/[pcre]/code/tags/pcre-4.5/internal.h
ViewVC logotype

Diff of /code/tags/pcre-4.5/internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

code/trunk/internal.h revision 19 by nigel, Sat Feb 24 21:38:33 2007 UTC code/tags/pcre-4.5/internal.h revision 74 by nigel, Sat Feb 24 21:40:32 2007 UTC
# Line 3  Line 3 
3  *************************************************/  *************************************************/
4    
5    
 #define PCRE_VERSION       "1.08 27-Mar-1998"  
   
   
6  /* This is a library of functions to support regular expressions whose syntax  /* This is a library of functions to support regular expressions whose syntax
7  and semantics are as close as possible to those of the Perl 5 language. See  and semantics are as close as possible to those of the Perl 5 language. See
8  the file Tech.Notes for some information on the internals.  the file Tech.Notes for some information on the internals.
9    
10  Written by: Philip Hazel <ph10@cam.ac.uk>  Written by: Philip Hazel <ph10@cam.ac.uk>
11    
12             Copyright (c) 1998 University of Cambridge             Copyright (c) 1997-2003 University of Cambridge
13    
14  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
15  Permission is granted to anyone to use this software for any purpose on any  Permission is granted to anyone to use this software for any purpose on any
# Line 28  restrictions: Line 25  restrictions:
25    
26  3. Altered versions must be plainly marked as such, and must not be  3. Altered versions must be plainly marked as such, and must not be
27     misrepresented as being the original software.     misrepresented as being the original software.
28    
29    4. If PCRE is embedded in any software that is released under the GNU
30       General Purpose Licence (GPL), then the terms of that licence shall
31       supersede any condition above with which it is incompatible.
32  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
33  */  */
34    
35  /* This header contains definitions that are shared between the different  /* This header contains definitions that are shared between the different
36  modules, but which are not relevant to the outside. */  modules, but which are not relevant to the outside. */
37    
38  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),  /* Get the definitions provided by running "configure" */
 define a macro for memmove() if USE_BCOPY is defined. */  
39    
40  #ifdef USE_BCOPY  #include "config.h"
 #undef  memmove        /* some systems may have a macro */  
 #define memmove(a, b, c) bcopy(b, a, c)  
 #endif  
41    
42  /* Standard C headers plus the external interface definition */  /* Standard C headers plus the external interface definition. The only time
43    setjmp and stdarg are used is when NO_RECURSE is set. */
44    
45  #include <ctype.h>  #include <ctype.h>
46  #include <limits.h>  #include <limits.h>
47  #include <setjmp.h>  #include <setjmp.h>
48    #include <stdarg.h>
49  #include <stddef.h>  #include <stddef.h>
50  #include <stdio.h>  #include <stdio.h>
51  #include <stdlib.h>  #include <stdlib.h>
52  #include <string.h>  #include <string.h>
53    
54    #ifndef PCRE_SPY
55    #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */
56    #endif
57    
58  #include "pcre.h"  #include "pcre.h"
59    
60    /* When compiling for use with the Virtual Pascal compiler, these functions
61    need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
62    option on the command line. */
63    
64    #ifdef VPCOMPAT
65    #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
66    #define memcpy(d,s,n)    _memcpy(d,s,n)
67    #define memmove(d,s,n)   _memmove(d,s,n)
68    #define memset(s,c,n)    _memset(s,c,n)
69    #else  /* VPCOMPAT */
70    
71    /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
72    define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
73    is set. Otherwise, include an emulating function for those systems that have
74    neither (there some non-Unix environments where this is the case). This assumes
75    that all calls to memmove are moving strings upwards in store, which is the
76    case in PCRE. */
77    
78    #if ! HAVE_MEMMOVE
79    #undef  memmove        /* some systems may have a macro */
80    #if HAVE_BCOPY
81    #define memmove(a, b, c) bcopy(b, a, c)
82    #else  /* HAVE_BCOPY */
83    void *
84    pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)
85    {
86    int i;
87    dest += n;
88    src += n;
89    for (i = 0; i < n; ++i) *(--dest) =  *(--src);
90    }
91    #define memmove(a, b, c) pcre_memmove(a, b, c)
92    #endif   /* not HAVE_BCOPY */
93    #endif   /* not HAVE_MEMMOVE */
94    #endif   /* not VPCOMPAT */
95    
96    
97    /* PCRE keeps offsets in its compiled code as 2-byte quantities by default.
