/[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 200 by ph10, Wed Aug 1 09:10:40 2007 UTC revision 613 by ph10, Sat Jul 2 16:59:52 2011 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) 1997-2011 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  #ifdef HAVE_CONFIG_H
77  #include <config.h>  #include "config.h"
78  #endif  #endif
79    
80  #define NLBLOCK md             /* Block containing newline information */  #define NLBLOCK md             /* Block containing newline information */
# Line 60  applications. */ Line 89  applications. */
89  #define SP "                   "  #define SP "                   "
90    
91    
   
92  /*************************************************  /*************************************************
93  *      Code parameters and static tables         *  *      Code parameters and static tables         *
94  *************************************************/  *************************************************/
# Line 78  never stored, so we push them well clear Line 106  never stored, so we push them well clear
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,                       /* NOTPROP, PROP, EXTUNI                  */    0, 0,                          /* \P, \p                                 */
122    0, 0, 0, 0, 0,                 /* \R, \H, \h, \V, \v                     */    0, 0, 0, 0, 0,                 /* \R, \H, \h, \V, \v                     */
123    0, 0, 0, 0, 0,                 /* \Z, \z, Opt, ^, $                      */    0,                             /* \X                                     */
124      0, 0, 0, 0, 0, 0,              /* \Z, \z, ^, ^M, $, $M                   */
125    1,                             /* Char                                   */    1,                             /* Char                                   */
126    1,                             /* Charnc                                 */    1,                             /* Chari                                  */
127    1,                             /* not                                    */    1,                             /* not                                    */
128      1,                             /* noti                                   */
129    /* Positive single-char repeats                                          */    /* Positive single-char repeats                                          */
130    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */
131    3, 3, 3,                       /* upto, minupto, exact                   */    3, 3, 3,                       /* upto, minupto, exact                   */
132    1, 1, 1, 3,                    /* *+, ++, ?+, upto+                      */    1, 1, 1, 3,                    /* *+, ++, ?+, upto+                      */
133      1, 1, 1, 1, 1, 1,              /* *I, *?I, +I, +?I, ?I, ??I              */
134      3, 3, 3,                       /* upto I, minupto I, exact I             */
135      1, 1, 1, 3,                    /* *+I, ++I, ?+I, upto+I                  */
136    /* Negative single-char repeats - only for chars < 256                   */    /* Negative single-char repeats - only for chars < 256                   */
137    1, 1, 1, 1, 1, 1,              /* NOT *, *?, +, +?, ?, ??                */    1, 1, 1, 1, 1, 1,              /* NOT *, *?, +, +?, ?, ??                */
138    3, 3, 3,                       /* NOT upto, minupto, exact               */    3, 3, 3,                       /* NOT upto, minupto, exact               */
139    1, 1, 1, 3,                    /* NOT *+, ++, ?+, updo+                  */    1, 1, 1, 3,                    /* NOT *+, ++, ?+, upto+                  */
140      1, 1, 1, 1, 1, 1,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */
141      3, 3, 3,                       /* NOT upto I, minupto I, exact I         */
142      1, 1, 1, 3,                    /* NOT *+I, ++I, ?+I, upto+I              */
143    /* Positive type repeats                                                 */    /* Positive type repeats                                                 */
144    1, 1, 1, 1, 1, 1,              /* Type *, *?, +, +?, ?, ??               */    1, 1, 1, 1, 1, 1,              /* Type *, *?, +, +?, ?, ??               */
145    3, 3, 3,                       /* Type upto, minupto, exact              */    3, 3, 3,                       /* Type upto, minupto, exact              */
# Line 114  static uschar coptable[] = { Line 151  static uschar coptable[] = {
151    0,                             /* NCLASS                                 */    0,                             /* NCLASS                                 */
152    0,                             /* XCLASS - variable length               */    0,                             /* XCLASS - variable length               */
153    0,                             /* REF                                    */    0,                             /* REF                                    */
154      0,                             /* REFI                                   */
155    0,                             /* RECURSE                                */    0,                             /* RECURSE                                */
156    0,                             /* CALLOUT                                */    0,                             /* CALLOUT                                */
157    0,                             /* Alt                                    */    0,                             /* Alt                                    */
158    0,                             /* Ket                                    */    0,                             /* Ket                                    */
159    0,                             /* KetRmax                                */    0,                             /* KetRmax                                */
160    0,                             /* KetRmin                                */    0,                             /* KetRmin                                */
161      0,                             /* KetRpos                                */
162    0,                             /* Assert                                 */    0,                             /* Assert                                 */
163    0,                             /* Assert not                             */    0,                             /* Assert not                             */
164    0,                             /* Assert behind                          */    0,                             /* Assert behind                          */
165    0,                             /* Assert behind not                      */    0,                             /* Assert behind not                      */
166    0,                             /* Reverse                                */    0,                             /* Reverse                                */
167    0, 0, 0, 0,                    /* ONCE, BRA, CBRA, COND                  */    0, 0, 0, 0, 0, 0,              /* ONCE, BRA, BRAPOS, CBRA, CBRAPOS, COND */
168    0, 0, 0,                       /* SBRA, SCBRA, SCOND                     */    0, 0, 0, 0, 0,                 /* SBRA, SBRAPOS, SCBRA, SCBRAPOS, SCOND  */
169    0,                             /* CREF                                   */    0, 0,                          /* CREF, NCREF                            */
170    0,                             /* RREF                                   */    0, 0,                          /* RREF, NRREF                            */
171    0,                             /* DEF                                    */    0,                             /* DEF                                    */
172    0, 0                           /* BRAZERO, BRAMINZERO                    */    0, 0, 0,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */
173      0, 0, 0,                       /* MARK, PRUNE, PRUNE_ARG                 */
174      0, 0, 0, 0,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG         */
175      0, 0, 0, 0,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */
176      0, 0                           /* CLOSE, SKIPZERO  */
177    };
178    
179    /* This table identifies those opcodes that inspect a character. It is used to
180    remember the fact that a character could have been inspected when the end of
181    the subject is reached. ***NOTE*** If the start of this table is modified, the
182    two tables that follow must also be modified. */
183    
184    static const uschar poptable[] = {
185      0,                             /* End                                    */
186      0, 0, 0, 1, 1,                 /* \A, \G, \K, \B, \b                     */
187      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */
188      1, 1, 1,                       /* Any, AllAny, Anybyte                   */
189      1, 1,                          /* \P, \p                                 */
190      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */
191      1,                             /* \X                                     */
192      0, 0, 0, 0, 0, 0,              /* \Z, \z, ^, ^M, $, $M                   */
193      1,                             /* Char                                   */
194      1,                             /* Chari                                  */
195      1,                             /* not                                    */
196      1,                             /* noti                                   */
197      /* Positive single-char repeats                                          */
198      1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */
199      1, 1, 1,                       /* upto, minupto, exact                   */
200      1, 1, 1, 1,                    /* *+, ++, ?+, upto+                      */
201      1, 1, 1, 1, 1, 1,              /* *I, *?I, +I, +?I, ?I, ??I              */
202      1, 1, 1,                       /* upto I, minupto I, exact I             */
203      1, 1, 1, 1,                    /* *+I, ++I, ?+I, upto+I                  */
