/[pcre]/code/trunk/pcre_dfa_exec.c
ViewVC logotype

Diff of /code/trunk/pcre_dfa_exec.c

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

revision 168 by ph10, Tue May 29 15:18:18 2007 UTC revision 518 by ph10, Tue May 18 15:47:01 2010 UTC
# Line 3  Line 3 
3  *************************************************/  *************************************************/
4    
5  /* PCRE is a library of functions to support regular expressions whose syntax  /* PCRE is a library of functions to support regular expressions whose syntax
6  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 (but see
7    below for why this module is different).
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2007 University of Cambridge             Copyright (c) 1997-2010 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 44  FSM). This is NOT Perl- compatible, but Line 45  FSM). This is NOT Perl- compatible, but
45  applications. */  applications. */
46    
47    
48    /* NOTE ABOUT PERFORMANCE: A user of this function sent some code that improved
49    the performance of his patterns greatly. I could not use it as it stood, as it
50    was not thread safe, and made assumptions about pattern sizes. Also, it caused
51    test 7 to loop, and test 9 to crash with a segfault.
52    
53    The issue is the check for duplicate states, which is done by a simple linear
54    search up the state list. (Grep for "duplicate" below to find the code.) For
55    many patterns, there will never be many states active at one time, so a simple
56    linear search is fine. In patterns that have many active states, it might be a
57    bottleneck. The suggested code used an indexing scheme to remember which states
58    had previously been used for each character, and avoided the linear search when
59    it knew there was no chance of a duplicate. This was implemented when adding
60    states to the state lists.
61    
62    I wrote some thread-safe, not-limited code to try something similar at the time
63    of checking for duplicates (instead of when adding states), using index vectors
64    on the stack. It did give a 13% improvement with one specially constructed
65    pattern for certain subject strings, but on other strings and on many of the
66    simpler patterns in the test suite it did worse. The major problem, I think,
67    was the extra time to initialize the index. This had to be done for each call
68    of internal_dfa_exec(). (The supplied patch used a static vector, initialized
69    only once - I suspect this was the cause of the problems with the tests.)
70    
71    Overall, I concluded that the gains in some cases did not outweigh the losses
72    in others, so I abandoned this code. */
73    
74    
75    
76    #ifdef HAVE_CONFIG_H
77    #include "config.h"
78    #endif
79    
80  #define NLBLOCK md             /* Block containing newline information */  #define NLBLOCK md             /* Block containing newline information */
81  #define PSSTART start_subject  /* Field containing processed string start */  #define PSSTART start_subject  /* Field containing processed string start */
82  #define PSEND   end_subject    /* Field containing processed string end */  #define PSEND   end_subject    /* Field containing processed string end */
# Line 56  applications. */ Line 89  applications. */
89  #define SP "                   "  #define SP "                   "
90    
91    
   
92  /*************************************************  /*************************************************
93  *      Code parameters and static tables         *  *      Code parameters and static tables         *
94  *************************************************/  *************************************************/
95    
96  /* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes  /* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes
97  into others, under special conditions. A gap of 20 between the blocks should be  into others, under special conditions. A gap of 20 between the blocks should be
98  enough. */  enough. The resulting opcodes don't have to be less than 256 because they are
99    never stored, so we push them well clear of the normal opcodes. */
100    
101  #define OP_PROP_EXTRA 100  #define OP_PROP_EXTRA       300
102  #define OP_EXTUNI_EXTRA 120  #define OP_EXTUNI_EXTRA     320
103  #define OP_ANYNL_EXTRA 140  #define OP_ANYNL_EXTRA      340
104    #define OP_HSPACE_EXTRA     360
105    #define OP_VSPACE_EXTRA     380
106    
107    
108  /* This table identifies those opcodes that are followed immediately by a  /* This table identifies those opcodes that are followed immediately by a
109  character that is to be tested in some way. This makes is possible to  character that is to be tested in some way. This makes it possible to
110  centralize the loading of these characters. In the case of Type * etc, the  centralize the loading of these characters. In the case of Type * etc, the
111  "character" is the opcode for \D, \d, \S, \s, \W, or \w, which will always be a  "character" is the opcode for \D, \d, \S, \s, \W, or \w, which will always be a
112  small value. ***NOTE*** If the start of this table is modified, the two tables  small value. Non-zero values in the table are the offsets from the opcode where
113  that follow must also be modified. */  the character is to be found. ***NOTE*** If the start of this table is
114    modified, the three tables that follow must also be modified. */
115    
116  static uschar coptable[] = {  static const uschar coptable[] = {
117    0,                             /* End                                    */    0,                             /* End                                    */
118    0, 0, 0, 0, 0,                 /* \A, \G, \K, \B, \b                     */    0, 0, 0, 0, 0,                 /* \A, \G, \K, \B, \b                     */
119    0, 0, 0, 0, 0, 0,              /* \D, \d, \S, \s, \W, \w                 */    0, 0, 0, 0, 0, 0,              /* \D, \d, \S, \s, \W, \w                 */
120    0, 0,                          /* Any, Anybyte                           */    0, 0, 0,                       /* Any, AllAny, Anybyte                   */
121    0, 0, 0, 0,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */    0, 0,                          /* \P, \p                                 */
122      0, 0, 0, 0, 0,                 /* \R, \H, \h, \V, \v                     */
123      0,                             /* \X                                     */
124    0, 0, 0, 0, 0,                 /* \Z, \z, Opt, ^, $                      */    0, 0, 0, 0, 0,                 /* \Z, \z, Opt, ^, $                      */
125    1,                             /* Char                                   */    1,                             /* Char                                   */
126    1,                             /* Charnc                                 */    1,                             /* Charnc                                 */
# Line 119  static uschar coptable[] = { Line 157  static uschar coptable[] = {
157    0,                             /* Reverse                                */    0,                             /* Reverse                                */
158    0, 0, 0, 0,                    /* ONCE, BRA, CBRA, COND                  */    0, 0, 0, 0,                    /* ONCE, BRA, CBRA, COND                  */
159    0, 0, 0,                       /* SBRA, SCBRA, SCOND                     */    0, 0, 0,                       /* SBRA, SCBRA, SCOND                     */
160    0,                             /* CREF                                   */    0, 0,                          /* CREF, NCREF                            */
161    0,                             /* RREF                                   */    0, 0,                          /* RREF, NRREF                            */
162    0,                             /* DEF                                    */    0,                             /* DEF                                    */
163    0, 0                           /* BRAZERO, BRAMINZERO                    */    0, 0,                          /* BRAZERO, BRAMINZERO                    */
164      0, 0, 0,                       /* MARK, PRUNE, PRUNE_ARG,                */
165      0, 0, 0, 0,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG,        */
166      0, 0, 0, 0, 0                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */
167    };
168    
169    /* This table identifies those opcodes that inspect a character. It is used to
170    remember the fact that a character could have been inspected when the end of
171    the subject is reached. ***NOTE*** If the start of this table is modified, the
172    two tables that follow must also be modified. */
173    
174    static const uschar poptable[] = {
175      0,                             /* End                                    */
176      0, 0, 0, 1, 1,                 /* \A, \G, \K, \B, \b                     */
177      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */
178      1, 1, 1,                       /* Any, AllAny, Anybyte                   */
179      1, 1,                          /* \P, \p                                 */
180      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */
181      1,                             /* \X                                     */
182      0, 0, 0, 0, 0,                 /* \Z, \z, Opt, ^, $                      */
183      1,                             /* Char                                   */
184      1,                             /* Charnc                                 */
185      1,                             /* not                                    */
186      /* Positive single-char repeats                                          */
187      1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */
188      1, 1, 1,                       /* upto, minupto, exact                   */
189      1, 1, 1, 1,                    /* *+, ++, ?+, upto+                      */
190      /* Negative single-char repeats - only for chars < 256                   */
191      1, 1, 1, 1, 1, 1,              /* NOT *, *?, +, +?, ?, ??                */
192      1, 1, 1,                       /* NOT upto, minupto, exact               */
193      1, 1, 1, 1,                    /* NOT *+, ++, ?+, upto+                  */
194      /* Positive type repeats                                                 */
195      1, 1, 1, 1, 1, 1,              /* Type *, *?, +, +?, ?, ??               */
196      1, 1, 1,                       /* Type upto, minupto, exact              */
197      1, 1, 1, 1,                    /* Type *+, ++, ?+, upto+                 */
198      /* Character class & ref repeats                                         */
199      1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */
200      1, 1,                          /* CRRANGE, CRMINRANGE                    */
201      1,                             /* CLASS                                  */
202      1,                             /* NCLASS                                 */
203      1,                             /* XCLASS - variable length               */
204      0,                             /* REF                                    */
205      0,                             /* RECURSE                                */
206      0,                             /* CALLOUT                                */
207      0,                             /* Alt                                    */
208      0,                             /* Ket                                    */
209      0,                             /* KetRmax                                */
210      0,                             /* KetRmin                                */
211      0,                             /* Assert                                 */
212      0,                             /* Assert not                             */
213      0,                             /* Assert behind                          */
214      0,                             /* Assert behind not                      */
215      0,                             /* Reverse                                */
216      0, 0, 0, 0,                    /* ONCE, BRA, CBRA, COND                  */