98    These are used, for example, to link from the start of a subpattern to its
99    alternatives and its end. The use of 2 bytes per offset limits the size of the
100    compiled regex to around 64K, which is big enough for almost everybody.
101    However, I received a request for an even bigger limit. For this reason, and
102    also to make the code easier to maintain, the storing and loading of offsets
103    from the byte string is now handled by the macros that are defined here.
104    
105    The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
106    the config.h file, but can be overridden by using -D on the command line. This
107    is automated on Unix systems via the "configure" command. */
108    
109    #if LINK_SIZE == 2
110    
111    #define PUT(a,n,d)   \
112      (a[n] = (d) >> 8), \
113      (a[(n)+1] = (d) & 255)
114    
115    #define GET(a,n) \
116      (((a)[n] << 8) | (a)[(n)+1])
117    
118    #define MAX_PATTERN_SIZE (1 << 16)
119    
120    
121    #elif LINK_SIZE == 3
122    
123    #define PUT(a,n,d)       \
124      (a[n] = (d) >> 16),    \
125      (a[(n)+1] = (d) >> 8), \
126      (a[(n)+2] = (d) & 255)
127    
128    #define GET(a,n) \
129      (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
130    
131    #define MAX_PATTERN_SIZE (1 << 24)
132    
133    
134    #elif LINK_SIZE == 4
135    
136    #define PUT(a,n,d)        \
137      (a[n] = (d) >> 24),     \
138      (a[(n)+1] = (d) >> 16), \
139      (a[(n)+2] = (d) >> 8),  \
140      (a[(n)+3] = (d) & 255)
141    
142    #define GET(a,n) \
143      (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
144    
145    #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */
146    
147    
148    #else
149    #error LINK_SIZE must be either 2, 3, or 4
150    #endif
151    
152    
153    /* Convenience macro defined in terms of the others */
154    
155    #define PUTINC(a,n,d)   PUT(a,n,d), a += LINK_SIZE
156    
157    
158    /* PCRE uses some other 2-byte quantities that do not change when the size of
159    offsets changes. There are used for repeat counts and for other things such as
160    capturing parenthesis numbers in back references. */
161    
162    #define PUT2(a,n,d)   \
163      a[n] = (d) >> 8; \
164      a[(n)+1] = (d) & 255
165    
166    #define GET2(a,n) \
167      (((a)[n] << 8) | (a)[(n)+1])
168    
169    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2
170    
171    
172  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
173  Standard C system should have one. */  Standard C system should have one. */
174    
# Line 60  Standard C system should have one. */ Line 176  Standard C system should have one. */
176  #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))  #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
177  #endif  #endif
178    
179  /* Private options flags start at the most significant end of the two bytes.  /* These are the public options that can change during matching. */
180  The public options defined in pcre.h start at the least significant end. Make  
181  sure they don't overlap! */  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
182    
183  #define PCRE_FIRSTSET           0x8000  /* first_char is set */  /* Private options flags start at the most significant end of the four bytes,
184  #define PCRE_STARTLINE          0x4000  /* start after \n for multiline */  but skip the top bit so we can use ints for convenience without getting tangled
185  #define PCRE_COMPILED_CASELESS  0x2000  /* like it says */  with negative values. The public options defined in pcre.h start at the least
186    significant end. Make sure they don't overlap, though now that we have expanded
187    to four bytes there is plenty of space. */
188    
189    #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */
190    #define PCRE_REQCHSET      0x20000000  /* req_byte is set */
191    #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
192    #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */
193    
194  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
195    
196  #define PCRE_STUDY_CASELESS 0x01     /* study was caseless */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
 #define PCRE_STUDY_MAPPED   0x02     /* a map of starting chars exists */  
197    
198  /* Masks for identifying the public options: all permitted at compile time,  /* Masks for identifying the public options which are permitted at compile
199  only some permitted at run or study time. */  time, run time or study time, respectively. */
200    
201  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
202    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
203     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
204       PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK)
205    
206  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
207    (PCRE_CASELESS|PCRE_ANCHORED|PCRE_MULTILINE|PCRE_NOTBOL|PCRE_NOTEOL| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK)