204      /* Negative single-char repeats - only for chars < 256                   */
205      1, 1, 1, 1, 1, 1,              /* NOT *, *?, +, +?, ?, ??                */
206      1, 1, 1,                       /* NOT upto, minupto, exact               */
207      1, 1, 1, 1,                    /* NOT *+, ++, ?+, upto+                  */
208      1, 1, 1, 1, 1, 1,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */
209      1, 1, 1,                       /* NOT upto I, minupto I, exact I         */
210      1, 1, 1, 1,                    /* NOT *+I, ++I, ?+I, upto+I              */
211      /* Positive type repeats                                                 */
212      1, 1, 1, 1, 1, 1,              /* Type *, *?, +, +?, ?, ??               */
213      1, 1, 1,                       /* Type upto, minupto, exact              */
214      1, 1, 1, 1,                    /* Type *+, ++, ?+, upto+                 */
215      /* Character class & ref repeats                                         */
216      1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */
217      1, 1,                          /* CRRANGE, CRMINRANGE                    */
218      1,                             /* CLASS                                  */
219      1,                             /* NCLASS                                 */
220      1,                             /* XCLASS - variable length               */
221      0,                             /* REF                                    */
222      0,                             /* REFI                                   */
223      0,                             /* RECURSE                                */
224      0,                             /* CALLOUT                                */
225      0,                             /* Alt                                    */
226      0,                             /* Ket                                    */
227      0,                             /* KetRmax                                */
228      0,                             /* KetRmin                                */
229      0,                             /* KetRpos                                */
230      0,                             /* Assert                                 */
231      0,                             /* Assert not                             */
232      0,                             /* Assert behind                          */
233      0,                             /* Assert behind not                      */
234      0,                             /* Reverse                                */
235      0, 0, 0, 0, 0, 0,              /* ONCE, BRA, BRAPOS, CBRA, CBRAPOS, COND */
236      0, 0, 0, 0, 0,                 /* SBRA, SBRAPOS, SCBRA, SCBRAPOS, SCOND  */
237      0, 0,                          /* CREF, NCREF                            */
238      0, 0,                          /* RREF, NRREF                            */
239      0,                             /* DEF                                    */
240      0, 0, 0,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */
241      0, 0, 0,                       /* MARK, PRUNE, PRUNE_ARG                 */
242      0, 0, 0, 0,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG         */
243      0, 0, 0, 0,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */
244      0, 0                           /* CLOSE, SKIPZERO                        */
245  };  };
246    
247  /* 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,
248  and \w */  and \w */
249    
250  static uschar toptable1[] = {  static const uschar toptable1[] = {
251    0, 0, 0, 0, 0, 0,    0, 0, 0, 0, 0, 0,
252    ctype_digit, ctype_digit,    ctype_digit, ctype_digit,
253    ctype_space, ctype_space,    ctype_space, ctype_space,
254    ctype_word,  ctype_word,    ctype_word,  ctype_word,
255    0                               /* OP_ANY */    0, 0                            /* OP_ANY, OP_ALLANY */
256  };  };
257    
258  static uschar toptable2[] = {  static const uschar toptable2[] = {
259    0, 0, 0, 0, 0, 0,    0, 0, 0, 0, 0, 0,
260    ctype_digit, 0,    ctype_digit, 0,
261    ctype_space, 0,    ctype_space, 0,
262    ctype_word,  0,    ctype_word,  0,
263    1                               /* OP_ANY */    1, 1                            /* OP_ANY, OP_ALLANY */
264  };  };
265    
266    
# Line 161  these structures in, is a vector of ints Line 272  these structures in, is a vector of ints
272  typedef struct stateblock {  typedef struct stateblock {
273    int offset;                     /* Offset to opcode */    int offset;                     /* Offset to opcode */
274    int count;                      /* Count for repeats */    int count;                      /* Count for repeats */
   int ims;                        /* ims flag bits */  
275    int data;                       /* Some use extra data */    int data;                       /* Some use extra data */
276  } stateblock;  } stateblock;
277    
278  #define INTS_PER_STATEBLOCK  (sizeof(stateblock)/sizeof(int))  #define INTS_PER_STATEBLOCK  (sizeof(stateblock)/sizeof(int))
279    
280    
281  #ifdef DEBUG  #ifdef PCRE_DEBUG
282  /*************************************************  /*************************************************
283  *             Print character string             *  *             Print character string             *
284  *************************************************/  *************************************************/
# Line 217  Arguments: Line 327  Arguments:
327    offsetcount       size of same    offsetcount       size of same
328    workspace         vector of workspace    workspace         vector of workspace
329    wscount           size of same    wscount           size of same
   ims               the current ims flags  
330    rlevel            function call recursion level    rlevel            function call recursion level
331    recursing         regex recursive call level    recursing         regex recursive call level
332    
333  Returns:            > 0 =>  Returns:            > 0 => number of match offset pairs placed in offsets
334                      = 0 =>                      = 0 => offsets overflowed; longest matches are present
335                       -1 => failed to match                       -1 => failed to match
336                     < -1 => some kind of unexpected problem                     < -1 => some kind of unexpected problem
337    
# Line 234  for the current character, one for the f Line 343  for the current character, one for the f
343      { \      { \
344      next_active_state->offset = (x); \      next_active_state->offset = (x); \
345      next_active_state->count  = (y); \      next_active_state->count  = (y); \
     next_active_state->ims    = ims; \  
346      next_active_state++; \      next_active_state++; \
347      DPRINTF(("%.*sADD_ACTIVE(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \      DPRINTF(("%.*sADD_ACTIVE(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \
348      } \      } \
# Line 245  for the current character, one for the f Line 353  for the current character, one for the f
353      { \      { \
354      next_active_state->offset = (x); \      next_active_state->offset = (x); \
355      next_active_state->count  = (y); \      next_active_state->count  = (y); \
     next_active_state->ims    = ims; \  
356      next_active_state->data   = (z); \      next_active_state->data   = (z); \
357      next_active_state++; \      next_active_state++; \
358      DPRINTF(("%.*sADD_ACTIVE_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \      DPRINTF(("%.*sADD_ACTIVE_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \
# Line 257  for the current character, one for the f Line 364  for the current character, one for the f
364      { \      { \
365      next_new_state->offset = (x); \      next_new_state->offset = (x); \
366      next_new_state->count  = (y); \      next_new_state->count  = (y); \
     next_new_state->ims    = ims; \  
367      next_new_state++; \      next_new_state++; \
368      DPRINTF(("%.*sADD_NEW(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \      DPRINTF(("%.*sADD_NEW(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \
369      } \      } \
# Line 268  for the current character, one for the f Line 374  for the current character, one for the f
374      { \      { \
375      next_new_state->offset = (x); \      next_new_state->offset = (x); \
376      next_new_state->count  = (y); \      next_new_state->count  = (y); \
     next_new_state->ims    = ims; \  
377      next_new_state->data   = (z); \      next_new_state->data   = (z); \
378      next_new_state++; \      next_new_state++; \
379      DPRINTF(("%.*sADD_NEW_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \      DPRINTF(("%.*sADD_NEW_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \
# Line 287  internal_dfa_exec( Line 392  internal_dfa_exec(
392    int offsetcount,    int offsetcount,
393    int *workspace,    int *workspace,
394    int wscount,    int wscount,
   int ims,  
395    int  rlevel,    int  rlevel,
396    int  recursing)    int  recursing)
397  {  {
# Line 335  next_new_state = new_states = active_sta Line 439  next_new_state = new_states = active_sta
439  new_count = 0;  new_count = 0;
440    
441  first_op = this_start_code + 1 + LINK_SIZE +  first_op = this_start_code + 1 + LINK_SIZE +
442    ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA)? 2:0);    ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA ||
443        *this_start_code == OP_CBRAPOS || *this_start_code == OP_SCBRAPOS)? 2:0);
444    
445  /* The first thing in any (sub) pattern is a bracket of some sort. Push all  /* The first thing in any (sub) pattern is a bracket of some sort. Push all
446  the alternative states onto the list, and find out where the end is. This  the alternative states onto the list, and find out where the end is. This
# Line 384  if (*first_op == OP_REVERSE) Line 489  if (*first_op == OP_REVERSE)
489    
490      {      {
491      gone_back = (current_subject - max_back < start_subject)?      gone_back = (current_subject - max_back < start_subject)?