217      0, 0, 0,                       /* SBRA, SCBRA, SCOND                     */
218      0, 0,                          /* CREF, NCREF                            */
219      0, 0,                          /* RREF, NRREF                            */
220      0,                             /* DEF                                    */
221      0, 0,                          /* BRAZERO, BRAMINZERO                    */
222      0, 0, 0,                       /* MARK, PRUNE, PRUNE_ARG,                */
223      0, 0, 0, 0,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG,        */
224      0, 0, 0, 0, 0                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */
225  };  };
226    
227  /* These 2 tables allow for compact code for testing for \D, \d, \S, \s, \W,  /* These 2 tables allow for compact code for testing for \D, \d, \S, \s, \W,
228  and \w */  and \w */
229    
230  static uschar toptable1[] = {  static const uschar toptable1[] = {
231    0, 0, 0, 0, 0, 0,    0, 0, 0, 0, 0, 0,
232    ctype_digit, ctype_digit,    ctype_digit, ctype_digit,
233    ctype_space, ctype_space,    ctype_space, ctype_space,
234    ctype_word,  ctype_word,    ctype_word,  ctype_word,
235    0                               /* OP_ANY */    0, 0                            /* OP_ANY, OP_ALLANY */
236  };  };
237    
238  static uschar toptable2[] = {  static const uschar toptable2[] = {
239    0, 0, 0, 0, 0, 0,    0, 0, 0, 0, 0, 0,
240    ctype_digit, 0,    ctype_digit, 0,
241    ctype_space, 0,    ctype_space, 0,
242    ctype_word,  0,    ctype_word,  0,
243    1                               /* OP_ANY */    1, 1                            /* OP_ANY, OP_ALLANY */
244  };  };
245    
246    
# Line 160  typedef struct stateblock { Line 259  typedef struct stateblock {
259  #define INTS_PER_STATEBLOCK  (sizeof(stateblock)/sizeof(int))  #define INTS_PER_STATEBLOCK  (sizeof(stateblock)/sizeof(int))
260    
261    
262  #ifdef DEBUG  #ifdef PCRE_DEBUG
263  /*************************************************  /*************************************************
264  *             Print character string             *  *             Print character string             *
265  *************************************************/  *************************************************/
# Line 213  Arguments: Line 312  Arguments:
312    rlevel            function call recursion level    rlevel            function call recursion level
313    recursing         regex recursive call level    recursing         regex recursive call level
314    
315  Returns:            > 0 =>  Returns:            > 0 => number of match offset pairs placed in offsets
316                      = 0 =>                      = 0 => offsets overflowed; longest matches are present
317                       -1 => failed to match                       -1 => failed to match
318                     < -1 => some kind of unexpected problem                     < -1 => some kind of unexpected problem
319    
# Line 380  if (*first_op == OP_REVERSE) Line 479  if (*first_op == OP_REVERSE)
479      current_subject -= gone_back;      current_subject -= gone_back;
480      }      }
481    
482      /* Save the earliest consulted character */
483    
484      if (current_subject < md->start_used_ptr)
485        md->start_used_ptr = current_subject;
486    
487    /* Now we can process the individual branches. */    /* Now we can process the individual branches. */
488    
489    end_code = this_start_code;    end_code = this_start_code;
# Line 444  for (;;) Line 548  for (;;)
548    int i, j;    int i, j;
549    int clen, dlen;    int clen, dlen;
550    unsigned int c, d;    unsigned int c, d;
551      int forced_fail = 0;
552      BOOL could_continue = FALSE;
553    
554    /* Make the new state list into the active state list and empty the    /* Make the new state list into the active state list and empty the
555    new state list. */    new state list. */
# Line 457  for (;;) Line 563  for (;;)
563    workspace[0] ^= 1;              /* Remember for the restarting feature */    workspace[0] ^= 1;              /* Remember for the restarting feature */
564    workspace[1] = active_count;    workspace[1] = active_count;
565    
566  #ifdef DEBUG  #ifdef PCRE_DEBUG
567    printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP);    printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP);
568    pchars((uschar *)ptr, strlen((char *)ptr), stdout);    pchars((uschar *)ptr, strlen((char *)ptr), stdout);
569    printf("\"\n");    printf("\"\n");
# Line 501  for (;;) Line 607  for (;;)
607      stateblock *current_state = active_states + i;      stateblock *current_state = active_states + i;
608      const uschar *code;      const uschar *code;
609      int state_offset = current_state->offset;      int state_offset = current_state->offset;
610      int count, codevalue;      int count, codevalue, rrc;
 #ifdef SUPPORT_UCP  
     int chartype, script;  
 #endif  
611    
612  #ifdef DEBUG  #ifdef PCRE_DEBUG
613      printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset);      printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset);
614      if (clen == 0) printf("EOL\n");      if (clen == 0) printf("EOL\n");
615        else if (c > 32 && c < 127) printf("'%c'\n", c);        else if (c > 32 && c < 127) printf("'%c'\n", c);
# Line 536  for (;;) Line 639  for (;;)
639          }          }
640        }        }
641    
642      /* Check for a duplicate state with the same count, and skip if found. */      /* Check for a duplicate state with the same count, and skip if found.
643        See the note at the head of this module about the possibility of improving
644        performance here. */
645    
646      for (j = 0; j < i; j++)      for (j = 0; j < i; j++)
647        {        {
# Line 553  for (;;) Line 658  for (;;)
658      code = start_code + state_offset;      code = start_code + state_offset;
659      codevalue = *code;      codevalue = *code;
660    
661        /* If this opcode inspects a character, but we are at the end of the
662        subject, remember the fact for use when testing for a partial match. */
663    
664        if (clen == 0 && poptable[codevalue] != 0)
665          could_continue = TRUE;
666    
667      /* If this opcode is followed by an inline character, load it. It is      /* If this opcode is followed by an inline character, load it. It is
668      tempting to test for the presence of a subject character here, but that      tempting to test for the presence of a subject character here, but that
669      is wrong, because sometimes zero repetitions of the subject are      is wrong, because sometimes zero repetitions of the subject are
670      permitted.      permitted.
671    
672      We also use this mechanism for opcodes such as OP_TYPEPLUS that take an      We also use this mechanism for opcodes such as OP_TYPEPLUS that take an
673      argument that is not a data character - but is always one byte long.      argument that is not a data character - but is always one byte long. We
674      Unfortunately, we have to take special action to deal with  \P, \p, and      have to take special action to deal with  \P, \p, \H, \h, \V, \v and \X in
675      \X in this case. To keep the other cases fast, convert these ones to new      this case. To keep the other cases fast, convert these ones to new opcodes.
676      opcodes. */      */
677    
678      if (coptable[codevalue] > 0)      if (coptable[codevalue] > 0)
679        {        {
# Line 580  for (;;) Line 691  for (;;)
691            case OP_PROP: codevalue += OP_PROP_EXTRA; break;            case OP_PROP: codevalue += OP_PROP_EXTRA; break;
692            case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break;            case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break;
693            case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break;            case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break;
694              case OP_NOT_HSPACE:
695              case OP_HSPACE: codevalue += OP_HSPACE_EXTRA; break;
696              case OP_NOT_VSPACE:
697              case OP_VSPACE: codevalue += OP_VSPACE_EXTRA; break;
698            default: break;            default: break;
699            }            }
700          }          }
# Line 595  for (;;) Line 710  for (;;)
710    
711      switch (codevalue)      switch (codevalue)
712        {        {
713    /* ========================================================================== */
714          /* These cases are never obeyed. This is a fudge that causes a compile-
715          time error if the vectors coptable or poptable, which are indexed by
716          opcode, are not the correct length. It seems to be the only way to do
717          such a check at compile time, as the sizeof() operator does not work
718          in the C preprocessor. */
719    
720          case OP_TABLE_LENGTH:
721          case OP_TABLE_LENGTH +
722            ((sizeof(coptable) == OP_TABLE_LENGTH) &&
723             (sizeof(poptable) == OP_TABLE_LENGTH)):
724          break;
725    
726  /* ========================================================================== */  /* ========================================================================== */
727        /* Reached a closing bracket. If not at the end of the pattern, carry        /* Reached a closing bracket. If not at the end of the pattern, carry
728        on with the next opcode. Otherwise, unless we have an empty string and        on with the next opcode. Otherwise, unless we have an empty string and
729        PCRE_NOTEMPTY is set, save the match data, shifting up all previous        PCRE_NOTEMPTY is set, or PCRE_NOTEMPTY_ATSTART is set and we are at the
730          start of the subject, save the match data, shifting up all previous
731        matches so we always have the longest first. */        matches so we always have the longest first. */
732    
733        case OP_KET:        case OP_KET:
# Line 613  for (;;) Line 741  for (;;)
741            ADD_ACTIVE(state_offset - GET(code, 1), 0);            ADD_ACTIVE(state_offset - GET(code, 1), 0);
742            }            }
743          }          }
744        else if (ptr > current_subject || (md->moptions & PCRE_NOTEMPTY) == 0)        else
745          {          {
746          if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0;          if (ptr > current_subject ||
747            else if (match_count > 0 && ++match_count * 2 >= offsetcount)              ((md->moptions & PCRE_NOTEMPTY) == 0 &&
748              match_count = 0;                ((md->moptions & PCRE_NOTEMPTY_ATSTART) == 0 ||
749          count = ((match_count == 0)? offsetcount : match_count * 2) - 2;                  current_subject > start_subject + md->start_offset)))
750          if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int));            {
751          if (offsetcount >= 2)            if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0;
752            {              else if (match_count > 0 && ++match_count * 2 >= offsetcount)
753            offsets[0] = current_subject - start_subject;                match_count = 0;
754            offsets[1] = ptr - start_subject;            count = ((match_count == 0)? offsetcount : match_count * 2) - 2;
755            DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP,            if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int));
756              offsets[1] - offsets[0], current_subject));            if (offsetcount >= 2)
757            }              {
758          if ((md->moptions & PCRE_DFA_SHORTEST) != 0)              offsets[0] = current_subject - start_subject;
759            {              offsets[1] = ptr - start_subject;
760            DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n"              DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP,
761              "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel,                offsets[1] - offsets[0], current_subject));
762              match_count, rlevel*2-2, SP));              }
763            return match_count;            if ((md->moptions & PCRE_DFA_SHORTEST) != 0)
764                {
765                DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n"