    PCRE_DOTALL|PCRE_DOLLAR_ENDONLY)  
208    
209  #define PUBLIC_STUDY_OPTIONS (PCRE_CASELESS)  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
210    
211  /* Magic number to provide a small check against being handed junk. */  /* Magic number to provide a small check against being handed junk. */
212    
213  #define MAGIC_NUMBER  0x50435245   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
214    
215    /* Negative values for the firstchar and reqchar variables */
216    
217    #define REQ_UNSET (-2)
218    #define REQ_NONE  (-1)
219    
220    /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a
221    variable-length repeat, or a anything other than literal characters. */
222    
223    #define REQ_CASELESS 0x0100    /* indicates caselessness */
224    #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
225    
226  /* Miscellaneous definitions */  /* Miscellaneous definitions */
227    
# Line 97  typedef int BOOL; Line 230  typedef int BOOL;
230  #define FALSE   0  #define FALSE   0
231  #define TRUE    1  #define TRUE    1
232    
233    /* Escape items that are just an encoding of a particular data value. Note that
234    ESC_n is defined as yet another macro, which is set in config.h to either \n
235    (the default) or \r (which some people want). */
236    
237    #ifndef ESC_e
238    #define ESC_e 27
239    #endif
240    
241    #ifndef ESC_f
242    #define ESC_f '\f'
243    #endif
244    
245    #ifndef ESC_n
246    #define ESC_n NEWLINE
247    #endif
248    
249    #ifndef ESC_r
250    #define ESC_r '\r'
251    #endif
252    
253    /* We can't officially use ESC_t because it is a POSIX reserved identifier
254    (presumably because of all the others like size_t). */
255    
256    #ifndef ESC_tee
257    #define ESC_tee '\t'
258    #endif
259    
260  /* These are escaped items that aren't just an encoding of a particular data  /* These are escaped items that aren't just an encoding of a particular data
261  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
262  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
263  definitions below, up to ESC_Z. The final one must be ESC_REF as subsequent  definitions below, up to ESC_z. There's a dummy for OP_ANY because it
264  values are used for \1, \2, \3, etc. There is a test in the code for an escape  corresponds to "." rather than an escape sequence. The final one must be
265  greater than ESC_b and less than ESC_X to detect the types that may be  ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two
266  repeated. If any new escapes are put in-between that don't consume a character,  tests in the code for an escape greater than ESC_b and less than ESC_Z to
267  that code will have to change. */  detect the types that may be repeated. These are the types that consume a
268    character. If any new escapes are put in between that don't consume a
269  enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,  character, that code will have to change. */
270    
271                      /* These are not Perl escapes, so can't appear in the */  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,
272         ESC_X,       /* simple table-lookup because they must be conditional */         ESC_w, ESC_dum1, ESC_C, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_REF };
273                      /* on PCRE_EXTRA. */  
274         ESC_Z,  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
275         ESC_REF };  contain UTF-8 characters with values greater than 255. */
276    
277    #define XCL_NOT    0x01    /* Flag: this is a negative class */
278    #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
279    
280    #define XCL_END       0    /* Marks end of individual items */
281    #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
282    #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
283    
284    
285  /* 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
286  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
287  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.
288    Note that whenever this list is updated, the two macro definitions that follow
289    must also be updated to match. */
290    
291  enum {  enum {
292    OP_END,            /* End of pattern */    OP_END,            /* 0 End of pattern */
293    
294    /* Values corresponding to backslashed metacharacters */    /* Values corresponding to backslashed metacharacters */
295    
296    OP_SOD,            /* Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
297    OP_NOT_WORD_BOUNDARY,  /* \B */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
298    OP_WORD_BOUNDARY,      /* \b */    OP_NOT_WORD_BOUNDARY,  /*  3 \B */
299    OP_NOT_DIGIT,          /* \D */    OP_WORD_BOUNDARY,      /*  4 \b */
300    OP_DIGIT,              /* \d */    OP_NOT_DIGIT,          /*  5 \D */
301    OP_NOT_WHITESPACE,     /* \S */    OP_DIGIT,              /*  6 \d */
302    OP_WHITESPACE,         /* \s */    OP_NOT_WHITESPACE,     /*  7 \S */