492        current_subject - start_subject : max_back;        (int)(current_subject - start_subject) : max_back;
493      current_subject -= gone_back;      current_subject -= gone_back;
494      }      }
495    
496      /* Save the earliest consulted character */
497    
498      if (current_subject < md->start_used_ptr)
499        md->start_used_ptr = current_subject;
500    
501    /* Now we can process the individual branches. */    /* Now we can process the individual branches. */
502    
503    end_code = this_start_code;    end_code = this_start_code;
# Line 396  if (*first_op == OP_REVERSE) Line 506  if (*first_op == OP_REVERSE)
506      int back = GET(end_code, 2+LINK_SIZE);      int back = GET(end_code, 2+LINK_SIZE);
507      if (back <= gone_back)      if (back <= gone_back)
508        {        {
509        int bstate = end_code - start_code + 2 + 2*LINK_SIZE;        int bstate = (int)(end_code - start_code + 2 + 2*LINK_SIZE);
510        ADD_NEW_DATA(-bstate, 0, gone_back - back);        ADD_NEW_DATA(-bstate, 0, gone_back - back);
511        }        }
512      end_code += GET(end_code, 1);      end_code += GET(end_code, 1);
# Line 429  else Line 539  else
539    else    else
540      {      {
541      int length = 1 + LINK_SIZE +      int length = 1 + LINK_SIZE +
542        ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA)? 2:0);        ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA ||
543            *this_start_code == OP_CBRAPOS || *this_start_code == OP_SCBRAPOS)?
544            2:0);
545      do      do
546        {        {
547        ADD_NEW(end_code - start_code + length, 0);        ADD_NEW((int)(end_code - start_code + length), 0);
548        end_code += GET(end_code, 1);        end_code += GET(end_code, 1);
549        length = 1 + LINK_SIZE;        length = 1 + LINK_SIZE;
550        }        }
# Line 452  for (;;) Line 564  for (;;)
564    int i, j;    int i, j;
565    int clen, dlen;    int clen, dlen;
566    unsigned int c, d;    unsigned int c, d;
567      int forced_fail = 0;
568      BOOL could_continue = FALSE;
569    
570    /* 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
571    new state list. */    new state list. */
# Line 465  for (;;) Line 579  for (;;)
579    workspace[0] ^= 1;              /* Remember for the restarting feature */    workspace[0] ^= 1;              /* Remember for the restarting feature */
580    workspace[1] = active_count;    workspace[1] = active_count;
581    
582  #ifdef DEBUG  #ifdef PCRE_DEBUG
583    printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP);    printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP);
584    pchars((uschar *)ptr, strlen((char *)ptr), stdout);    pchars((uschar *)ptr, strlen((char *)ptr), stdout);
585    printf("\"\n");    printf("\"\n");
# Line 507  for (;;) Line 621  for (;;)
621    for (i = 0; i < active_count; i++)    for (i = 0; i < active_count; i++)
622      {      {
623      stateblock *current_state = active_states + i;      stateblock *current_state = active_states + i;
624        BOOL caseless = FALSE;
625      const uschar *code;      const uschar *code;
626      int state_offset = current_state->offset;      int state_offset = current_state->offset;
627      int count, codevalue;      int count, codevalue, rrc;
 #ifdef SUPPORT_UCP  
     int chartype, script;  
 #endif  
628    
629  #ifdef DEBUG  #ifdef PCRE_DEBUG
630      printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset);      printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset);
631      if (clen == 0) printf("EOL\n");      if (clen == 0) printf("EOL\n");
632        else if (c > 32 && c < 127) printf("'%c'\n", c);        else if (c > 32 && c < 127) printf("'%c'\n", c);
633          else printf("0x%02x\n", c);          else printf("0x%02x\n", c);
634  #endif  #endif
635    
     /* This variable is referred to implicity in the ADD_xxx macros. */  
   
     ims = current_state->ims;  
   
636      /* A negative offset is a special case meaning "hold off going to this      /* A negative offset is a special case meaning "hold off going to this
637      (negated) state until the number of characters in the data field have      (negated) state until the number of characters in the data field have
638      been skipped". */      been skipped". */
# Line 544  for (;;) Line 652  for (;;)
652          }          }
653        }        }
654    
655      /* 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.
656        See the note at the head of this module about the possibility of improving
657        performance here. */
658    
659      for (j = 0; j < i; j++)      for (j = 0; j < i; j++)
660        {        {
# Line 561  for (;;) Line 671  for (;;)
671      code = start_code + state_offset;      code = start_code + state_offset;
672      codevalue = *code;      codevalue = *code;
673    
674        /* If this opcode inspects a character, but we are at the end of the
675        subject, remember the fact for use when testing for a partial match. */
676    
677        if (clen == 0 && poptable[codevalue] != 0)
678          could_continue = TRUE;
679    
680      /* 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
681      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
682      is wrong, because sometimes zero repetitions of the subject are      is wrong, because sometimes zero repetitions of the subject are
# Line 607  for (;;) Line 723  for (;;)
723    
724      switch (codevalue)      switch (codevalue)
725        {        {
726    /* ========================================================================== */
727          /* These cases are never obeyed. This is a fudge that causes a compile-
728          time error if the vectors coptable or poptable, which are indexed by
729          opcode, are not the correct length. It seems to be the only way to do
730          such a check at compile time, as the sizeof() operator does not work
731          in the C preprocessor. */
732    
733          case OP_TABLE_LENGTH:
734          case OP_TABLE_LENGTH +
735            ((sizeof(coptable) == OP_TABLE_LENGTH) &&
736             (sizeof(poptable) == OP_TABLE_LENGTH)):
737          break;
738    
739  /* ========================================================================== */  /* ========================================================================== */
740        /* 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
741        on with the next opcode. Otherwise, unless we have an empty string and        on with the next opcode. For repeating opcodes, also add the repeat
742        PCRE_NOTEMPTY is set, save the match data, shifting up all previous        state. Note that KETRPOS will always be encountered at the end of the
743          subpattern, because the possessive subpattern repeats are always handled
744          using recursive calls. Thus, it never adds any new states.