766                  "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel,
767                  match_count, rlevel*2-2, SP));
768                return match_count;
769                }
770            }            }
771          }          }
772        break;        break;
# Line 680  for (;;) Line 814  for (;;)
814        break;        break;
815    
816        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
817          case OP_SKIPZERO:
818          code += 1 + GET(code, 2);
819          while (*code == OP_ALT) code += GET(code, 1);
820          ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0);
821          break;
822    
823          /*-----------------------------------------------------------------*/
824        case OP_CIRC:        case OP_CIRC:
825        if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) ||        if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) ||
826            ((ims & PCRE_MULTILINE) != 0 &&            ((ims & PCRE_MULTILINE) != 0 &&
# Line 718  for (;;) Line 859  for (;;)
859    
860        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
861        case OP_ANY:        case OP_ANY:
862        if (clen > 0 && ((ims & PCRE_DOTALL) != 0 || !IS_NEWLINE(ptr)))        if (clen > 0 && !IS_NEWLINE(ptr))
863            { ADD_NEW(state_offset + 1, 0); }
864          break;
865    
866          /*-----------------------------------------------------------------*/
867          case OP_ALLANY:
868          if (clen > 0)
869          { ADD_NEW(state_offset + 1, 0); }          { ADD_NEW(state_offset + 1, 0); }
870        break;        break;
871    
# Line 733  for (;;) Line 880  for (;;)
880        if ((md->moptions & PCRE_NOTEOL) == 0)        if ((md->moptions & PCRE_NOTEOL) == 0)
881          {          {
882          if (clen == 0 ||          if (clen == 0 ||
883              (IS_NEWLINE(ptr) &&              ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) &&
884                 ((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen)                 ((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen)
885              ))              ))
886            { ADD_ACTIVE(state_offset + 1, 0); }            { ADD_ACTIVE(state_offset + 1, 0); }
# Line 770  for (;;) Line 917  for (;;)
917          if (ptr > start_subject)          if (ptr > start_subject)
918            {            {
919            const uschar *temp = ptr - 1;            const uschar *temp = ptr - 1;
920              if (temp < md->start_used_ptr) md->start_used_ptr = temp;
921  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
922            if (utf8) BACKCHAR(temp);            if (utf8) BACKCHAR(temp);
923  #endif  #endif
924            GETCHARTEST(d, temp);            GETCHARTEST(d, temp);
925    #ifdef SUPPORT_UCP
926              if ((md->poptions & PCRE_UCP) != 0)
927                {
928                if (d == '_') left_word = TRUE; else
929                  {
930                  int cat = UCD_CATEGORY(d);
931                  left_word = (cat == ucp_L || cat == ucp_N);
932                  }
933                }
934              else
935    #endif
936            left_word = d < 256 && (ctypes[d] & ctype_word) != 0;            left_word = d < 256 && (ctypes[d] & ctype_word) != 0;
937            }            }
938          else left_word = 0;          else left_word = FALSE;
939    
940          if (clen > 0) right_word = c < 256 && (ctypes[c] & ctype_word) != 0;          if (clen > 0)
941            else right_word = 0;            {
942    #ifdef SUPPORT_UCP
943              if ((md->poptions & PCRE_UCP) != 0)
944                {
945                if (c == '_') right_word = TRUE; else
946                  {
947                  int cat = UCD_CATEGORY(c);
948                  right_word = (cat == ucp_L || cat == ucp_N);
949                  }
950                }
951              else
952    #endif
953              right_word = c < 256 && (ctypes[c] & ctype_word) != 0;
954              }
955            else right_word = FALSE;
956    
957          if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY))          if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY))
958            { ADD_ACTIVE(state_offset + 1, 0); }            { ADD_ACTIVE(state_offset + 1, 0); }
# Line 798  for (;;) Line 971  for (;;)
971        if (clen > 0)        if (clen > 0)
972          {          {
973          BOOL OK;          BOOL OK;
974          int category = _pcre_ucp_findprop(c, &chartype, &script);          const ucd_record * prop = GET_UCD(c);
975          switch(code[1])          switch(code[1])
976            {            {
977            case PT_ANY:            case PT_ANY:
# Line 806  for (;;) Line 979  for (;;)
979            break;            break;
980    
981            case PT_LAMP:            case PT_LAMP:
982            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
983                   prop->chartype == ucp_Lt;
984            break;            break;
985    
986            case PT_GC:            case PT_GC:
987            OK = category == code[2];            OK = _pcre_ucp_gentype[prop->chartype] == code[2];
988            break;            break;
989    
990            case PT_PC:            case PT_PC:
991            OK = chartype == code[2];            OK = prop->chartype == code[2];
992            break;            break;
993    
994            case PT_SC:            case PT_SC:
995            OK = script == code[2];            OK = prop->script == code[2];
996            break;            break;
997    
998              /* These are specials for combination cases. */
999    
1000              case PT_ALNUM:
1001              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1002                   _pcre_ucp_gentype[prop->chartype] == ucp_N;
1003              break;
1004    
1005              case PT_SPACE:    /* Perl space */
1006              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1007                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1008              break;
1009    
1010              case PT_PXSPACE:  /* POSIX space */
1011              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1012                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1013                   c == CHAR_FF || c == CHAR_CR;
1014              break;
1015    
1016              case PT_WORD:
1017              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1018                   _pcre_ucp_gentype[prop->chartype] == ucp_N ||
1019                   c == CHAR_UNDERSCORE;
1020              break;
1021    
1022            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1023    
# Line 838  for (;;) Line 1036  for (;;)
1036  /* ========================================================================== */  /* ========================================================================== */
1037        /* These opcodes likewise inspect the subject character, but have an        /* These opcodes likewise inspect the subject character, but have an
1038        argument that is not a data character. It is one of these opcodes:        argument that is not a data character. It is one of these opcodes:
1039        OP_ANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE, OP_WORDCHAR,        OP_ANY, OP_ALLANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE,
1040        OP_NOT_WORDCHAR. The value is loaded into d. */        OP_WORDCHAR, OP_NOT_WORDCHAR. The value is loaded into d. */
1041    
1042        case OP_TYPEPLUS:        case OP_TYPEPLUS:
1043        case OP_TYPEMINPLUS:        case OP_TYPEMINPLUS:
# Line 850  for (;;) Line 1048  for (;;)
1048          {          {
1049          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1050              (c < 256 &&              (c < 256 &&
1051                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1052                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1053            {            {
1054            if (count > 0 && codevalue == OP_TYPEPOSPLUS)            if (count > 0 && codevalue == OP_TYPEPOSPLUS)
# Line 876  for (;;) Line 1071  for (;;)
1071          {          {
1072          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1073              (c < 256 &&              (c < 256 &&
1074                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1075                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1076            {            {
1077            if (codevalue == OP_TYPEPOSQUERY)            if (codevalue == OP_TYPEPOSQUERY)
# Line 901  for (;;) Line 1093  for (;;)
1093          {          {
1094          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1095              (c < 256 &&              (c < 256 &&
1096                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1097                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1098            {            {
1099            if (codevalue == OP_TYPEPOSSTAR)            if (codevalue == OP_TYPEPOSSTAR)
# Line 924  for (;;) Line 1113  for (;;)
1113          {          {
1114          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1115              (c < 256 &&              (c < 256 &&
1116                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1117                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1118            {            {
1119            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
# Line 948  for (;;) Line 1134  for (;;)
1134          {          {
1135          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1136              (c < 256 &&              (c < 256 &&
1137                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1138                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1139            {            {
1140            if (codevalue == OP_TYPEPOSUPTO)            if (codevalue == OP_TYPEPOSUPTO)
# Line 982  for (;;) Line 1165  for (;;)
1165        if (clen > 0)        if (clen > 0)
1166          {          {
1167          BOOL OK;          BOOL OK;
1168          int category = _pcre_ucp_findprop(c, &chartype, &script);          const ucd_record * prop = GET_UCD(c);
1169          switch(code[2])          switch(code[2])
1170            {            {
1171            case PT_ANY:            case PT_ANY:
# Line 990  for (;;) Line 1173  for (;;)
1173            break;            break;
1174    
1175            case PT_LAMP:            case PT_LAMP:
1176            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1177                prop->chartype == ucp_Lt;
1178            break;            break;
1179    
1180            case PT_GC:            case PT_GC:
1181            OK = category == code[3];            OK = _pcre_ucp_gentype[prop->chartype] == code[3];
1182            break;            break;
1183    
1184            case PT_PC:            case PT_PC:
1185            OK = chartype == code[3];            OK = prop->chartype == code[3];
1186            break;            break;
1187    
1188            case PT_SC:            case PT_SC:
1189            OK = script == code[3];            OK = prop->script == code[3];
1190            break;            break;
1191    
1192              /* These are specials for combination cases. */
1193    
1194              case PT_ALNUM:
1195              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1196                   _pcre_ucp_gentype[prop->chartype] == ucp_N;
1197              break;
1198    
1199              case PT_SPACE:    /* Perl space */
1200              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1201                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1202              break;
1203    
1204              case PT_PXSPACE:  /* POSIX space */
1205              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1206                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1207                   c == CHAR_FF || c == CHAR_CR;
1208              break;
1209    
1210              case PT_WORD:
1211              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1212                   _pcre_ucp_gentype[prop->chartype] == ucp_N ||
1213                   c == CHAR_UNDERSCORE;
1214              break;
1215    
1216            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1217    