303    OP_NOT_WORDCHAR,       /* \W */    OP_WHITESPACE,         /*  8 \s */
304    OP_WORDCHAR,           /* \w */    OP_NOT_WORDCHAR,       /*  9 \W */
305    OP_CUT,            /* The analogue of Prolog's "cut" operation (extension) */    OP_WORDCHAR,           /* 10 \w */
306    OP_EOD,            /* End of data: \Z. */    OP_ANY,            /* 11 Match any character */
307      OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */
308    OP_CIRC,           /* Start of line - varies with multiline switch */    OP_EODN,           /* 13 End of data or \n at end of data: \Z. */
309    OP_DOLL,           /* End of line - varies with multiline switch */    OP_EOD,            /* 14 End of data: \z */
310    OP_ANY,            /* Match any character */  
311    OP_CHARS,          /* Match string of characters */    OP_OPT,            /* 15 Set runtime options */
312    OP_NOT,            /* Match anything but the following char */    OP_CIRC,           /* 16 Start of line - varies with multiline switch */
313      OP_DOLL,           /* 17 End of line - varies with multiline switch */
314    OP_STAR,           /* The maximizing and minimizing versions of */    OP_CHARS,          /* 18 Match string of characters */
315    OP_MINSTAR,        /* all these opcodes must come in pairs, with */    OP_NOT,            /* 19 Match anything but the following char */
316    OP_PLUS,           /* the minimizing one second. */  
317    OP_MINPLUS,        /* This first set applies to single characters */    OP_STAR,           /* 20 The maximizing and minimizing versions of */
318    OP_QUERY,    OP_MINSTAR,        /* 21 all these opcodes must come in pairs, with */
319    OP_MINQUERY,    OP_PLUS,           /* 22 the minimizing one second. */
320    OP_UPTO,           /* From 0 to n matches */    OP_MINPLUS,        /* 23 This first set applies to single characters */
321    OP_MINUPTO,    OP_QUERY,          /* 24 */
322    OP_EXACT,          /* Exactly n matches */    OP_MINQUERY,       /* 25 */
323      OP_UPTO,           /* 26 From 0 to n matches */
324    OP_NOTSTAR,        /* The maximizing and minimizing versions of */    OP_MINUPTO,        /* 27 */
325    OP_NOTMINSTAR,     /* all these opcodes must come in pairs, with */    OP_EXACT,          /* 28 Exactly n matches */
326    OP_NOTPLUS,        /* the minimizing one second. */  
327    OP_NOTMINPLUS,     /* This first set applies to "not" single characters */    OP_NOTSTAR,        /* 29 The maximizing and minimizing versions of */
328    OP_NOTQUERY,    OP_NOTMINSTAR,     /* 30 all these opcodes must come in pairs, with */
329    OP_NOTMINQUERY,    OP_NOTPLUS,        /* 31 the minimizing one second. */
330    OP_NOTUPTO,        /* From 0 to n matches */    OP_NOTMINPLUS,     /* 32 This set applies to "not" single characters */
331    OP_NOTMINUPTO,    OP_NOTQUERY,       /* 33 */
332    OP_NOTEXACT,       /* Exactly n matches */    OP_NOTMINQUERY,    /* 34 */
333      OP_NOTUPTO,        /* 35 From 0 to n matches */
334    OP_TYPESTAR,       /* The maximizing and minimizing versions of */    OP_NOTMINUPTO,     /* 36 */
335    OP_TYPEMINSTAR,    /* all these opcodes must come in pairs, with */    OP_NOTEXACT,       /* 37 Exactly n matches */
336    OP_TYPEPLUS,       /* the minimizing one second. These codes must */  
337    OP_TYPEMINPLUS,    /* be in exactly the same order as those above. */    OP_TYPESTAR,       /* 38 The maximizing and minimizing versions of */
338    OP_TYPEQUERY,      /* This set applies to character types such as \d */    OP_TYPEMINSTAR,    /* 39 all these opcodes must come in pairs, with */
339    OP_TYPEMINQUERY,    OP_TYPEPLUS,       /* 40 the minimizing one second. These codes must */
340    OP_TYPEUPTO,       /* From 0 to n matches */    OP_TYPEMINPLUS,    /* 41 be in exactly the same order as those above. */
341    OP_TYPEMINUPTO,    OP_TYPEQUERY,      /* 42 This set applies to character types such as \d */
342    OP_TYPEEXACT,      /* Exactly n matches */    OP_TYPEMINQUERY,   /* 43 */
343      OP_TYPEUPTO,       /* 44 From 0 to n matches */
344    OP_CRSTAR,         /* The maximizing and minimizing versions of */    OP_TYPEMINUPTO,    /* 45 */
345    OP_CRMINSTAR,      /* all these opcodes must come in pairs, with */    OP_TYPEEXACT,      /* 46 Exactly n matches */
346    OP_CRPLUS,         /* the minimizing one second. These codes must */  
347    OP_CRMINPLUS,      /* be in exactly the same order as those above. */    OP_CRSTAR,         /* 47 The maximizing and minimizing versions of */
348    OP_CRQUERY,        /* These are for character classes and back refs */    OP_CRMINSTAR,      /* 48 all these opcodes must come in pairs, with */
349    OP_CRMINQUERY,    OP_CRPLUS,         /* 49 the minimizing one second. These codes must */
350    OP_CRRANGE,        /* These are different to the three seta above. */    OP_CRMINPLUS,      /* 50 be in exactly the same order as those above. */
351    OP_CRMINRANGE,    OP_CRQUERY,        /* 51 These are for character classes and back refs */