745    
746          At the end of the (sub)pattern, unless we have an empty string and
747          PCRE_NOTEMPTY is set, or PCRE_NOTEMPTY_ATSTART is set and we are at the
748          start of the subject, save the match data, shifting up all previous
749        matches so we always have the longest first. */        matches so we always have the longest first. */
750    
751        case OP_KET:        case OP_KET:
752        case OP_KETRMIN:        case OP_KETRMIN:
753        case OP_KETRMAX:        case OP_KETRMAX:
754          case OP_KETRPOS:
755        if (code != end_code)        if (code != end_code)
756          {          {
757          ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0);          ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0);
# Line 625  for (;;) Line 760  for (;;)
760            ADD_ACTIVE(state_offset - GET(code, 1), 0);            ADD_ACTIVE(state_offset - GET(code, 1), 0);
761            }            }
762          }          }
763        else if (ptr > current_subject || (md->moptions & PCRE_NOTEMPTY) == 0)        else
764          {          {
765          if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0;          if (ptr > current_subject ||
766            else if (match_count > 0 && ++match_count * 2 >= offsetcount)              ((md->moptions & PCRE_NOTEMPTY) == 0 &&
767              match_count = 0;                ((md->moptions & PCRE_NOTEMPTY_ATSTART) == 0 ||
768          count = ((match_count == 0)? offsetcount : match_count * 2) - 2;                  current_subject > start_subject + md->start_offset)))
769          if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int));            {
770          if (offsetcount >= 2)            if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0;
771            {              else if (match_count > 0 && ++match_count * 2 >= offsetcount)
772            offsets[0] = current_subject - start_subject;                match_count = 0;
773            offsets[1] = ptr - start_subject;            count = ((match_count == 0)? offsetcount : match_count * 2) - 2;
774            DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP,            if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int));
775              offsets[1] - offsets[0], current_subject));            if (offsetcount >= 2)
776            }              {
777          if ((md->moptions & PCRE_DFA_SHORTEST) != 0)              offsets[0] = (int)(current_subject - start_subject);
778            {              offsets[1] = (int)(ptr - start_subject);
779            DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n"              DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP,
780              "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel,                offsets[1] - offsets[0], current_subject));
781              match_count, rlevel*2-2, SP));              }
782            return match_count;            if ((md->moptions & PCRE_DFA_SHORTEST) != 0)
783                {
784                DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n"
785                  "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel,
786                  match_count, rlevel*2-2, SP));
787                return match_count;
788                }
789            }            }
790          }          }
791        break;        break;
# Line 656  for (;;) Line 797  for (;;)
797        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
798        case OP_ALT:        case OP_ALT:
799        do { code += GET(code, 1); } while (*code == OP_ALT);        do { code += GET(code, 1); } while (*code == OP_ALT);
800        ADD_ACTIVE(code - start_code, 0);        ADD_ACTIVE((int)(code - start_code), 0);
801        break;        break;
802    
803        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
# Line 664  for (;;) Line 805  for (;;)
805        case OP_SBRA:        case OP_SBRA:
806        do        do
807          {          {
808          ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0);          ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
809          code += GET(code, 1);          code += GET(code, 1);
810          }          }
811        while (*code == OP_ALT);        while (*code == OP_ALT);
# Line 673  for (;;) Line 814  for (;;)
814        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
815        case OP_CBRA:        case OP_CBRA:
816        case OP_SCBRA:        case OP_SCBRA:
817        ADD_ACTIVE(code - start_code + 3 + LINK_SIZE,  0);        ADD_ACTIVE((int)(code - start_code + 3 + LINK_SIZE),  0);
818        code += GET(code, 1);        code += GET(code, 1);
819        while (*code == OP_ALT)        while (*code == OP_ALT)
820          {          {
821          ADD_ACTIVE(code - start_code + 1 + LINK_SIZE,  0);          ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE),  0);
822          code += GET(code, 1);          code += GET(code, 1);
823          }          }
824        break;        break;
# Line 688  for (;;) Line 829  for (;;)
829        ADD_ACTIVE(state_offset + 1, 0);        ADD_ACTIVE(state_offset + 1, 0);
830        code += 1 + GET(code, 2);        code += 1 + GET(code, 2);
831        while (*code == OP_ALT) code += GET(code, 1);        while (*code == OP_ALT) code += GET(code, 1);
832        ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0);        ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
833          break;
834    
835          /*-----------------------------------------------------------------*/
836          case OP_SKIPZERO:
837          code += 1 + GET(code, 2);
838          while (*code == OP_ALT) code += GET(code, 1);
839          ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
840        break;        break;
841    
842        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
843        case OP_CIRC:        case OP_CIRC:
844        if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) ||        if (ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0)
           ((ims & PCRE_MULTILINE) != 0 &&  
             ptr != end_subject &&  
             WAS_NEWLINE(ptr)))  
845          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
846        break;        break;
847    
848        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
849        case OP_EOD:        case OP_CIRCM:
850        if (ptr >= end_subject) { ADD_ACTIVE(state_offset + 1, 0); }        if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) ||
851              (ptr != end_subject && WAS_NEWLINE(ptr)))
852            { ADD_ACTIVE(state_offset + 1, 0); }
853        break;        break;
854    
855        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
856        case OP_OPT:        case OP_EOD:
857        ims = code[1];        if (ptr >= end_subject)
858        ADD_ACTIVE(state_offset + 2, 0);          {
859            if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
860              could_continue = TRUE;
861            else { ADD_ACTIVE(state_offset + 1, 0); }
862            }
863        break;        break;
864    
865        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
# Line 730  for (;;) Line 881  for (;;)
881    
882        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
883        case OP_ANY:        case OP_ANY:
884        if (clen > 0 && ((ims & PCRE_DOTALL) != 0 || !IS_NEWLINE(ptr)))        if (clen > 0 && !IS_NEWLINE(ptr))
885            { ADD_NEW(state_offset + 1, 0); }
886          break;
887    
888          /*-----------------------------------------------------------------*/
889          case OP_ALLANY:
890          if (clen > 0)
891          { ADD_NEW(state_offset + 1, 0); }          { ADD_NEW(state_offset + 1, 0); }
892        break;        break;
893    
894        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
895        case OP_EODN:        case OP_EODN:
896        if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen))        if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
897            could_continue = TRUE;
898          else if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen))
899          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
900        break;        break;
901    
# Line 744  for (;;) Line 903  for (;;)
903        case OP_DOLL:        case OP_DOLL:
904        if ((md->moptions & PCRE_NOTEOL) == 0)        if ((md->moptions & PCRE_NOTEOL) == 0)
905          {          {
906          if (clen == 0 ||          if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
907              (IS_NEWLINE(ptr) &&            could_continue = TRUE;
908                 ((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen)          else if (clen == 0 ||
909                ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) &&
910                   (ptr == end_subject - md->nllen)
911              ))              ))
912            { ADD_ACTIVE(state_offset + 1, 0); }            { ADD_ACTIVE(state_offset + 1, 0); }
913          }          }
914        else if ((ims & PCRE_MULTILINE) != 0 && IS_NEWLINE(ptr))        break;
915    
916          /*-----------------------------------------------------------------*/
917          case OP_DOLLM:
918          if ((md->moptions & PCRE_NOTEOL) == 0)
919            {
920            if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
921              could_continue = TRUE;
922            else if (clen == 0 ||
923                ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr)))
924              { ADD_ACTIVE(state_offset + 1, 0); }
925            }
926          else if (IS_NEWLINE(ptr))
927          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
928        break;        break;
929    
# Line 782  for (;;) Line 955  for (;;)
955          if (ptr > start_subject)          if (ptr > start_subject)
956            {            {
957            const uschar *temp = ptr - 1;            const uschar *temp = ptr - 1;
958              if (temp < md->start_used_ptr) md->start_used_ptr = temp;
959  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
960            if (utf8) BACKCHAR(temp);            if (utf8) BACKCHAR(temp);
961  #endif  #endif
962            GETCHARTEST(d, temp);            GETCHARTEST(d, temp);
963    #ifdef SUPPORT_UCP
964              if ((md->poptions & PCRE_UCP) != 0)
965                {
966                if (d == '_') left_word = TRUE; else
967                  {
968                  int cat = UCD_CATEGORY(d);
969                  left_word = (cat == ucp_L || cat == ucp_N);
970                  }
971                }
972              else
973    #endif
974            left_word = d < 256 && (ctypes[d] & ctype_word) != 0;            left_word = d < 256 && (ctypes[d] & ctype_word) != 0;
975            }            }
976          else left_word = 0;          else left_word = FALSE;
977    
978          if (clen > 0) right_word = c < 256 && (ctypes[c] & ctype_word) != 0;          if (clen > 0)
