */
737    
738    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
739    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
740                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
741                             character > 255 is encountered. */                             character > 255 is encountered. */
742    
743    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
744                             class. This does both positive and negative. */                             class. This does both positive and negative. */
745    
746    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 81 Match a back reference */
747    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
748    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
749    
750    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
751    OP_KET,            /* 66 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
752    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
753    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
754    
755    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
756    
757    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
758    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
759    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
760    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
761    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
762    
763    /* 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,
764    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. */
765    
766    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */    OP_ONCE,           /* 93 Atomic group */
767    OP_COND,           /* 75 Conditional group */    OP_BRA,            /* 94 Start of non-capturing bracket */
768    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */    OP_CBRA,           /* 95 Start of capturing bracket */
769      OP_COND,           /* 96 Conditional group */
770    OP_BRAZERO,        /* 77 These two must remain together and in this */  
771    OP_BRAMINZERO,     /* 78 order. */    /* These three must follow the previous three, in the same order. There's a
772      check for >= SBRA to distinguish the two sets. */
773    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
774                             than can fit into an opcode. */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
775      OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
776    OP_BRA             /* 80 This and greater values are used for brackets that    OP_SCOND,          /* 99 Conditional group, check empty */
777                             extract substrings up to EXTRACT_BASIC_MAX. After  
778                             that, use is made of OP_BRANUMBER. */    OP_CREF,           /* 100 Used to hold a capture number as condition */
779  };    OP_RREF,           /* 101 Used to hold a recursion number as condition */
780      OP_DEF,            /* 102 The DEFINE condition */
781    
782      OP_BRAZERO,        /* 103 These two must remain together and in this */
783      OP_BRAMINZERO,     /* 104 order. */
784    
785  /* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and    /* These are backtracking control verbs */
 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. */  
786    
787  #define EXTRACT_BASIC_MAX  100    OP_PRUNE,          /* 105 */
788      OP_SKIP,           /* 106 */
789      OP_THEN,           /* 107 */
790      OP_COMMIT,         /* 108 */
791    
792      /* These are forced failure and success verbs */
793    
794      OP_FAIL,           /* 109 */
795      OP_ACCEPT,         /* 110 */
796    
797      /* This is used to skip a subpattern with a {0} quantifier */
798    
799      OP_SKIPZERO        /* 111 */
800    };
801    
802    
803  /* 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
804  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
805    
806  #define OP_NAME_LIST \  #define OP_NAME_LIST \
807    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
808    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
809    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
810    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
811    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
812    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
813      "*+","++", "?+", "{",                                           \
814    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
815      "*+","++", "?+", "{",                                           \
816    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
817      "*+","++", "?+", "{",                                           \
818    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
819    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
820    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
821    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
822    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
823      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
824      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
825      "Skip zero"
826    
827    
828  /* 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 632  in UTF-8 mode. The code that uses this t Line 836  in UTF-8 mode. The code that uses this t
836    
837  #define OP_LENGTHS \  #define OP_LENGTHS \
838    1,                             /* End                                    */ \    1,                             /* End                                    */ \
839    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                     */ \
840    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
841    2, 2, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
842      3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
843      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
844    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
845    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
846    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 642  in UTF-8 mode. The code that uses this t Line 848  in UTF-8 mode. The code that uses this t
848    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
849    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
850    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
851      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
852    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
853    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
854    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
855      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
856    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
857    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
858    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
859      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
860    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
861    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
862    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 666  in UTF-8 mode. The code that uses this t Line 875  in UTF-8 mode. The code that uses this t
875    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
876    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
877    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
878    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
879      1+LINK_SIZE,                   /* BRA                                    */ \
880      3+LINK_SIZE,                   /* CBRA                                   */ \
881    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
882      1+LINK_SIZE,                   /* SBRA                                   */ \
883      3+LINK_SIZE,                   /* SCBRA                                  */ \
884      1+LINK_SIZE,                   /* SCOND                                  */ \
885    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
886      3,                             /* RREF                                   */ \
887      1,                             /* DEF                                    */ \
888    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
889    3,                             /* BRANUMBER                              */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
890    1+LINK_SIZE                    /* BRA                                    */ \    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
891    
892    
893  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
894    
895  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
896    
897  /* 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
898  tracked. */  tracked. */
# Line 685  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 901  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
901         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
902         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
903         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
904         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
905           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
906           ERR60, ERR61, ERR62, ERR63, ERR64 };
907    
908  /* 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
909  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 707  NOTE NOTE NOTE: Line 925  NOTE NOTE NOTE:
925  typedef struct real_pcre {  typedef struct real_pcre {
926    pcre_uint32 magic_number;    pcre_uint32 magic_number;
927    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
928    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
929    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
930      pcre_uint16 dummy1;             /* For future use */