352      OP_CRMINQUERY,     /* 52 */
353    OP_CLASS,          /* Match a character class */    OP_CRRANGE,        /* 53 These are different to the three seta above. */
354    OP_NEGCLASS,       /* Match a character class, specified negatively */    OP_CRMINRANGE,     /* 54 */
355    OP_REF,            /* Match a back reference */  
356      OP_CLASS,          /* 55 Match a character class, chars < 256 only */
357    OP_ALT,            /* Start of alternation */    OP_NCLASS,         /* 56 Same, but the bitmap was created from a negative
358    OP_KET,            /* End of group that doesn't have an unbounded repeat */                             class - the difference is relevant only when a UTF-8
359    OP_KETRMAX,        /* These two must remain together and in this */                             character > 255 is encountered. */
360    OP_KETRMIN,        /* order. They are for groups the repeat for ever. */  
361      OP_XCLASS,         /* 57 Extended class for handling UTF-8 chars within the
362    OP_ASSERT,                             class. This does both positive and negative. */
363    OP_ASSERT_NOT,  
364    OP_ONCE,           /* Once matched, don't back up into the subpattern */    OP_REF,            /* 58 Match a back reference */
365      OP_RECURSE,        /* 59 Match a numbered subpattern (possibly recursive) */
366      OP_CALLOUT,        /* 60 Call out to external function if provided */
367    
368      OP_ALT,            /* 61 Start of alternation */
369      OP_KET,            /* 62 End of group that doesn't have an unbounded repeat */
370      OP_KETRMAX,        /* 63 These two must remain together and in this */
371      OP_KETRMIN,        /* 64 order. They are for groups the repeat for ever. */
372    
373      /* The assertions must come before ONCE and COND */
374    
375      OP_ASSERT,         /* 65 Positive lookahead */
376      OP_ASSERT_NOT,     /* 66 Negative lookahead */
377      OP_ASSERTBACK,     /* 67 Positive lookbehind */
378      OP_ASSERTBACK_NOT, /* 68 Negative lookbehind */
379      OP_REVERSE,        /* 69 Move pointer back - used in lookbehind assertions */
380    
381      /* ONCE and COND must come after the assertions, with ONCE first, as there's
382      a test for >= ONCE for a subpattern that isn't an assertion. */
383    
384      OP_ONCE,           /* 70 Once matched, don't back up into the subpattern */
385      OP_COND,           /* 71 Conditional group */
386      OP_CREF,           /* 72 Used to hold an extraction string number (cond ref) */
387    
388      OP_BRAZERO,        /* 73 These two must remain together and in this */
389      OP_BRAMINZERO,     /* 74 order. */
390    
391      OP_BRANUMBER,      /* 75 Used for extracting brackets whose number is greater
392                               than can fit into an opcode. */
393    
394      OP_BRA             /* 76 This and greater values are used for brackets that
395                               extract substrings up to a basic limit. After that,
396                               use is made of OP_BRANUMBER. */
397    };
398    
399    OP_BRAZERO,        /* These two must remain together and in this */  /* WARNING: There is an implicit assumption in study.c that all opcodes are
400    OP_BRAMINZERO,     /* order. */  less than 128 in value. This makes handling UTF-8 character sequences easier.
401    */
402    
   OP_BRA             /* This and greater values are used for brackets that  
                         extract substrings. */  
 };  
403    
404  /* The highest extraction number. This is limited by the number of opcodes  /* This macro defines textual names for all the opcodes. There are used only
405  left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */  for debugging, in pcre.c when DEBUG is defined, and also in pcretest.c. The
406    macro is referenced only in printint.c. */
407    
408    #define OP_NAME_LIST \
409      "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \
410      "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte", "\\Z", "\\z",     \
411      "Opt", "^", "$", "chars", "not",                                \
412      "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
413      "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
414      "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
415      "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
416      "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
417      "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
418      "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\
419      "Brazero", "Braminzero", "Branumber", "Bra"
420    
421    
422    /* This macro defines the length of fixed length operations in the compiled
423    regex. The lengths are used when searching for specific things, and also in the
424    debugging printing of a compiled regex. We use a macro so that it can be
425    incorporated both into pcre.c and pcretest.c without being publicly exposed.