979            else right_word = 0;            {
980    #ifdef SUPPORT_UCP
981              if ((md->poptions & PCRE_UCP) != 0)
982                {
983                if (c == '_') right_word = TRUE; else
984                  {
985                  int cat = UCD_CATEGORY(c);
986                  right_word = (cat == ucp_L || cat == ucp_N);
987                  }
988                }
989              else
990    #endif
991              right_word = c < 256 && (ctypes[c] & ctype_word) != 0;
992              }
993            else right_word = FALSE;
994    
995          if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY))          if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY))
996            { ADD_ACTIVE(state_offset + 1, 0); }            { ADD_ACTIVE(state_offset + 1, 0); }
# Line 810  for (;;) Line 1009  for (;;)
1009        if (clen > 0)        if (clen > 0)
1010          {          {
1011          BOOL OK;          BOOL OK;
1012          int category = _pcre_ucp_findprop(c, &chartype, &script);          const ucd_record * prop = GET_UCD(c);
1013          switch(code[1])          switch(code[1])
1014            {            {
1015            case PT_ANY:            case PT_ANY:
# Line 818  for (;;) Line 1017  for (;;)
1017            break;            break;
1018    
1019            case PT_LAMP:            case PT_LAMP:
1020            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1021                   prop->chartype == ucp_Lt;
1022            break;            break;
1023    
1024            case PT_GC:            case PT_GC:
1025            OK = category == code[2];            OK = _pcre_ucp_gentype[prop->chartype] == code[2];
1026            break;            break;
1027    
1028            case PT_PC:            case PT_PC:
1029            OK = chartype == code[2];            OK = prop->chartype == code[2];
1030            break;            break;
1031    
1032            case PT_SC:            case PT_SC:
1033            OK = script == code[2];            OK = prop->script == code[2];
1034              break;
1035    
1036              /* These are specials for combination cases. */
1037    
1038              case PT_ALNUM:
1039              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1040                   _pcre_ucp_gentype[prop->chartype] == ucp_N;
1041              break;
1042    
1043              case PT_SPACE:    /* Perl space */
1044              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1045                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1046              break;
1047    
1048              case PT_PXSPACE:  /* POSIX space */
1049              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1050                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1051                   c == CHAR_FF || c == CHAR_CR;
1052              break;
1053    
1054              case PT_WORD:
1055              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1056                   _pcre_ucp_gentype[prop->chartype] == ucp_N ||
1057                   c == CHAR_UNDERSCORE;
1058            break;            break;
1059    
1060            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
# Line 850  for (;;) Line 1074  for (;;)
1074  /* ========================================================================== */  /* ========================================================================== */
1075        /* These opcodes likewise inspect the subject character, but have an        /* These opcodes likewise inspect the subject character, but have an
1076        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:
1077        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,
1078        OP_NOT_WORDCHAR. The value is loaded into d. */        OP_WORDCHAR, OP_NOT_WORDCHAR. The value is loaded into d. */
1079    
1080        case OP_TYPEPLUS:        case OP_TYPEPLUS:
1081        case OP_TYPEMINPLUS:        case OP_TYPEMINPLUS:
# Line 862  for (;;) Line 1086  for (;;)
1086          {          {
1087          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1088              (c < 256 &&              (c < 256 &&
1089                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1090                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1091            {            {
1092            if (count > 0 && codevalue == OP_TYPEPOSPLUS)            if (count > 0 && codevalue == OP_TYPEPOSPLUS)
# Line 888  for (;;) Line 1109  for (;;)
1109          {          {
1110          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1111              (c < 256 &&              (c < 256 &&
1112                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1113                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1114            {            {
1115            if (codevalue == OP_TYPEPOSQUERY)            if (codevalue == OP_TYPEPOSQUERY)
# Line 913  for (;;) Line 1131  for (;;)
1131          {          {
1132          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1133              (c < 256 &&              (c < 256 &&
1134                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1135                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1136            {            {
1137            if (codevalue == OP_TYPEPOSSTAR)            if (codevalue == OP_TYPEPOSSTAR)
# Line 936  for (;;) Line 1151  for (;;)
1151          {          {
1152          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1153              (c < 256 &&              (c < 256 &&
1154                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1155                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1156            {            {
1157            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
# Line 960  for (;;) Line 1172  for (;;)
1172          {          {
1173          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1174              (c < 256 &&              (c < 256 &&
1175                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1176                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1177            {            {
1178            if (codevalue == OP_TYPEPOSUPTO)            if (codevalue == OP_TYPEPOSUPTO)
# Line 994  for (;;) Line 1203  for (;;)
1203        if (clen > 0)        if (clen > 0)
1204          {          {
1205          BOOL OK;          BOOL OK;
1206          int category = _pcre_ucp_findprop(c, &chartype, &script);          const ucd_record * prop = GET_UCD(c);
1207          switch(code[2])          switch(code[2])
1208            {            {
1209            case PT_ANY:            case PT_ANY:
# Line 1002  for (;;) Line 1211  for (;;)
1211            break;            break;
1212    
1213            case PT_LAMP:            case PT_LAMP:
1214            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1215                prop->chartype == ucp_Lt;
1216            break;            break;
1217    
1218            case PT_GC:            case PT_GC:
1219            OK = category == code[3];            OK = _pcre_ucp_gentype[prop->chartype] == code[3];
1220            break;            break;
1221    
1222            case PT_PC:            case PT_PC:
1223            OK = chartype == code[3];            OK = prop->chartype == code[3];
1224            break;            break;
1225    
1226            case PT_SC:            case PT_SC:
1227            OK = script == code[3];            OK = prop->script == code[3];
1228              break;
1229    
1230              /* These are specials for combination cases. */
1231    
1232              case PT_ALNUM:
1233              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1234                   _pcre_ucp_gentype[prop->chartype] == ucp_N;
1235              break;
1236    
1237              case PT_SPACE:    /* Perl space */
1238              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1239                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1240              break;
1241    
1242              case PT_PXSPACE:  /* POSIX space */
1243              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1244                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1245                   c == CHAR_FF || c == CHAR_CR;
1246              break;
1247    
1248              case PT_WORD:
1249              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1250                   _pcre_ucp_gentype[prop->chartype] == ucp_N ||
1251                   c == CHAR_UNDERSCORE;
1252            break;            break;
1253    
1254            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
# Line 1043  for (;;) Line 1277  for (;;)
1277        case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS:        case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS:
1278        count = current_state->count;  /* Already matched */        count = current_state->count;  /* Already matched */
1279        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1280        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)
1281          {          {
1282          const uschar *nptr = ptr + clen;          const uschar *nptr = ptr + clen;
1283          int ncount = 0;          int ncount = 0;
# Line 1057  for (;;) Line 1291  for (;;)
1291            int nd;            int nd;
1292            int ndlen = 1;            int ndlen = 1;
1293            GETCHARLEN(nd, nptr, ndlen);            GETCHARLEN(nd, nptr, ndlen);
1294            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if (UCD_CATEGORY(nd) != ucp_M) break;
1295            ncount++;            ncount++;
1296            nptr += ndlen;            nptr += ndlen;
1297            }            }
# Line 1078  for (;;) Line 1312  for (;;)
1312          int ncount = 0;          int ncount = 0;
1313          switch (c)          switch (c)
1314            {            {
           case 0x000d:  
           if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;  
           /* Fall through */  
           case 0x000a:  
1315            case 0x000b:            case 0x000b:
1316            case 0x000c:            case 0x000c:
1317            case 0x0085:            case 0x0085:
1318            case 0x2028:            case 0x2028:
1319            case 0x2029:            case 0x2029:
1320              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1321              goto ANYNL01;
1322    
1323              case 0x000d:
1324              if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;
1325              /* Fall through */
1326    
1327              ANYNL01:
1328              case 0x000a:
1329            if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS)            if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS)
1330              {              {
1331              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