1218            default:            default:
# Line 1031  for (;;) Line 1239  for (;;)
1239        case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS:        case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS:
1240        count = current_state->count;  /* Already matched */        count = current_state->count;  /* Already matched */
1241        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1242        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)
1243          {          {
1244          const uschar *nptr = ptr + clen;          const uschar *nptr = ptr + clen;
1245          int ncount = 0;          int ncount = 0;
# Line 1045  for (;;) Line 1253  for (;;)
1253            int nd;            int nd;
1254            int ndlen = 1;            int ndlen = 1;
1255            GETCHARLEN(nd, nptr, ndlen);            GETCHARLEN(nd, nptr, ndlen);
1256            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if (UCD_CATEGORY(nd) != ucp_M) break;
1257            ncount++;            ncount++;
1258            nptr += ndlen;            nptr += ndlen;
1259            }            }
# Line 1066  for (;;) Line 1274  for (;;)
1274          int ncount = 0;          int ncount = 0;
1275          switch (c)          switch (c)
1276            {            {
1277              case 0x000b:
1278              case 0x000c:
1279              case 0x0085:
1280              case 0x2028:
1281              case 0x2029:
1282              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1283              goto ANYNL01;
1284    
1285            case 0x000d:            case 0x000d:
1286            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;
1287            /* Fall through */            /* Fall through */
1288    
1289              ANYNL01:
1290              case 0x000a:
1291              if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS)
1292                {
1293                active_count--;           /* Remove non-match possibility */
1294                next_active_state--;
1295                }
1296              count++;
1297              ADD_NEW_DATA(-state_offset, count, ncount);
1298              break;
1299    
1300              default:
1301              break;
1302              }
1303            }
1304          break;
1305    
1306          /*-----------------------------------------------------------------*/
1307          case OP_VSPACE_EXTRA + OP_TYPEPLUS:
1308          case OP_VSPACE_EXTRA + OP_TYPEMINPLUS:
1309          case OP_VSPACE_EXTRA + OP_TYPEPOSPLUS:
1310          count = current_state->count;  /* Already matched */
1311          if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1312          if (clen > 0)
1313            {
1314            BOOL OK;
1315            switch (c)
1316              {
1317            case 0x000a:            case 0x000a:
1318            case 0x000b:            case 0x000b:
1319            case 0x000c:            case 0x000c:
1320              case 0x000d:
1321            case 0x0085:            case 0x0085:
1322            case 0x2028:            case 0x2028:
1323            case 0x2029:            case 0x2029:
1324            if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS)            OK = TRUE;
1325              break;
1326    
1327              default:
1328              OK = FALSE;
1329              break;
1330              }
1331    
1332            if (OK == (d == OP_VSPACE))
1333              {
1334              if (count > 0 && codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSPLUS)
1335              {              {
1336              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
1337              next_active_state--;              next_active_state--;
1338              }              }
1339            count++;            count++;
1340            ADD_NEW_DATA(-state_offset, count, ncount);            ADD_NEW_DATA(-state_offset, count, 0);
1341              }
1342            }
1343          break;
1344    
1345          /*-----------------------------------------------------------------*/
1346          case OP_HSPACE_EXTRA + OP_TYPEPLUS:
1347          case OP_HSPACE_EXTRA + OP_TYPEMINPLUS:
1348          case OP_HSPACE_EXTRA + OP_TYPEPOSPLUS:
1349          count = current_state->count;  /* Already matched */
1350          if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1351          if (clen > 0)
1352            {
1353            BOOL OK;
1354            switch (c)
1355              {
1356              case 0x09:      /* HT */
1357              case 0x20:      /* SPACE */
1358              case 0xa0:      /* NBSP */
1359              case 0x1680:    /* OGHAM SPACE MARK */
1360              case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1361              case 0x2000:    /* EN QUAD */
1362              case 0x2001:    /* EM QUAD */
1363              case 0x2002:    /* EN SPACE */
1364              case 0x2003:    /* EM SPACE */
1365              case 0x2004:    /* THREE-PER-EM SPACE */
1366              case 0x2005:    /* FOUR-PER-EM SPACE */
1367              case 0x2006:    /* SIX-PER-EM SPACE */
1368              case 0x2007:    /* FIGURE SPACE */
1369              case 0x2008:    /* PUNCTUATION SPACE */
1370              case 0x2009:    /* THIN SPACE */
1371              case 0x200A:    /* HAIR SPACE */
1372              case 0x202f:    /* NARROW NO-BREAK SPACE */
1373              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1374              case 0x3000:    /* IDEOGRAPHIC SPACE */
1375              OK = TRUE;
1376            break;            break;
1377    
1378            default:            default:
1379              OK = FALSE;
1380            break;            break;
1381            }            }
1382    
1383            if (OK == (d == OP_HSPACE))
1384              {
1385              if (count > 0 && codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSPLUS)
1386                {
1387                active_count--;           /* Remove non-match possibility */
1388                next_active_state--;
1389                }
1390              count++;
1391              ADD_NEW_DATA(-state_offset, count, 0);
1392              }
1393          }          }
1394        break;        break;
1395    
# Line 1108  for (;;) Line 1412  for (;;)
1412        if (clen > 0)        if (clen > 0)
1413          {          {
1414          BOOL OK;          BOOL OK;
1415          int category = _pcre_ucp_findprop(c, &chartype, &script);          const ucd_record * prop = GET_UCD(c);
1416          switch(code[2])          switch(code[2])
1417            {            {
1418            case PT_ANY:            case PT_ANY:
# Line 1116  for (;;) Line 1420  for (;;)
1420            break;            break;
1421    
1422            case PT_LAMP:            case PT_LAMP:
1423            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1424                prop->chartype == ucp_Lt;
1425            break;            break;
1426    
1427            case PT_GC:            case PT_GC:
1428            OK = category == code[3];            OK = _pcre_ucp_gentype[prop->chartype] == code[3];
1429            break;            break;
1430    
1431            case PT_PC:            case PT_PC:
1432            OK = chartype == code[3];            OK = prop->chartype == code[3];
1433            break;            break;
1434    
1435            case PT_SC:            case PT_SC:
1436            OK = script == code[3];            OK = prop->script == code[3];
1437            break;            break;
1438    
1439              /* These are specials for combination cases. */
1440    
1441              case PT_ALNUM:
1442              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1443                   _pcre_ucp_gentype[prop->chartype] == ucp_N;
1444              break;
1445    
1446              case PT_SPACE:    /* Perl space */
1447              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1448                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1449              break;
1450    
1451              case PT_PXSPACE:  /* POSIX space */
1452              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1453                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1454                   c == CHAR_FF || c == CHAR_CR;
1455              break;
1456    
1457              case PT_WORD:
1458              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1459                   _pcre_ucp_gentype[prop->chartype] == ucp_N ||
1460                   c == CHAR_UNDERSCORE;
1461              break;
1462    
1463            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1464    
# Line 1166  for (;;) Line 1495  for (;;)
1495        QS2:        QS2:
1496    
1497        ADD_ACTIVE(state_offset + 2, 0);        ADD_ACTIVE(state_offset + 2, 0);
1498        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)
1499          {          {
1500          const uschar *nptr = ptr + clen;          const uschar *nptr = ptr + clen;
1501          int ncount = 0;          int ncount = 0;
# Line 1181  for (;;) Line 1510  for (;;)
1510            int nd;            int nd;
1511            int ndlen = 1;            int ndlen = 1;
1512            GETCHARLEN(nd, nptr, ndlen);            GETCHARLEN(nd, nptr, ndlen);
1513            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if (UCD_CATEGORY(nd) != ucp_M) break;
1514            ncount++;            ncount++;
1515            nptr += ndlen;            nptr += ndlen;
1516            }            }
# Line 1209  for (;;) Line 1538  for (;;)
1538          int ncount = 0;          int ncount = 0;
1539          switch (c)          switch (c)
1540            {            {
1541              case 0x000b:
1542              case 0x000c:
1543              case 0x0085:
1544              case 0x2028:
1545              case 0x2029:
1546              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1547              goto ANYNL02;
1548    
1549            case 0x000d:            case 0x000d:
1550            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;
1551            /* Fall through */            /* Fall through */
1552    
1553              ANYNL02:
1554              case 0x000a:
1555              if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR ||
1556                  codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY)
1557                {
1558                active_count--;           /* Remove non-match possibility */
1559                next_active_state--;
1560                }
1561              ADD_NEW_DATA(-(state_offset + count), 0, ncount);
1562              break;
1563    
1564              default:
1565              break;
1566              }
1567            }
1568          break;
1569    
1570          /*-----------------------------------------------------------------*/
1571          case OP_VSPACE_EXTRA + OP_TYPEQUERY:
1572          case OP_VSPACE_EXTRA + OP_TYPEMINQUERY:
1573          case OP_VSPACE_EXTRA + OP_TYPEPOSQUERY:
1574          count = 2;
1575          goto QS4;
1576    
1577          case OP_VSPACE_EXTRA + OP_TYPESTAR:
1578          case OP_VSPACE_EXTRA + OP_TYPEMINSTAR:
1579          case OP_VSPACE_EXTRA + OP_TYPEPOSSTAR:
1580          count = 0;
1581    
1582          QS4:
1583          ADD_ACTIVE(state_offset + 2, 0);
1584          if (clen > 0)
1585            {
1586            BOOL OK;
1587            switch (c)
1588              {
1589            case 0x000a:            case 0x000a:
1590            case 0x000b:            case 0x000b:
1591            case 0x000c:            case 0x000c:
1592              case 0x000d:
1593            case 0x0085:            case 0x0085:
1594            case 0x2028:            case 0x2028:
1595            case 0x2029:            case 0x2029:
1596            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR ||            OK = TRUE;
1597                codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY)            break;
1598    
1599              default:
1600              OK = FALSE;
1601              break;
1602              }
1603            if (OK == (d == OP_VSPACE))
1604              {
1605              if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSSTAR ||
1606                  codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSQUERY)