426    
427    As things have been extended, some of these are no longer fixed lenths, but are
428    minima instead. For example, the length of a single-character repeat may vary
429    in UTF-8 mode. The code that uses this table must know about such things. */
430    
431    #define OP_LENGTHS \
432      1,                             /* End                                    */ \
433      1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \
434      1, 1, 1, 1, 2, 1, 1,           /* Any, Anybyte, \Z, \z, Opt, ^, $        */ \
435      2,                             /* Chars - the minimum length             */ \
436      2,                             /* not                                    */ \
437      /* Positive single-char repeats                            ** These are  */ \
438      2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
439      4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
440      /* Negative single-char repeats - only for chars < 256                   */ \
441      2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
442      4, 4, 4,                       /* NOT upto, minupto, exact               */ \
443      /* Positive type repeats                                                 */ \
444      2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
445      4, 4, 4,                       /* Type upto, minupto, exact              */ \
446      /* Character class & ref repeats                                         */ \
447      1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
448      5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
449     33,                             /* CLASS                                  */ \
450     33,                             /* NCLASS                                 */ \
451      0,                             /* XCLASS - variable length               */ \
452      3,                             /* REF                                    */ \
453      1+LINK_SIZE,                   /* RECURSE                                */ \
454      2,                             /* CALLOUT                                */ \
455      1+LINK_SIZE,                   /* Alt                                    */ \
456      1+LINK_SIZE,                   /* Ket                                    */ \
457      1+LINK_SIZE,                   /* KetRmax                                */ \
458      1+LINK_SIZE,                   /* KetRmin                                */ \
459      1+LINK_SIZE,                   /* Assert                                 */ \
460      1+LINK_SIZE,                   /* Assert not                             */ \
461      1+LINK_SIZE,                   /* Assert behind                          */ \
462      1+LINK_SIZE,                   /* Assert behind not                      */ \
463      1+LINK_SIZE,                   /* Reverse                                */ \
464      1+LINK_SIZE,                   /* Once                                   */ \
465      1+LINK_SIZE,                   /* COND                                   */ \
466      3,                             /* CREF                                   */ \
467      1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
468      3,                             /* BRANUMBER                              */ \
469      1+LINK_SIZE                    /* BRA                                    */ \
470    
471    
472    /* The highest extraction number before we have to start using additional
473    bytes. (Originally PCRE didn't have support for extraction counts highter than
474    this number.) The value is limited by the number of opcodes left after OP_BRA,
475    i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional
476    opcodes. */
477    
478  #define EXTRACT_MAX  99  #define EXTRACT_BASIC_MAX  150
479    
480    /* A magic value for OP_CREF to indicate the "in recursion" condition. */
481    
482    #define CREF_RECURSE  0xffff
483    
484  /* The texts of compile-time error messages are defined as macros here so that  /* The texts of compile-time error messages are defined as macros here so that
485  they can be accessed by the POSIX wrapper and converted into error codes.  Yes,  they can be accessed by the POSIX wrapper and converted into error codes.  Yes,
# Line 222  just to accommodate the POSIX wrapper. * Line 498  just to accommodate the POSIX wrapper. *
498  #define ERR10 "operand of unlimited repeat could match the empty string"  #define ERR10 "operand of unlimited repeat could match the empty string"
499  #define ERR11 "internal error: unexpected repeat"  #define ERR11 "internal error: unexpected repeat"
500  #define ERR12 "unrecognized character after (?"  #define ERR12 "unrecognized character after (?"