# Line 1095  for (;;) Line 1334  for (;;)
1334            count++;            count++;
1335            ADD_NEW_DATA(-state_offset, count, ncount);            ADD_NEW_DATA(-state_offset, count, ncount);
1336            break;            break;
1337    
1338            default:            default:
1339            break;            break;
1340            }            }
# Line 1210  for (;;) Line 1450  for (;;)
1450        if (clen > 0)        if (clen > 0)
1451          {          {
1452          BOOL OK;          BOOL OK;
1453          int category = _pcre_ucp_findprop(c, &chartype, &script);          const ucd_record * prop = GET_UCD(c);
1454          switch(code[2])          switch(code[2])
1455            {            {
1456            case PT_ANY:            case PT_ANY:
# Line 1218  for (;;) Line 1458  for (;;)
1458            break;            break;
1459    
1460            case PT_LAMP:            case PT_LAMP:
1461            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1462                prop->chartype == ucp_Lt;
1463            break;            break;
1464    
1465            case PT_GC:            case PT_GC:
1466            OK = category == code[3];            OK = _pcre_ucp_gentype[prop->chartype] == code[3];
1467            break;            break;
1468    
1469            case PT_PC:            case PT_PC:
1470            OK = chartype == code[3];            OK = prop->chartype == code[3];
1471            break;            break;
1472    
1473            case PT_SC:            case PT_SC:
1474            OK = script == code[3];            OK = prop->script == code[3];
1475              break;
1476    
1477              /* These are specials for combination cases. */
1478    
1479              case PT_ALNUM:
1480              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1481                   _pcre_ucp_gentype[prop->chartype] == ucp_N;
1482              break;
1483    
1484              case PT_SPACE:    /* Perl space */
1485              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1486                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1487              break;
1488    
1489              case PT_PXSPACE:  /* POSIX space */
1490              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1491                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1492                   c == CHAR_FF || c == CHAR_CR;
1493              break;
1494    
1495              case PT_WORD:
1496              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1497                   _pcre_ucp_gentype[prop->chartype] == ucp_N ||
1498                   c == CHAR_UNDERSCORE;
1499            break;            break;
1500    
1501            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
# Line 1268  for (;;) Line 1533  for (;;)
1533        QS2:        QS2:
1534    
1535        ADD_ACTIVE(state_offset + 2, 0);        ADD_ACTIVE(state_offset + 2, 0);
1536        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)
1537          {          {
1538          const uschar *nptr = ptr + clen;          const uschar *nptr = ptr + clen;
1539          int ncount = 0;          int ncount = 0;
# Line 1283  for (;;) Line 1548  for (;;)
1548            int nd;            int nd;
1549            int ndlen = 1;            int ndlen = 1;
1550            GETCHARLEN(nd, nptr, ndlen);            GETCHARLEN(nd, nptr, ndlen);
1551            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if (UCD_CATEGORY(nd) != ucp_M) break;
1552            ncount++;            ncount++;
1553            nptr += ndlen;            nptr += ndlen;
1554            }            }
# Line 1311  for (;;) Line 1576  for (;;)
1576          int ncount = 0;          int ncount = 0;
1577          switch (c)          switch (c)
1578            {            {
           case 0x000d:  
           if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;  
           /* Fall through */  
           case 0x000a:  
1579            case 0x000b:            case 0x000b:
1580            case 0x000c:            case 0x000c:
1581            case 0x0085:            case 0x0085:
1582            case 0x2028:            case 0x2028:
1583            case 0x2029:            case 0x2029:
1584              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1585              goto ANYNL02;
1586    
1587              case 0x000d:
1588              if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;
1589              /* Fall through */
1590    
1591              ANYNL02:
1592              case 0x000a:
1593            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR ||            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR ||
1594                codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY)                codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY)
1595              {              {
# Line 1328  for (;;) Line 1598  for (;;)
1598              }              }
1599            ADD_NEW_DATA(-(state_offset + count), 0, ncount);            ADD_NEW_DATA(-(state_offset + count), 0, ncount);
1600            break;            break;
1601    
1602            default:            default:
1603            break;            break;
1604            }            }
# Line 1451  for (;;) Line 1722  for (;;)
1722        if (clen > 0)        if (clen > 0)
1723          {          {
1724          BOOL OK;          BOOL OK;
1725          int category = _pcre_ucp_findprop(c, &chartype, &script);          const ucd_record * prop = GET_UCD(c);
1726          switch(code[4])          switch(code[4])
1727            {            {
1728            case PT_ANY:            case PT_ANY:
# Line 1459  for (;;) Line 1730  for (;;)
1730            break;            break;
1731    
1732            case PT_LAMP:            case PT_LAMP:
1733            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1734                prop->chartype == ucp_Lt;
1735            break;            break;
1736    
1737            case PT_GC:            case PT_GC:
1738            OK = category == code[5];            OK = _pcre_ucp_gentype[prop->chartype] == code[5];
1739            break;            break;
1740    
1741            case PT_PC:            case PT_PC:
1742            OK = chartype == code[5];            OK = prop->chartype == code[5];
1743            break;            break;
1744    
1745            case PT_SC:            case PT_SC:
1746            OK = script == code[5];            OK = prop->script == code[5];
1747              break;
1748    
1749              /* These are specials for combination cases. */
1750    
1751              case PT_ALNUM:
1752              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1753                   _pcre_ucp_gentype[prop->chartype] == ucp_N;
1754              break;
1755    
1756              case PT_SPACE:    /* Perl space */
1757              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1758                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1759              break;
1760    
1761              case PT_PXSPACE:  /* POSIX space */
1762              OK = _pcre_ucp_gentype[prop->chartype] == ucp_Z ||
1763                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1764                   c == CHAR_FF || c == CHAR_CR;
1765              break;
1766    
1767              case PT_WORD:
1768              OK = _pcre_ucp_gentype[prop->chartype] == ucp_L ||
1769                   _pcre_ucp_gentype[prop->chartype] == ucp_N ||
1770                   c == CHAR_UNDERSCORE;
1771            break;            break;
1772    
1773            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
# Line 1504  for (;;) Line 1800  for (;;)
1800        if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT)        if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT)
1801          { ADD_ACTIVE(state_offset + 4, 0); }          { ADD_ACTIVE(state_offset + 4, 0); }
1802        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1803        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)
1804          {          {
1805          const uschar *nptr = ptr + clen;          const uschar *nptr = ptr + clen;
1806          int ncount = 0;          int ncount = 0;
# Line 1518  for (;;) Line 1814  for (;;)
1814            int nd;            int nd;
1815            int ndlen = 1;            int ndlen = 1;
1816            GETCHARLEN(nd, nptr, ndlen);            GETCHARLEN(nd, nptr, ndlen);
1817            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if (UCD_CATEGORY(nd) != ucp_M) break;
1818            ncount++;            ncount++;
1819            nptr += ndlen;            nptr += ndlen;
1820            }            }
# Line 1543  for (;;) Line 1839  for (;;)
1839          int ncount = 0;          int ncount = 0;
1840          switch (c)          switch (c)
1841            {            {
           case 0x000d:  
           if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;  
           /* Fall through */  
           case 0x000a:  
1842            case 0x000b:            case 0x000b:
1843            case 0x000c:            case 0x000c:
1844            case 0x0085:            case 0x0085:
1845            case 0x2028:            case 0x2028:
1846            case 0x2029:            case 0x2029:
1847              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1848              goto ANYNL03;
1849    
1850              case 0x000d:
1851              if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;
1852              /* Fall through */
1853    
1854              ANYNL03:
1855              case 0x000a:
1856            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO)            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO)
1857              {              {
1858              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
# Line 1562  for (;;) Line 1863  for (;;)
1863            else            else
1864              { ADD_NEW_DATA(-state_offset, count, ncount); }              { ADD_NEW_DATA(-state_offset, count, ncount); }
1865            break;            break;
1866    
1867            default:            default:
1868            break;            break;
1869            }            }
# Line 1677  for (;;) Line 1979  for (;;)
1979        break;        break;
1980    
1981        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
1982        case OP_CHARNC:        case OP_CHARI:
1983        if (clen == 0) break;        if (clen == 0) break;
1984    
1985  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1692  for (;;) Line 1994  for (;;)
1994            other case of the character. */            other case of the character. */
1995    
1996  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1997            othercase = _pcre_ucp_othercase(c);            othercase = UCD_OTHERCASE(c);
1998  #else  #else
1999            othercase = NOTACHAR;            othercase = NOTACHAR;
2000  #endif  #endif