1607              {              {
1608              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
1609              next_active_state--;              next_active_state--;
1610              }              }
1611            ADD_NEW_DATA(-(state_offset + count), 0, ncount);            ADD_NEW_DATA(-(state_offset + count), 0, 0);
1612              }
1613            }
1614          break;
1615    
1616          /*-----------------------------------------------------------------*/
1617          case OP_HSPACE_EXTRA + OP_TYPEQUERY:
1618          case OP_HSPACE_EXTRA + OP_TYPEMINQUERY:
1619          case OP_HSPACE_EXTRA + OP_TYPEPOSQUERY:
1620          count = 2;
1621          goto QS5;
1622    
1623          case OP_HSPACE_EXTRA + OP_TYPESTAR:
1624          case OP_HSPACE_EXTRA + OP_TYPEMINSTAR:
1625          case OP_HSPACE_EXTRA + OP_TYPEPOSSTAR:
1626          count = 0;
1627    
1628          QS5:
1629          ADD_ACTIVE(state_offset + 2, 0);
1630          if (clen > 0)
1631            {
1632            BOOL OK;
1633            switch (c)
1634              {
1635              case 0x09:      /* HT */
1636              case 0x20:      /* SPACE */
1637              case 0xa0:      /* NBSP */
1638              case 0x1680:    /* OGHAM SPACE MARK */
1639              case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1640              case 0x2000:    /* EN QUAD */
1641              case 0x2001:    /* EM QUAD */
1642              case 0x2002:    /* EN SPACE */
1643              case 0x2003:    /* EM SPACE */
1644              case 0x2004:    /* THREE-PER-EM SPACE */
1645              case 0x2005:    /* FOUR-PER-EM SPACE */
1646              case 0x2006:    /* SIX-PER-EM SPACE */
1647              case 0x2007:    /* FIGURE SPACE */
1648              case 0x2008:    /* PUNCTUATION SPACE */
1649              case 0x2009:    /* THIN SPACE */
1650              case 0x200A:    /* HAIR SPACE */
1651              case 0x202f:    /* NARROW NO-BREAK SPACE */
1652              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1653              case 0x3000:    /* IDEOGRAPHIC SPACE */
1654              OK = TRUE;
1655            break;            break;
1656    
1657            default:            default:
1658              OK = FALSE;
1659            break;            break;
1660            }            }
1661    
1662            if (OK == (d == OP_HSPACE))
1663              {
1664              if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSSTAR ||
1665                  codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSQUERY)
1666                {
1667                active_count--;           /* Remove non-match possibility */
1668                next_active_state--;
1669                }
1670              ADD_NEW_DATA(-(state_offset + count), 0, 0);
1671              }
1672          }          }
1673        break;        break;
1674    
# Line 1244  for (;;) Line 1684  for (;;)
1684        if (clen > 0)        if (clen > 0)
1685          {          {
1686          BOOL OK;          BOOL OK;
1687          int category = _pcre_ucp_findprop(c, &chartype, &script);          const ucd_record * prop = GET_UCD(c);
1688          switch(code[4])          switch(code[4])
1689            {            {
1690            case PT_ANY:            case PT_ANY:
# Line 1252  for (;;) Line 1692  for (;;)
1692            break;            break;
1693    
1694            case PT_LAMP:            case PT_LAMP:
1695            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1696                prop->chartype == ucp_Lt;
1697            break;            break;
1698    
1699            case PT_GC:            case PT_GC:
1700            OK = category == code[5];            OK = _pcre_ucp_gentype[prop->chartype] == code[5];
1701            break;            break;
1702    
1703            case PT_PC:            case PT_PC:
1704            OK = chartype == code[5];            OK = prop->chartype == code[5];
1705            break;            break;
1706    
1707            case PT_SC:            case PT_SC:
1708            OK = script == code[5];            OK = prop->script == code[5];
1709            break;            break;
1710    
1711              /* These are specials for combination cases. */
1712    
1713              case PT_ALNUM:
1714              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1715                   _pcre_ucp_gentype[prop->chartype] == ucp_N;
1716              break;
1717    
1718              case PT_SPACE:    /* Perl space */
1719              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1720                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1721              break;
1722    
1723              case PT_PXSPACE:  /* POSIX space */
1724              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1725                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1726                   c == CHAR_FF || c == CHAR_CR;
1727              break;
1728    
1729              case PT_WORD:
1730              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1731                   _pcre_ucp_gentype[prop->chartype] == ucp_N ||
1732                   c == CHAR_UNDERSCORE;
1733              break;
1734    
1735            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1736    
# Line 1297  for (;;) Line 1762  for (;;)
1762        if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT)        if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT)
1763          { ADD_ACTIVE(state_offset + 4, 0); }          { ADD_ACTIVE(state_offset + 4, 0); }
1764        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1765        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)
1766          {          {
1767          const uschar *nptr = ptr + clen;          const uschar *nptr = ptr + clen;
1768          int ncount = 0;          int ncount = 0;
# Line 1311  for (;;) Line 1776  for (;;)
1776            int nd;            int nd;
1777            int ndlen = 1;            int ndlen = 1;
1778            GETCHARLEN(nd, nptr, ndlen);            GETCHARLEN(nd, nptr, ndlen);
1779            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if (UCD_CATEGORY(nd) != ucp_M) break;
1780            ncount++;            ncount++;
1781            nptr += ndlen;            nptr += ndlen;
1782            }            }
# Line 1336  for (;;) Line 1801  for (;;)
1801          int ncount = 0;          int ncount = 0;
1802          switch (c)          switch (c)
1803            {            {
1804              case 0x000b:
1805              case 0x000c:
1806              case 0x0085:
1807              case 0x2028:
1808              case 0x2029:
1809              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1810              goto ANYNL03;
1811    
1812            case 0x000d:            case 0x000d:
1813            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;
1814            /* Fall through */            /* Fall through */
1815    
1816              ANYNL03:
1817              case 0x000a:
1818              if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO)
1819                {
1820                active_count--;           /* Remove non-match possibility */
1821                next_active_state--;
1822                }
1823              if (++count >= GET2(code, 1))
1824                { ADD_NEW_DATA(-(state_offset + 4), 0, ncount); }
1825              else
1826                { ADD_NEW_DATA(-state_offset, count, ncount); }
1827              break;
1828    
1829              default:
1830              break;
1831              }
1832            }
1833          break;
1834    
1835          /*-----------------------------------------------------------------*/
1836          case OP_VSPACE_EXTRA + OP_TYPEEXACT:
1837          case OP_VSPACE_EXTRA + OP_TYPEUPTO:
1838          case OP_VSPACE_EXTRA + OP_TYPEMINUPTO:
1839          case OP_VSPACE_EXTRA + OP_TYPEPOSUPTO:
1840          if (codevalue != OP_VSPACE_EXTRA + OP_TYPEEXACT)
1841            { ADD_ACTIVE(state_offset + 4, 0); }
1842          count = current_state->count;  /* Number already matched */
1843          if (clen > 0)
1844            {
1845            BOOL OK;
1846            switch (c)
1847              {
1848            case 0x000a:            case 0x000a:
1849            case 0x000b:            case 0x000b:
1850            case 0x000c:            case 0x000c:
1851              case 0x000d:
1852            case 0x0085:            case 0x0085:
1853            case 0x2028:            case 0x2028:
1854            case 0x2029:            case 0x2029:
1855            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO)            OK = TRUE;
1856              break;
1857    
1858              default:
1859              OK = FALSE;
1860              }
1861    
1862            if (OK == (d == OP_VSPACE))
1863              {
1864              if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSUPTO)
1865              {              {
1866              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
1867              next_active_state--;              next_active_state--;
1868              }              }
1869            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
1870              { ADD_NEW_DATA(-(state_offset + 4), 0, ncount); }              { ADD_NEW_DATA(-(state_offset + 4), 0, 0); }
1871            else            else
1872              { ADD_NEW_DATA(-state_offset, count, ncount); }              { ADD_NEW_DATA(-state_offset, count, 0); }
1873              }
1874            }
1875          break;
1876    
1877          /*-----------------------------------------------------------------*/
1878          case OP_HSPACE_EXTRA + OP_TYPEEXACT:
1879          case OP_HSPACE_EXTRA + OP_TYPEUPTO:
1880          case OP_HSPACE_EXTRA + OP_TYPEMINUPTO:
1881          case OP_HSPACE_EXTRA + OP_TYPEPOSUPTO:
1882          if (codevalue != OP_HSPACE_EXTRA + OP_TYPEEXACT)
1883            { ADD_ACTIVE(state_offset + 4, 0); }
1884          count = current_state->count;  /* Number already matched */
1885          if (clen > 0)
1886            {
1887            BOOL OK;
1888            switch (c)
1889              {
1890              case 0x09:      /* HT */
1891              case 0x20:      /* SPACE */
1892              case 0xa0:      /* NBSP */
1893              case 0x1680:    /* OGHAM SPACE MARK */
1894              case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1895              case 0x2000:    /* EN QUAD */
1896              case 0x2001:    /* EM QUAD */
1897              case 0x2002:    /* EN SPACE */
1898              case 0x2003:    /* EM SPACE */
1899              case 0x2004:    /* THREE-PER-EM SPACE */
1900              case 0x2005:    /* FOUR-PER-EM SPACE */
1901              case 0x2006:    /* SIX-PER-EM SPACE */
1902              case 0x2007:    /* FIGURE SPACE */
1903              case 0x2008:    /* PUNCTUATION SPACE */
1904              case 0x2009:    /* THIN SPACE */
1905              case 0x200A:    /* HAIR SPACE */
1906              case 0x202f:    /* NARROW NO-BREAK SPACE */
1907              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1908              case 0x3000:    /* IDEOGRAPHIC SPACE */
1909              OK = TRUE;
1910            break;            break;
1911    
1912            default:            default:
1913              OK = FALSE;
1914            break;            break;
1915            }            }
1916    
1917            if (OK == (d == OP_HSPACE))
1918              {
1919              if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSUPTO)
1920                {
1921                active_count--;           /* Remove non-match possibility */
1922                next_active_state--;
1923                }
1924              if (++count >= GET2(code, 1))
1925                { ADD_NEW_DATA(-(state_offset + 4), 0, 0); }
1926              else
1927                { ADD_NEW_DATA(-state_offset, count, 0); }
1928              }
1929          }          }
1930        break;        break;
1931    
# Line 1388  for (;;) Line 1956  for (;;)
1956            other case of the character. */            other case of the character. */
1957    
1958  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1959            othercase = _pcre_ucp_othercase(c);            othercase = UCD_OTHERCASE(c);