501  #define ERR13 "too many capturing parenthesized sub-patterns"  #define ERR13 "POSIX named classes are supported only within a class"
502  #define ERR14 "missing )"  #define ERR14 "missing )"
503  #define ERR15 "back reference to non-existent subpattern"  #define ERR15 "reference to non-existent subpattern"
504  #define ERR16 "erroffset passed as NULL"  #define ERR16 "erroffset passed as NULL"
505  #define ERR17 "unknown option bit(s) set"  #define ERR17 "unknown option bit(s) set"
506  #define ERR18 "missing ) after comment"  #define ERR18 "missing ) after comment"
507  #define ERR19 "too many sets of parentheses"  #define ERR19 "parentheses nested too deeply"
508  #define ERR20 "regular expression too large"  #define ERR20 "regular expression too large"
509  #define ERR21 "failed to get memory"  #define ERR21 "failed to get memory"
510  #define ERR22 "unmatched brackets"  #define ERR22 "unmatched parentheses"
511  #define ERR23 "internal error: code overflow"  #define ERR23 "internal error: code overflow"
512    #define ERR24 "unrecognized character after (?<"
513    #define ERR25 "lookbehind assertion is not fixed length"
514    #define ERR26 "malformed number after (?("
515    #define ERR27 "conditional group contains more than two branches"
516    #define ERR28 "assertion expected after (?("
517    #define ERR29 "(?R or (?digits must be followed by )"
518    #define ERR30 "unknown POSIX class name"
519    #define ERR31 "POSIX collating elements are not supported"
520    #define ERR32 "this version of PCRE is not compiled with PCRE_UTF8 support"
521    #define ERR33 "spare error"
522    #define ERR34 "character value in \\x{...} sequence is too large"
523    #define ERR35 "invalid condition (?(0)"
524    #define ERR36 "\\C not allowed in lookbehind assertion"
525    #define ERR37 "PCRE does not support \\L, \\l, \\N, \\P, \\p, \\U, \\u, or \\X"
526    #define ERR38 "number after (?C is > 255"
527    #define ERR39 "closing ) for (?C expected"
528    #define ERR40 "recursive call could loop indefinitely"
529    #define ERR41 "unrecognized character after (?P"
530    #define ERR42 "syntax error after (?P"
531    #define ERR43 "two named groups have the same name"
532    #define ERR44 "invalid UTF-8 string"
533    
534  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
535  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
# Line 243  Unix, where it is defined in sys/types, Line 540  Unix, where it is defined in sys/types,
540    
541  typedef unsigned char uschar;  typedef unsigned char uschar;
542    
543  /* The real format of the start of the pcre block; the actual code vector  /* The real format of the start of the pcre block; the index of names and the
544  runs on as long as necessary after the end. */  code vector run on as long as necessary after the end. */
545    
546  typedef struct real_pcre {  typedef struct real_pcre {
547    unsigned int  magic_number;    unsigned long int magic_number;
548    unsigned short int options;    size_t size;                        /* Total that was malloced */
549    unsigned char top_bracket;    const unsigned char *tables;        /* Pointer to tables */
550    unsigned char top_backref;    unsigned long int options;
551    unsigned char first_char;    unsigned short int top_bracket;
552    unsigned char code[1];    unsigned short int top_backref;
553      unsigned short int first_byte;
554      unsigned short int req_byte;
555      unsigned short int name_entry_size; /* Size of any name items; 0 => none */
556      unsigned short int name_count;      /* Number of name items */
557  } real_pcre;  } real_pcre;
558    
559  /* The real format of the extra block returned by pcre_study(). */  /* The format of the block used to store data from pcre_study(). */
   
 typedef struct real_pcre_extra {  
   unsigned char options;  
   unsigned char start_bits[32];  
 } real_pcre_extra;  
   
 /* Global tables from chartables.c */  
560    
561  extern uschar pcre_lcc[];  typedef struct pcre_study_data {
562  extern uschar pcre_fcc[];    size_t size;                        /* Total that was malloced */
563  extern uschar pcre_cbits[];    uschar options;
564  extern uschar pcre_ctypes[];    uschar start_bits[32];
565    } pcre_study_data;
566    
567    /* Structure for passing "static" information around between the functions
568    doing the compiling, so that they are thread-safe. */
569    
570    typedef struct compile_data {
571      const uschar *lcc;            /* Points to lower casing table */
572      const uschar *fcc;            /* Points to case-flipping table */
573      const uschar *cbits;          /* Points to character type table */
574      const uschar *ctypes;         /* Points to table of type maps */
575      const uschar *start_code;     /* The start of the compiled code */
576      uschar *name_table;           /* The name/number table */
577      int  names_found;             /* Number of entries so far */
578      int  name_entry_size;         /* Size of each entry */
579      int  top_backref;             /* Maximum back reference */
580      unsigned int backref_map;     /* Bitmap of low back refs */
581      int  req_varyopt;             /* "After variable item" flag for reqbyte */
582    } compile_data;
583    
584    /* Structure for maintaining a chain of pointers to the currently incomplete
585    branches, for testing for left recursion. */
586    
587    typedef struct branch_chain {
588      struct branch_chain *outer;
589      uschar *current;
590    } branch_chain;
591    
592    /* Structure for items in a linked list that represents an explicit recursive
593    call within the pattern. */
594    
595    typedef struct recursion_info {
596      struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
597      int group_num;                /* Number of group that was called */
598      const uschar *after_call;     /* "Return value": points after the call in the expr */
599      const uschar *save_start;     /* Old value of md->start_match */
600      int *offset_save;             /* Pointer to start of saved offsets */
601      int saved_max;                /* Number of saved offsets */
602    } recursion_info;
603    
604    /* When compiling in a mode that doesn't use recursive calls to match(),
605    a structure is used to remember local variables on the heap. It is defined in
606    pcre.c, close to the match() function, so that it is easy to keep it in step
607    with any changes of local variable. However, the pointer to the current frame
608    must be saved in some "static" place over a longjmp(). We declare the
609    structure here so that we can put a pointer in the match_data structure.