931    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
932    pcre_uint16 top_backref;    pcre_uint16 top_backref;
933    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 740  typedef struct compile_data { Line 958  typedef struct compile_data {
958    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
959    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
960    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
961      const uschar *start_workspace;/* The start of working space */
962    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
963    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
964      const uschar *end_pattern;    /* The end of the pattern */
965      uschar *hwm;                  /* High watermark of workspace */
966    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
967    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
968    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
969      int  bracount;                /* Count of capturing parens as we compile */
970      int  final_bracount;          /* Saved value after first pass */
971    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
972    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
973      int  external_options;        /* External (initial) options */
974      int  external_flags;          /* External flag bits to be set */
975    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
976    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
977      int  nltype;                  /* Newline type */
978      int  nllen;                   /* Newline string length */
979      uschar nl[4];                 /* Newline string when fixed length */
980  } compile_data;  } compile_data;
981    
982  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 766  typedef struct recursion_info { Line 994  typedef struct recursion_info {
994    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
995    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
996    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 */
997    const uschar *save_start;     /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
998    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
999    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1000  } recursion_info;  } recursion_info;
1001    
1002  /* 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
1003  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
1004  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. */
1005  with any changes of local variable. However, the pointer to the current frame  
1006  must be saved in some "static" place over a longjmp(). We declare the  typedef struct eptrblock {
1007  structure here so that we can put a pointer in the match_data structure.    struct eptrblock *epb_prev;
1008  NOTE: This isn't used for a "normal" compilation of pcre. */    USPTR epb_saved_eptr;
1009    } eptrblock;
1010    
 struct heapframe;  
1011    
1012  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1013  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
1014    
1015  typedef struct match_data {  typedef struct match_data {
1016    unsigned long int match_call_count; /* As it says */    unsigned long int match_call_count;      /* As it says */
1017    unsigned long int match_limit;/* As it says */    unsigned long int match_limit;           /* As it says */
1018      unsigned long int match_limit_recursion; /* As it says */
1019    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
1020    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
1021    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1022      int    nltype;                /* Newline type */
1023      int    nllen;                 /* Newline string length */
1024      uschar nl[4];                 /* Newline string when fixed */
1025    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1026    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
1027    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1028    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1029    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1030    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1031      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1032    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1033    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1034    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1035    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1036      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1037    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1038    const uschar *start_subject;  /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1039    const uschar *end_subject;    /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1040    const uschar *start_match;    /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1041    const uschar *end_match_ptr;  /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1042    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1043    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1044    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1045      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1046      int    eptrn;                 /* Next free eptrblock */
1047    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1048    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 */  
1049  } match_data;  } match_data;
1050    
1051  /* 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 823  typedef struct dfa_match_data { Line 1058  typedef struct dfa_match_data {
1058    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1059    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1060    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1061      int    nltype;                /* Newline type */
1062      int    nllen;                 /* Newline string length */
1063      uschar nl[4];                 /* Newline string when fixed */
1064    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1065  } dfa_match_data;  } dfa_match_data;
1066    
# Line 832  typedef struct dfa_match_data { Line 1070  typedef struct dfa_match_data {
1070  #define ctype_letter  0x02  #define ctype_letter  0x02
1071  #define ctype_digit   0x04  #define ctype_digit   0x04
1072  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1073  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1074  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1075    
1076  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
# Line 859  total length. */ Line 1097  total length. */
1097  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
1098  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1099    
1100  /* 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
1101  pcre_ucp_findchar(). */  codes. Each entry used to point directly to a name, but to reduce the number of
1102    relocations in shared libraries, it now has an offset into a single string
1103    instead. */
1104    
1105  typedef struct {  typedef struct {
1106    const char *name;    pcre_uint16 name_offset;
1107    int value;    pcre_uint16 type;
1108      pcre_uint16 value;
1109  } ucp_type_table;  } ucp_type_table;
1110    
1111    
# Line 880  extern const uschar _pcre_utf8_table4[]; Line 1121  extern const uschar _pcre_utf8_table4[];
1121    
1122  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1123    
1124    extern const char   _pcre_utt_names[];
1125  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1126  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1127    
# Line 892  extern const uschar _pcre_OP_lengths[]; Line 1134  extern const uschar _pcre_OP_lengths[];
1134  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
1135  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1136    
1137  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1138  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        int *, BOOL);
1139                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1140  extern int         _pcre_ucp_findchar(const int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1141  extern int         _pcre_valid_utf8(const uschar *, int);                        const pcre_study_data *, pcre_study_data *);
1142  extern BOOL        _pcre_xclass(int, const uschar *);  extern int          _pcre_valid_utf8(const uschar *, int);
1143    extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1144                          int *, BOOL);
1145    extern BOOL         _pcre_xclass(int, const uschar *);
1146    
1147    
1148    /* Unicode character database (UCD) */
1149    
1150    typedef struct {
1151      uschar script;
1152      uschar chartype;
1153      pcre_int32 other_case;
1154    } ucd_record;
1155    
1156    extern const ucd_record  _pcre_ucd_records[];
1157    extern const uschar      _pcre_ucd_stage1[];
1158    extern const pcre_uint16 _pcre_ucd_stage2[];
1159    extern const int         _pcre_ucp_gentype[];
1160    
1161    
1162    /* UCD access macros */
1163    
1164    #define UCD_BLOCK_SIZE 128
1165    #define GET_UCD(ch) (_pcre_ucd_records + \
1166            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1167            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1168    
1169    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1170    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1171    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1172    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1173    
1174  #endif  #endif
1175    

Legend:
Removed from v.85  
changed lines
  Added in v.369

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12