1960  #else  #else
1961            othercase = NOTACHAR;            othercase = NOTACHAR;
1962  #endif  #endif
# Line 1413  for (;;) Line 1981  for (;;)
1981        to wait for them to pass before continuing. */        to wait for them to pass before continuing. */
1982    
1983        case OP_EXTUNI:        case OP_EXTUNI:
1984        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)
1985          {          {
1986          const uschar *nptr = ptr + clen;          const uschar *nptr = ptr + clen;
1987          int ncount = 0;          int ncount = 0;
# Line 1421  for (;;) Line 1989  for (;;)
1989            {            {
1990            int nclen = 1;            int nclen = 1;
1991            GETCHARLEN(c, nptr, nclen);            GETCHARLEN(c, nptr, nclen);
1992            if (_pcre_ucp_findprop(c, &chartype, &script) != ucp_M) break;            if (UCD_CATEGORY(c) != ucp_M) break;
1993            ncount++;            ncount++;
1994            nptr += nclen;            nptr += nclen;
1995            }            }
# Line 1438  for (;;) Line 2006  for (;;)
2006        case OP_ANYNL:        case OP_ANYNL:
2007        if (clen > 0) switch(c)        if (clen > 0) switch(c)
2008          {          {
         case 0x000a:  
2009          case 0x000b:          case 0x000b:
2010          case 0x000c:          case 0x000c:
2011          case 0x0085:          case 0x0085:
2012          case 0x2028:          case 0x2028:
2013          case 0x2029:          case 0x2029:
2014            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
2015    
2016            case 0x000a:
2017          ADD_NEW(state_offset + 1, 0);          ADD_NEW(state_offset + 1, 0);
2018          break;          break;
2019    
2020          case 0x000d:          case 0x000d:
2021          if (ptr + 1 < end_subject && ptr[1] == 0x0a)          if (ptr + 1 < end_subject && ptr[1] == 0x0a)
2022            {            {
# Line 1460  for (;;) Line 2031  for (;;)
2031        break;        break;
2032    
2033        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2034          case OP_NOT_VSPACE:
2035          if (clen > 0) switch(c)
2036            {
2037            case 0x000a:
2038            case 0x000b:
2039            case 0x000c:
2040            case 0x000d:
2041            case 0x0085:
2042            case 0x2028:
2043            case 0x2029:
2044            break;
2045    
2046            default:
2047            ADD_NEW(state_offset + 1, 0);
2048            break;
2049            }
2050          break;
2051    
2052          /*-----------------------------------------------------------------*/
2053          case OP_VSPACE:
2054          if (clen > 0) switch(c)
2055            {
2056            case 0x000a:
2057            case 0x000b:
2058            case 0x000c:
2059            case 0x000d:
2060            case 0x0085:
2061            case 0x2028:
2062            case 0x2029:
2063            ADD_NEW(state_offset + 1, 0);
2064            break;
2065    
2066            default: break;
2067            }
2068          break;
2069    
2070          /*-----------------------------------------------------------------*/
2071          case OP_NOT_HSPACE:
2072          if (clen > 0) switch(c)
2073            {
2074            case 0x09:      /* HT */
2075            case 0x20:      /* SPACE */
2076            case 0xa0:      /* NBSP */
2077            case 0x1680:    /* OGHAM SPACE MARK */
2078            case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
2079            case 0x2000:    /* EN QUAD */
2080            case 0x2001:    /* EM QUAD */
2081            case 0x2002:    /* EN SPACE */
2082            case 0x2003:    /* EM SPACE */
2083            case 0x2004:    /* THREE-PER-EM SPACE */
2084            case 0x2005:    /* FOUR-PER-EM SPACE */
2085            case 0x2006:    /* SIX-PER-EM SPACE */
2086            case 0x2007:    /* FIGURE SPACE */
2087            case 0x2008:    /* PUNCTUATION SPACE */
2088            case 0x2009:    /* THIN SPACE */
2089            case 0x200A:    /* HAIR SPACE */
2090            case 0x202f:    /* NARROW NO-BREAK SPACE */
2091            case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2092            case 0x3000:    /* IDEOGRAPHIC SPACE */
2093            break;
2094    
2095            default:
2096            ADD_NEW(state_offset + 1, 0);
2097            break;
2098            }
2099          break;
2100    
2101          /*-----------------------------------------------------------------*/
2102          case OP_HSPACE:
2103          if (clen > 0) switch(c)
2104            {
2105            case 0x09:      /* HT */
2106            case 0x20:      /* SPACE */
2107            case 0xa0:      /* NBSP */
2108            case 0x1680:    /* OGHAM SPACE MARK */
2109            case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
2110            case 0x2000:    /* EN QUAD */
2111            case 0x2001:    /* EM QUAD */
2112            case 0x2002:    /* EN SPACE */
2113            case 0x2003:    /* EM SPACE */
2114            case 0x2004:    /* THREE-PER-EM SPACE */
2115            case 0x2005:    /* FOUR-PER-EM SPACE */
2116            case 0x2006:    /* SIX-PER-EM SPACE */
2117            case 0x2007:    /* FIGURE SPACE */
2118            case 0x2008:    /* PUNCTUATION SPACE */
2119            case 0x2009:    /* THIN SPACE */
2120            case 0x200A:    /* HAIR SPACE */
2121            case 0x202f:    /* NARROW NO-BREAK SPACE */
2122            case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2123            case 0x3000:    /* IDEOGRAPHIC SPACE */
2124            ADD_NEW(state_offset + 1, 0);
2125            break;
2126            }
2127          break;
2128    
2129          /*-----------------------------------------------------------------*/
2130        /* Match a negated single character. This is only used for one-byte        /* Match a negated single character. This is only used for one-byte
2131        characters, that is, we know that d < 256. The character we are        characters, that is, we know that d < 256. The character we are
2132        checking (c) can be multibyte. */        checking (c) can be multibyte. */
# Line 1490  for (;;) Line 2157  for (;;)
2157            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2158              {              {
2159  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2160              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2161  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2162              }              }
2163            else            else
# Line 1528  for (;;) Line 2195  for (;;)
2195            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2196              {              {
2197  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2198              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2199  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2200              }              }
2201            else            else
# Line 1564  for (;;) Line 2231  for (;;)
2231            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2232              {              {
2233  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2234              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2235  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2236              }              }
2237            else            else
# Line 1596  for (;;) Line 2263  for (;;)
2263            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2264              {              {
2265  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2266              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2267  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2268              }              }
2269            else            else
# Line 1631  for (;;) Line 2298  for (;;)
2298            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2299              {              {
2300  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2301              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2302  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2303              }              }
2304            else            else
# Line 1739  for (;;) Line 2406  for (;;)
2406    
2407  /* ========================================================================== */  /* ========================================================================== */
2408        /* These are the opcodes for fancy brackets of various kinds. We have        /* These are the opcodes for fancy brackets of various kinds. We have
2409        to use recursion in order to handle them. */        to use recursion in order to handle them. The "always failing" assertion
2410          (?!) is optimised to OP_FAIL when compiling, so we have to support that,
2411          though the other "backtracking verbs" are not supported. */
2412    
2413          case OP_FAIL:
2414          forced_fail++;    /* Count FAILs for multiple states */
2415          break;
2416    
2417        case OP_ASSERT:        case OP_ASSERT:
2418        case OP_ASSERT_NOT:        case OP_ASSERT_NOT:
# Line 1766  for (;;) Line 2439  for (;;)
2439            rlevel,                               /* function recursion level */            rlevel,                               /* function recursion level */
2440            recursing);                           /* pass on regex recursion */            recursing);                           /* pass on regex recursion */
2441    
2442            if (rc == PCRE_ERROR_DFA_UITEM) return rc;
2443          if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK))          if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK))
2444              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }
2445          }          }
# Line 1777  for (;;) Line 2451  for (;;)
2451          {          {
2452          int local_offsets[1000];          int local_offsets[1000];
2453          int local_workspace[1000];          int local_workspace[1000];
2454          int condcode = code[LINK_SIZE+1];          int codelink = GET(code, 1);
2455            int condcode;
2456    
2457            /* Because of the way auto-callout works during compile, a callout item
2458            is inserted between OP_COND and an assertion condition. This does not
2459            happen for the other conditions. */
2460    
2461            if (code[LINK_SIZE+1] == OP_CALLOUT)
2462              {
2463              rrc = 0;
2464              if (pcre_callout != NULL)
2465                {
2466                pcre_callout_block cb;
2467                cb.version          = 1;   /* Version 1 of the callout block */
2468                cb.callout_number   = code[LINK_SIZE+2];
2469                cb.offset_vector    = offsets;
2470                cb.subject          = (PCRE_SPTR)start_subject;
2471                cb.subject_length   = end_subject - start_subject;
2472                cb.start_match      = current_subject - start_subject;
2473                cb.current_position = ptr - start_subject;
2474                cb.pattern_position = GET(code, LINK_SIZE + 3);
2475                cb.next_item_length = GET(code, 3 + 2*LINK_SIZE);
2476                cb.capture_top      = 1;
2477                cb.capture_last     = -1;
2478                cb.callout_data     = md->callout_data;
2479                if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc;   /* Abandon */
2480                }
2481              if (rrc > 0) break;                      /* Fail this thread */
2482              code += _pcre_OP_lengths[OP_CALLOUT];    /* Skip callout data */
2483              }
2484    
2485            condcode = code[LINK_SIZE+1];
2486    
2487          /* Back reference conditions are not supported */          /* Back reference conditions are not supported */
2488    
2489          if (condcode == OP_CREF) return PCRE_ERROR_DFA_UCOND;          if (condcode == OP_CREF || condcode == OP_NCREF)
2490              return PCRE_ERROR_DFA_UCOND;
2491    
2492          /* The DEFINE condition is always false */          /* The DEFINE condition is always false */
2493    
2494          if (condcode == OP_DEF)          if (condcode == OP_DEF)
2495            {            { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
           ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0);  
           }  
2496    
2497          /* The only supported version of OP_RREF is for the value RREF_ANY,          /* The only supported version of OP_RREF is for the value RREF_ANY,
2498          which means "test if in any recursion". We can't test for specifically          which means "test if in any recursion". We can't test for specifically
2499          recursed groups. */          recursed groups. */
2500    
2501          else if (condcode == OP_RREF)          else if (condcode == OP_RREF || condcode == OP_NRREF)
2502            {            {
2503            int value = GET2(code, LINK_SIZE+2);            int value = GET2(code, LINK_SIZE+2);
2504            if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND;            if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND;
2505            if (recursing > 0) { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); }            if (recursing > 0)
2506              else { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); }              { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); }
2507              else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
2508            }            }
2509    
2510          /* Otherwise, the condition is an assertion */          /* Otherwise, the condition is an assertion */
# Line 1825  for (;;) Line 2530  for (;;)
2530              rlevel,                               /* function recursion level */              rlevel,                               /* function recursion level */
2531              recursing);                           /* pass on regex recursion */              recursing);                           /* pass on regex recursion */
2532    
2533              if (rc == PCRE_ERROR_DFA_UITEM) return rc;
2534            if ((rc >= 0) ==            if ((rc >= 0) ==
2535                  (condcode == OP_ASSERT || condcode == OP_ASSERTBACK))                  (condcode == OP_ASSERT || condcode == OP_ASSERTBACK))
2536              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }
2537            else            else
2538              { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); }              { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
2539            }            }
2540          }          }
2541        break;        break;
# Line 1981  for (;;) Line 2687  for (;;)
2687        /* Handle callouts */        /* Handle callouts */
2688    
2689        case OP_CALLOUT:        case OP_CALLOUT:
2690          rrc = 0;
2691        if (pcre_callout != NULL)        if (pcre_callout != NULL)
2692          {          {
         int rrc;  
2693          pcre_callout_block cb;          pcre_callout_block cb;
2694          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 1;   /* Version 1 of the callout block */
2695          cb.callout_number   = code[1];          cb.callout_number   = code[1];
# Line 1998  for (;;) Line 2704  for (;;)
2704          cb.capture_last     = -1;          cb.capture_last     = -1;
2705          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
2706          if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc;   /* Abandon */          if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc;   /* Abandon */
         if (rrc == 0) { ADD_ACTIVE(state_offset + 2 + 2*LINK_SIZE, 0); }  
2707          }          }
2708          if (rrc == 0)
2709            { ADD_ACTIVE(state_offset + _pcre_OP_lengths[OP_CALLOUT], 0); }
2710        break;        break;
2711    
2712    
# Line 2015  for (;;) Line 2722  for (;;)
2722    /* We have finished the processing at the current subject character. If no    /* We have finished the processing at the current subject character. If no
2723    new states have been set for the next character, we have found all the    new states have been set for the next character, we have found all the
2724    matches that we are going to find. If we are at the top level and partial    matches that we are going to find. If we are at the top level and partial
2725    matching has been requested, check for appropriate conditions. */    matching has been requested, check for appropriate conditions.
2726    
2727      The "forced_ fail" variable counts the number of (*F) encountered for the
2728      character. If it is equal to the original active_count (saved in
2729      workspace[1]) it means that (*F) was found on every active state. In this
2730      case we don't want to give a partial match.
2731    
2732      The "could_continue" variable is true if a state could have continued but
2733      for the fact that the end of the subject was reached. */
2734    
2735    if (new_count <= 0)    if (new_count <= 0)
2736      {      {
2737      if (match_count < 0 &&                     /* No matches found */      if (rlevel == 1 &&                               /* Top level, and */
2738          rlevel == 1 &&                         /* Top level match function */          could_continue &&                            /* Some could go on */
2739          (md->moptions & PCRE_PARTIAL) != 0 &&  /* Want partial matching */          forced_fail != workspace[1] &&               /* Not all forced fail & */
2740          ptr >= end_subject &&                  /* Reached end of subject */          (                                            /* either... */
2741          ptr > current_subject)                 /* Matched non-empty string */          (md->moptions & PCRE_PARTIAL_HARD) != 0      /* Hard partial */
2742            ||                                           /* or... */
2743            ((md->moptions & PCRE_PARTIAL_SOFT) != 0 &&  /* Soft partial and */
2744             match_count < 0)                            /* no matches */
2745            ) &&                                         /* And... */
2746            ptr >= end_subject &&                     /* Reached end of subject */
2747            ptr > current_subject)                    /* Matched non-empty string */
2748        {        {
2749        if (offsetcount >= 2)        if (offsetcount >= 2)
2750          {          {
2751          offsets[0] = current_subject - start_subject;          offsets[0] = md->start_used_ptr - start_subject;
2752          offsets[1] = end_subject - start_subject;          offsets[1] = end_subject - start_subject;
2753          }          }
2754        match_count = PCRE_ERROR_PARTIAL;        match_count = PCRE_ERROR_PARTIAL;
# Line 2082  Returns:          > 0 => number of match Line 2803  Returns:          > 0 => number of match
2803                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
2804  */  */
2805    
2806  PCRE_EXP_DEFN int  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
2807  pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data,
2808    const char *subject, int length, int start_offset, int options, int *offsets,    const char *subject, int length, int start_offset, int options, int *offsets,
2809    int offsetcount, int *workspace, int wscount)    int offsetcount, int *workspace, int wscount)
# Line 2169  md->start_code = (const uschar *)argumen Line 2890  md->start_code = (const uschar *)argumen
2890      re->name_table_offset + re->name_count * re->name_entry_size;      re->name_table_offset + re->name_count * re->name_entry_size;
2891  md->start_subject = (const unsigned char *)subject;  md->start_subject = (const unsigned char *)subject;
2892  md->end_subject = end_subject;  md->end_subject = end_subject;
2893    md->start_offset = start_offset;
2894  md->moptions = options;  md->moptions = options;
2895  md->poptions = re->options;  md->poptions = re->options;
2896    
2897    /* If the BSR option is not set at match time, copy what was set
2898    at compile time. */
2899    
2900    if ((md->moptions & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == 0)
2901      {
2902      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
2903        md->moptions |= re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE);
2904    #ifdef BSR_ANYCRLF
2905      else md->moptions |= PCRE_BSR_ANYCRLF;
2906    #endif
2907      }
2908    
2909  /* Handle different types of newline. The three bits give eight cases. If  /* Handle different types of newline. The three bits give eight cases. If
2910  nothing is set at run time, whatever was used at compile time applies. */  nothing is set at run time, whatever was used at compile time applies. */
2911    
# Line 2179  switch ((((options & PCRE_NEWLINE_BITS) Line 2913  switch ((((options & PCRE_NEWLINE_BITS)
2913           PCRE_NEWLINE_BITS)           PCRE_NEWLINE_BITS)
2914    {    {
2915    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
2916    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = CHAR_CR; break;
2917    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = CHAR_NL; break;
2918    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
2919         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break;
2920    case PCRE_NEWLINE_ANY: newline = -1; break;    case PCRE_NEWLINE_ANY: newline = -1; break;
2921    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
2922    default: return PCRE_ERROR_BADNEWLINE;    default: return PCRE_ERROR_BADNEWLINE;
# Line 2242  if (md->tables == NULL) md->tables = _pc Line 2976  if (md->tables == NULL) md->tables = _pc
2976  used in a loop when finding where to start. */  used in a loop when finding where to start. */
2977    
2978  lcc = md->tables + lcc_offset;  lcc = md->tables + lcc_offset;
2979  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
2980  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
2981    
2982  /* Set up the first character to match, if available. The first_byte value is  /* Set up the first character to match, if available. The first_byte value is
# Line 2253  studied, there may be a bitmap of possib Line 2987  studied, there may be a bitmap of possib
2987    
2988  if (!anchored)  if (!anchored)
2989    {    {
2990    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
2991      {      {
2992      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
2993      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
# Line 2261  if (!anchored) Line 2995  if (!anchored)
2995      }      }
2996    else    else
2997      {      {
2998      if (startline && study != NULL &&      if (!startline && study != NULL &&
2999           (study->options & PCRE_STUDY_MAPPED) != 0)           (study->flags & PCRE_STUDY_MAPPED) != 0)
3000        start_bits = study->start_bits;        start_bits = study->start_bits;
3001      }      }
3002    }    }
# Line 2270  if (!anchored) Line 3004  if (!anchored)
3004  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
3005  character" set. */  character" set. */
3006    
3007  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
3008    {    {
3009    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
3010    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
# Line 2278  if ((re->options & PCRE_REQCHSET) != 0) Line 3012  if ((re->options & PCRE_REQCHSET) != 0)
3012    }    }
3013    
3014  /* Call the main matching function, looping for a non-anchored regex after a  /* Call the main matching function, looping for a non-anchored regex after a
3015  failed match. Unless restarting, optimize by moving to the first match  failed match. If not restarting, perform certain optimizations at the start of
3016  character if possible, when not anchored. Then unless wanting a partial match,  a match. */
 check for a required later character. */  
3017    
3018  for (;;)  for (;;)
3019    {    {
# Line 2290  for (;;) Line 3023  for (;;)
3023      {      {
3024      const uschar *save_end_subject = end_subject;      const uschar *save_end_subject = end_subject;
3025    
3026      /* Advance to a unique first char if possible. If firstline is TRUE, the      /* If firstline is TRUE, the start of the match is constrained to the first
3027      start of the match is constrained to the first line of a multiline string.      line of a multiline string. Implement this by temporarily adjusting
3028      Implement this by temporarily adjusting end_subject so that we stop      end_subject so that we stop scanning at a newline. If the match fails at