610    NOTE: This isn't used for a "normal" compilation of pcre. */
611    
612    struct heapframe;
613    
614    /* Structure for passing "static" information around between the functions
615    doing the matching, so that they are thread-safe. */
616    
617    typedef struct match_data {
618      unsigned long int match_call_count; /* As it says */
619      unsigned long int match_limit;/* As it says */
620      int   *offset_vector;         /* Offset vector */
621      int    offset_end;            /* One past the end */
622      int    offset_max;            /* The maximum usable for return data */
623      const uschar *lcc;            /* Points to lower casing table */
624      const uschar *ctypes;         /* Points to table of type maps */
625      BOOL   offset_overflow;       /* Set if too many extractions */
626      BOOL   notbol;                /* NOTBOL flag */
627      BOOL   noteol;                /* NOTEOL flag */
628      BOOL   utf8;                  /* UTF8 flag */
629      BOOL   endonly;               /* Dollar not before final \n */
630      BOOL   notempty;              /* Empty string match not wanted */
631      const uschar *start_code;     /* For use when recursing */
632      const uschar *start_subject;  /* Start of the subject string */
633      const uschar *end_subject;    /* End of the subject string */
634      const uschar *start_match;    /* Start of this match attempt */
635      const uschar *end_match_ptr;  /* Subject position at end match */
636      int    end_offset_top;        /* Highwater mark at end of match */
637      int    capture_last;          /* Most recent capture number */
638      int    start_offset;          /* The start offset value */
639      recursion_info *recursive;    /* Linked list of recursion data */
640      void  *callout_data;          /* To pass back to callouts */
641      struct heapframe *thisframe;  /* Used only when compiling for no recursion */
642    } match_data;
643    
644  /* Bit definitions for entries in pcre_ctypes[]. */  /* Bit definitions for entries in the pcre_ctypes table. */
645    
646  #define ctype_space   0x01  #define ctype_space   0x01
647  #define ctype_letter  0x02  #define ctype_letter  0x02
# Line 278  extern uschar pcre_ctypes[]; Line 650  extern uschar pcre_ctypes[];
650  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphameric or '_' */
651  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
652    
653  /* Offsets for the bitmap tables */  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
654    of bits for a class map. Some classes are built by combining these tables. */
655    
656  #define cbit_digit    0  #define cbit_space     0      /* [:space:] or \s */
657  #define cbit_letter  32  #define cbit_xdigit   32      /* [:xdigit:] */
658  #define cbit_word    64  #define cbit_digit    64      /* [:digit:] or \d */
659  #define cbit_space   96  #define cbit_upper    96      /* [:upper:] */
660  #define cbit_length 128      /* Length of the cbits table */  #define cbit_lower   128      /* [:lower:] */
661    #define cbit_word    160      /* [:word:] or \w */
662    #define cbit_graph   192      /* [:graph:] */
663    #define cbit_print   224      /* [:print:] */
664    #define cbit_punct   256      /* [:punct:] */
665    #define cbit_cntrl   288      /* [:cntrl:] */
666    #define cbit_length  320      /* Length of the cbits table */
667    
668    /* Offsets of the various tables from the base tables pointer, and
669    total length. */
670    
671    #define lcc_offset      0
672    #define fcc_offset    256
673    #define cbits_offset  512
674    #define ctypes_offset (cbits_offset + cbit_length)
675    #define tables_length (ctypes_offset + 256)
676    
677  /* End of internal.h */  /* End of internal.h */

Legend:
Removed from v.19  
changed lines
  Added in v.74

  ViewVC Help
Powered by ViewVC 1.1.5