3029      scanning at a newline. If the match fails at the newline, later code breaks      the newline, later code breaks this loop. */
     this loop. */  
3030    
3031      if (firstline)      if (firstline)
3032        {        {
3033        const uschar *t = current_subject;        USPTR t = current_subject;
3034    #ifdef SUPPORT_UTF8
3035          if (utf8)
3036            {
3037            while (t < md->end_subject && !IS_NEWLINE(t))
3038              {
3039              t++;
3040              while (t < end_subject && (*t & 0xc0) == 0x80) t++;
3041              }
3042            }
3043          else
3044    #endif
3045        while (t < md->end_subject && !IS_NEWLINE(t)) t++;        while (t < md->end_subject && !IS_NEWLINE(t)) t++;
3046        end_subject = t;        end_subject = t;
3047        }        }
3048    
3049      if (first_byte >= 0)      /* There are some optimizations that avoid running the match if a known
3050        starting point is not found. However, there is an option that disables
3051        these, for testing and for ensuring that all callouts do actually occur. */
3052    
3053        if ((options & PCRE_NO_START_OPTIMIZE) == 0)
3054        {        {
3055        if (first_byte_caseless)        /* Advance to a known first byte. */
         while (current_subject < end_subject &&  
                lcc[*current_subject] != first_byte)  
           current_subject++;  
       else  
         while (current_subject < end_subject && *current_subject != first_byte)  
           current_subject++;  
       }  
3056    
3057      /* Or to just after a linebreak for a multiline match if possible */        if (first_byte >= 0)
3058            {
3059            if (first_byte_caseless)
3060              while (current_subject < end_subject &&
3061                     lcc[*current_subject] != first_byte)
3062                current_subject++;
3063            else
3064              while (current_subject < end_subject &&
3065                     *current_subject != first_byte)
3066                current_subject++;
3067            }
3068    
3069      else if (startline)        /* Or to just after a linebreak for a multiline match if possible */
3070        {  
3071        if (current_subject > md->start_subject + start_offset)        else if (startline)
3072          {          {
3073          while (current_subject <= end_subject && !WAS_NEWLINE(current_subject))          if (current_subject > md->start_subject + start_offset)
3074            current_subject++;            {
3075    #ifdef SUPPORT_UTF8
3076              if (utf8)
3077                {
3078                while (current_subject < end_subject &&
3079                       !WAS_NEWLINE(current_subject))
3080                  {
3081                  current_subject++;
3082                  while(current_subject < end_subject &&
3083                        (*current_subject & 0xc0) == 0x80)
3084                    current_subject++;
3085                  }
3086                }
3087              else
3088    #endif
3089              while (current_subject < end_subject && !WAS_NEWLINE(current_subject))
3090                current_subject++;
3091    
3092              /* If we have just passed a CR and the newline option is ANY or
3093              ANYCRLF, and we are now at a LF, advance the match position by one
3094              more character. */
3095    
3096          /* If we have just passed a CR and the newline option is ANY or            if (current_subject[-1] == CHAR_CR &&
3097          ANYCRLF, and we are now at a LF, advance the match position by one more                 (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
3098          character. */                 current_subject < end_subject &&
3099                   *current_subject == CHAR_NL)
3100          if (current_subject[-1] == '\r' &&              current_subject++;
3101               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&            }
              current_subject < end_subject &&  
              *current_subject == '\n')  
           current_subject++;  
3102          }          }
       }  
3103    
3104      /* Or to a non-unique first char after study */        /* Or to a non-unique first char after study */
3105    
3106      else if (start_bits != NULL)        else if (start_bits != NULL)
       {  
       while (current_subject < end_subject)  
3107          {          {
3108          register unsigned int c = *current_subject;          while (current_subject < end_subject)
3109          if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++;            {
3110            else break;            register unsigned int c = *current_subject;
3111              if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++;
3112                else break;
3113              }
3114          }          }
3115        }        }
3116    
3117      /* Restore fudged end_subject */      /* Restore fudged end_subject */
3118    
3119      end_subject = save_end_subject;      end_subject = save_end_subject;
     }  
   
   /* If req_byte is set, we know that that character must appear in the subject  
   for the match to succeed. If the first character is set, req_byte must be  
   later in the subject; otherwise the test starts at the match point. This  
   optimization can save a huge amount of work in patterns with nested unlimited  
   repeats that aren't going to match. Writing separate code for cased/caseless  
   versions makes it go faster, as does using an autoincrement and backing off  
   on a match.  
   
   HOWEVER: when the subject string is very, very long, searching to its end can  
   take a long time, and give bad performance on quite ordinary patterns. This  
   showed up when somebody was matching /^C/ on a 32-megabyte string... so we  
   don't do this when the string is sufficiently long.  
   
   ALSO: this processing is disabled when partial matching is requested.  
   */  
   
   if (req_byte >= 0 &&  
       end_subject - current_subject < REQ_BYTE_MAX &&  
       (options & PCRE_PARTIAL) == 0)  
     {  
     register const uschar *p = current_subject + ((first_byte >= 0)? 1 : 0);  
3120    
3121      /* We don't need to repeat the search if we haven't yet reached the      /* The following two optimizations are disabled for partial matching or if
3122      place we found it at last time. */      disabling is explicitly requested (and of course, by the test above, this
3123        code is not obeyed when restarting after a partial match). */
3124    
3125      if (p > req_byte_ptr)      if ((options & PCRE_NO_START_OPTIMIZE) == 0 &&
3126            (options & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) == 0)
3127        {        {
3128        if (req_byte_caseless)        /* If the pattern was studied, a minimum subject length may be set. This
3129          {        is a lower bound; no actual string of that length may actually match the
3130          while (p < end_subject)        pattern. Although the value is, strictly, in characters, we treat it as
3131            {        bytes to avoid spending too much time in this optimization. */
3132            register int pp = *p++;  
3133            if (pp == req_byte || pp == req_byte2) { p--; break; }        if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 &&
3134            }            (pcre_uint32)(end_subject - current_subject) < study->minlength)
3135          }          return PCRE_ERROR_NOMATCH;
3136        else  
3137          /* If req_byte is set, we know that that character must appear in the
3138          subject for the match to succeed. If the first character is set, req_byte
3139          must be later in the subject; otherwise the test starts at the match
3140          point. This optimization can save a huge amount of work in patterns with
3141          nested unlimited repeats that aren't going to match. Writing separate
3142          code for cased/caseless versions makes it go faster, as does using an
3143          autoincrement and backing off on a match.
3144    
3145          HOWEVER: when the subject string is very, very long, searching to its end
3146          can take a long time, and give bad performance on quite ordinary
3147          patterns. This showed up when somebody was matching /^C/ on a 32-megabyte
3148          string... so we don't do this when the string is sufficiently long. */
3149    
3150          if (req_byte >= 0 && end_subject - current_subject < REQ_BYTE_MAX)
3151          {          {
3152          while (p < end_subject)          register const uschar *p = current_subject + ((first_byte >= 0)? 1 : 0);
3153    
3154            /* We don't need to repeat the search if we haven't yet reached the
3155            place we found it at last time. */
3156    
3157            if (p > req_byte_ptr)
3158            {            {
3159            if (*p++ == req_byte) { p--; break; }            if (req_byte_caseless)
3160            }              {
3161          }              while (p < end_subject)
3162                  {
3163                  register int pp = *p++;
3164                  if (pp == req_byte || pp == req_byte2) { p--; break; }
3165                  }
3166                }
3167              else
3168                {
3169                while (p < end_subject)
3170                  {
3171                  if (*p++ == req_byte) { p--; break; }
3172                  }
3173                }
3174    
3175        /* If we can't find the required character, break the matching loop,            /* If we can't find the required character, break the matching loop,
3176        which will cause a return or PCRE_ERROR_NOMATCH. */            which will cause a return or PCRE_ERROR_NOMATCH. */
3177    
3178        if (p >= end_subject) break;            if (p >= end_subject) break;
3179    
3180        /* If we have found the required character, save the point where we            /* If we have found the required character, save the point where we
3181        found it, so that we don't search again next time round the loop if            found it, so that we don't search again next time round the loop if
3182        the start hasn't passed this character yet. */            the start hasn't passed this character yet. */
3183    
3184        req_byte_ptr = p;            req_byte_ptr = p;
3185              }
3186            }
3187        }        }
3188      }      }   /* End of optimizations that are done when not restarting */
3189    
3190    /* OK, now we can do the business */    /* OK, now we can do the business */
3191    
3192      md->start_used_ptr = current_subject;
3193    
3194    rc = internal_dfa_exec(    rc = internal_dfa_exec(
3195      md,                                /* fixed match data */      md,                                /* fixed match data */
3196      md->start_code,                    /* this subexpression's code */      md->start_code,                    /* this subexpression's code */
# Line 2440  for (;;) Line 3221  for (;;)
3221      }      }
3222    if (current_subject > end_subject) break;    if (current_subject > end_subject) break;
3223    
3224    /* If we have just passed a CR and the newline option is CRLF or ANY or    /* If we have just passed a CR and we are now at a LF, and the pattern does
3225    ANYCRLF, and we are now at a LF, advance the match position by one more    not contain any explicit matches for \r or \n, and the newline option is CRLF
3226    character. */    or ANY or ANYCRLF, advance the match position by one more character. */
3227    
3228    if (current_subject[-1] == '\r' &&    if (current_subject[-1] == CHAR_CR &&
3229         (md->nltype == NLTYPE_ANY ||        current_subject < end_subject &&
3230          md->nltype == NLTYPE_ANYCRLF ||        *current_subject == CHAR_NL &&
3231          md->nllen == 2) &&        (re->flags & PCRE_HASCRORLF) == 0 &&
3232         current_subject < end_subject &&          (md->nltype == NLTYPE_ANY ||
3233         *current_subject == '\n')           md->nltype == NLTYPE_ANYCRLF ||
3234             md->nllen == 2))
3235      current_subject++;      current_subject++;
3236    
3237    }   /* "Bumpalong" loop */    }   /* "Bumpalong" loop */

Legend:
Removed from v.168  
changed lines
  Added in v.518

  ViewVC Help
Powered by ViewVC 1.1.5