# Line 1717  for (;;) Line 2019  for (;;)
2019        to wait for them to pass before continuing. */        to wait for them to pass before continuing. */
2020    
2021        case OP_EXTUNI:        case OP_EXTUNI:
2022        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)
2023          {          {
2024          const uschar *nptr = ptr + clen;          const uschar *nptr = ptr + clen;
2025          int ncount = 0;          int ncount = 0;
# Line 1725  for (;;) Line 2027  for (;;)
2027            {            {
2028            int nclen = 1;            int nclen = 1;
2029            GETCHARLEN(c, nptr, nclen);            GETCHARLEN(c, nptr, nclen);
2030            if (_pcre_ucp_findprop(c, &chartype, &script) != ucp_M) break;            if (UCD_CATEGORY(c) != ucp_M) break;
2031            ncount++;            ncount++;
2032            nptr += nclen;            nptr += nclen;
2033            }            }
# Line 1742  for (;;) Line 2044  for (;;)
2044        case OP_ANYNL:        case OP_ANYNL:
2045        if (clen > 0) switch(c)        if (clen > 0) switch(c)
2046          {          {
         case 0x000a:  
2047          case 0x000b:          case 0x000b:
2048          case 0x000c:          case 0x000c:
2049          case 0x0085:          case 0x0085:
2050          case 0x2028:          case 0x2028:
2051          case 0x2029:          case 0x2029:
2052            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
2053    
2054            case 0x000a:
2055          ADD_NEW(state_offset + 1, 0);          ADD_NEW(state_offset + 1, 0);
2056          break;          break;
2057    
2058          case 0x000d:          case 0x000d:
2059          if (ptr + 1 < end_subject && ptr[1] == 0x0a)          if (ptr + 1 < end_subject && ptr[1] == 0x0a)
2060            {            {
# Line 1860  for (;;) Line 2165  for (;;)
2165        break;        break;
2166    
2167        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2168        /* Match a negated single character. This is only used for one-byte        /* Match a negated single character casefully. This is only used for
2169        characters, that is, we know that d < 256. The character we are        one-byte characters, that is, we know that d < 256. The character we are
2170        checking (c) can be multibyte. */        checking (c) can be multibyte. */
2171    
2172        case OP_NOT:        case OP_NOT:
2173        if (clen > 0)        if (clen > 0 && c != d) { ADD_NEW(state_offset + dlen + 1, 0); }
         {  
         unsigned int otherd = ((ims & PCRE_CASELESS) != 0)? fcc[d] : d;  
         if (c != d && c != otherd) { ADD_NEW(state_offset + dlen + 1, 0); }  
         }  
2174        break;        break;
2175    
2176        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2177          /* Match a negated single character caselessly. This is only used for
2178          one-byte characters, that is, we know that d < 256. The character we are
2179          checking (c) can be multibyte. */
2180    
2181          case OP_NOTI:
2182          if (clen > 0 && c != d && c != fcc[d])
2183            { ADD_NEW(state_offset + dlen + 1, 0); }
2184          break;
2185    
2186          /*-----------------------------------------------------------------*/
2187          case OP_PLUSI:
2188          case OP_MINPLUSI:
2189          case OP_POSPLUSI:
2190          case OP_NOTPLUSI:
2191          case OP_NOTMINPLUSI:
2192          case OP_NOTPOSPLUSI:
2193          caseless = TRUE;
2194          codevalue -= OP_STARI - OP_STAR;
2195    
2196          /* Fall through */
2197        case OP_PLUS:        case OP_PLUS:
2198        case OP_MINPLUS:        case OP_MINPLUS:
2199        case OP_POSPLUS:        case OP_POSPLUS:
# Line 1884  for (;;) Line 2205  for (;;)
2205        if (clen > 0)        if (clen > 0)
2206          {          {
2207          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2208          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2209            {            {
2210  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2211            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2212              {              {
2213  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2214              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2215  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2216              }              }
2217            else            else
# Line 1912  for (;;) Line 2233  for (;;)
2233        break;        break;
2234    
2235        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2236          case OP_QUERYI:
2237          case OP_MINQUERYI:
2238          case OP_POSQUERYI:
2239          case OP_NOTQUERYI:
2240          case OP_NOTMINQUERYI:
2241          case OP_NOTPOSQUERYI:
2242          caseless = TRUE;
2243          codevalue -= OP_STARI - OP_STAR;
2244          /* Fall through */
2245        case OP_QUERY:        case OP_QUERY:
2246        case OP_MINQUERY:        case OP_MINQUERY:
2247        case OP_POSQUERY:        case OP_POSQUERY:
# Line 1922  for (;;) Line 2252  for (;;)
2252        if (clen > 0)        if (clen > 0)
2253          {          {
2254          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2255          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2256            {            {
2257  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2258            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2259              {              {
2260  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2261              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2262  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2263              }              }
2264            else            else
# Line 1948  for (;;) Line 2278  for (;;)
2278        break;        break;
2279    
2280        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2281          case OP_STARI:
2282          case OP_MINSTARI:
2283          case OP_POSSTARI:
2284          case OP_NOTSTARI:
2285          case OP_NOTMINSTARI:
2286          case OP_NOTPOSSTARI:
2287          caseless = TRUE;
2288          codevalue -= OP_STARI - OP_STAR;
2289          /* Fall through */
2290        case OP_STAR:        case OP_STAR:
2291        case OP_MINSTAR:        case OP_MINSTAR:
2292        case OP_POSSTAR:        case OP_POSSTAR:
# Line 1958  for (;;) Line 2297  for (;;)
2297        if (clen > 0)        if (clen > 0)
2298          {          {
2299          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2300          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2301            {            {
2302  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2303            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2304              {              {
2305  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2306              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2307  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2308              }              }
2309            else            else
# Line 1984  for (;;) Line 2323  for (;;)
2323        break;        break;
2324    
2325        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2326          case OP_EXACTI:
2327          case OP_NOTEXACTI:
2328          caseless = TRUE;
2329          codevalue -= OP_STARI - OP_STAR;
2330          /* Fall through */
2331        case OP_EXACT:        case OP_EXACT:
2332        case OP_NOTEXACT:        case OP_NOTEXACT:
2333        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
2334        if (clen > 0)        if (clen > 0)
2335          {          {
2336          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2337          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2338            {            {
2339  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2340            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2341              {              {
2342  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2343              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2344  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2345              }              }
2346            else            else
# Line 2014  for (;;) Line 2358  for (;;)
2358        break;        break;
2359    
2360        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2361          case OP_UPTOI:
2362          case OP_MINUPTOI:
2363          case OP_POSUPTOI:
2364          case OP_NOTUPTOI:
2365          case OP_NOTMINUPTOI:
2366          case OP_NOTPOSUPTOI:
2367          caseless = TRUE;
2368          codevalue -= OP_STARI - OP_STAR;
2369          /* Fall through */
2370        case OP_UPTO:        case OP_UPTO:
2371        case OP_MINUPTO:        case OP_MINUPTO:
2372        case OP_POSUPTO:        case OP_POSUPTO:
# Line 2025  for (;;) Line 2378  for (;;)
2378        if (clen > 0)        if (clen > 0)
2379          {          {
2380          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2381          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2382            {            {
2383  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2384            if (utf8 && d >= 128)            if (utf8 && d >= 128)
2385              {              {
2386  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2387              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2388  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2389              }              }
2390            else            else
# Line 2092  for (;;) Line 2445  for (;;)
2445          points to the byte after the end of the class. If there is a          points to the byte after the end of the class. If there is a
2446          quantifier, this is where it will be. */          quantifier, this is where it will be. */
2447    
2448          next_state_offset = ecode - start_code;          next_state_offset = (int)(ecode - start_code);
2449    
2450          switch (*ecode)          switch (*ecode)
2451            {            {
# Line 2139  for (;;) Line 2492  for (;;)
2492    
2493  /* ========================================================================== */  /* ========================================================================== */
2494        /* These are the opcodes for fancy brackets of various kinds. We have        /* These are the opcodes for fancy brackets of various kinds. We have
2495        to use recursion in order to handle them. */        to use recursion in order to handle them. The "always failing" assertion
2496          (?!) is optimised to OP_FAIL when compiling, so we have to support that,
2497          though the other "backtracking verbs" are not supported. */
2498    
2499          case OP_FAIL:
2500          forced_fail++;    /* Count FAILs for multiple states */
2501          break;
2502    
2503        case OP_ASSERT:        case OP_ASSERT:
2504        case OP_ASSERT_NOT:        case OP_ASSERT_NOT:
# Line 2157  for (;;) Line 2516  for (;;)
2516            md,                                   /* static match data */            md,                                   /* static match data */
2517            code,                                 /* this subexpression's code */            code,                                 /* this subexpression's code */
2518            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2519            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2520            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2521            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2522            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2523            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
           ims,                                  /* the current ims flags */  
2524            rlevel,                               /* function recursion level */            rlevel,                               /* function recursion level */
2525            recursing);                           /* pass on regex recursion */            recursing);                           /* pass on regex recursion */
2526    
2527            if (rc == PCRE_ERROR_DFA_UITEM) return rc;
2528          if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK))          if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK))
2529              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }              { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); }
2530          }          }
2531        break;        break;
2532    
# Line 2177  for (;;) Line 2536  for (;;)
2536          {          {
2537          int local_offsets[1000];          int local_offsets[1000];
2538          int local_workspace[1000];          int local_workspace[1000];
2539          int condcode = code[LINK_SIZE+1];          int codelink = GET(code, 1);
2540            int condcode;
2541    
2542            /* Because of the way auto-callout works during compile, a callout item
2543            is inserted between OP_COND and an assertion condition. This does not
2544            happen for the other conditions. */
2545    
2546            if (code[LINK_SIZE+1] == OP_CALLOUT)
2547              {
2548              rrc = 0;
2549              if (pcre_callout != NULL)
2550                {
2551                pcre_callout_block cb;
2552                cb.version          = 1;   /* Version 1 of the callout block */
2553                cb.callout_number   = code[LINK_SIZE+2];
2554                cb.offset_vector    = offsets;
2555                cb.subject          = (PCRE_SPTR)start_subject;
2556                cb.subject_length   = (int)(end_subject - start_subject);
2557                cb.start_match      = (int)(current_subject - start_subject);
2558                cb.current_position = (int)(ptr - start_subject);
2559                cb.pattern_position = GET(code, LINK_SIZE + 3);
2560                cb.next_item_length = GET(code, 3 + 2*LINK_SIZE);
2561                cb.capture_top      = 1;
2562                cb.capture_last     = -1;
2563                cb.callout_data     = md->callout_data;
2564                if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc;   /* Abandon */
2565                }
2566              if (rrc > 0) break;                      /* Fail this thread */
2567              code += _pcre_OP_lengths[OP_CALLOUT];    /* Skip callout data */
2568              }
2569    
2570            condcode = code[LINK_SIZE+1];
2571    
2572          /* Back reference conditions are not supported */          /* Back reference conditions are not supported */
2573    
2574          if (condcode == OP_CREF) return PCRE_ERROR_DFA_UCOND;          if (condcode == OP_CREF || condcode == OP_NCREF)
2575              return PCRE_ERROR_DFA_UCOND;
2576    
2577          /* The DEFINE condition is always false */          /* The DEFINE condition is always false */
2578    
2579          if (condcode == OP_DEF)          if (condcode == OP_DEF)
2580            {            { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
           ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0);  
           }  
2581    
2582          /* 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,
2583          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
2584          recursed groups. */          recursed groups. */
2585    
2586          else if (condcode == OP_RREF)          else if (condcode == OP_RREF || condcode == OP_NRREF)
2587            {            {
2588            int value = GET2(code, LINK_SIZE+2);            int value = GET2(code, LINK_SIZE+2);
2589            if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND;            if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND;
2590            if (recursing > 0) { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); }            if (recursing > 0)
2591              else { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); }              { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); }
2592              else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
2593            }            }
2594    
2595          /* Otherwise, the condition is an assertion */          /* Otherwise, the condition is an assertion */
# Line 2216  for (;;) Line 2606  for (;;)
2606              md,                                   /* fixed match data */              md,                                   /* fixed match data */
2607              asscode,                              /* this subexpression's code */              asscode,                              /* this subexpression's code */
2608              ptr,                                  /* where we currently are */              ptr,                                  /* where we currently are */
2609              ptr - start_subject,                  /* start offset */              (int)(ptr - start_subject),           /* start offset */
2610              local_offsets,                        /* offset vector */              local_offsets,                        /* offset vector */
2611              sizeof(local_offsets)/sizeof(int),    /* size of same */              sizeof(local_offsets)/sizeof(int),    /* size of same */
2612              local_workspace,                      /* workspace vector */              local_workspace,                      /* workspace vector */
2613              sizeof(local_workspace)/sizeof(int),  /* size of same */              sizeof(local_workspace)/sizeof(int),  /* size of same */
             ims,                                  /* the current ims flags */  
2614              rlevel,                               /* function recursion level */              rlevel,                               /* function recursion level */
2615              recursing);                           /* pass on regex recursion */              recursing);                           /* pass on regex recursion */
2616    
2617              if (rc == PCRE_ERROR_DFA_UITEM) return rc;
2618            if ((rc >= 0) ==            if ((rc >= 0) ==
2619                  (condcode == OP_ASSERT || condcode == OP_ASSERTBACK))                  (condcode == OP_ASSERT || condcode == OP_ASSERTBACK))
2620              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }              { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); }
2621            else            else
2622              { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); }              { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
2623            }            }
2624          }          }
2625        break;        break;
# Line 2248  for (;;) Line 2638  for (;;)
2638            md,                                   /* fixed match data */            md,                                   /* fixed match data */
2639            start_code + GET(code, 1),            /* this subexpression's code */            start_code + GET(code, 1),            /* this subexpression's code */
2640            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2641            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2642            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2643            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2644            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2645            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
           ims,                                  /* the current ims flags */  
2646            rlevel,                               /* function recursion level */            rlevel,                               /* function recursion level */
2647            recursing + 1);                       /* regex recurse level */            recursing + 1);                       /* regex recurse level */
2648    
# Line 2291  for (;;) Line 2680  for (;;)
2680        break;        break;
2681    
2682        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2683          case OP_BRAPOS:
2684          case OP_SBRAPOS:
2685          case OP_CBRAPOS:
2686          case OP_SCBRAPOS:
2687          case OP_BRAPOSZERO:
2688            {
2689            int charcount, matched_count;
2690            const uschar *local_ptr = ptr;
2691            BOOL allow_zero;
2692    
2693            if (codevalue == OP_BRAPOSZERO)
2694              {
2695              allow_zero = TRUE;
2696              codevalue = *(++code);  /* Codevalue will be one of above BRAs */
2697              }
2698            else allow_zero = FALSE;
2699    
2700            /* Loop to match the subpattern as many times as possible as if it were
2701            a complete pattern. */
2702    
2703            for (matched_count = 0;; matched_count++)
2704              {
2705              int local_offsets[2];
2706              int local_workspace[1000];
2707    
2708              int rc = internal_dfa_exec(
2709                md,                                   /* fixed match data */
2710                code,                                 /* this subexpression's code */
2711                local_ptr,                            /* where we currently are */
2712                (int)(ptr - start_subject),           /* start offset */
2713                local_offsets,                        /* offset vector */
2714                sizeof(local_offsets)/sizeof(int),    /* size of same */
2715                local_workspace,                      /* workspace vector */
2716                sizeof(local_workspace)/sizeof(int),  /* size of same */
2717                rlevel,                               /* function recursion level */
2718                recursing);                           /* pass on regex recursion */
2719    
2720              /* Failed to match */
2721    
2722              if (rc < 0)
2723                {
2724                if (rc != PCRE_ERROR_NOMATCH) return rc;
2725                break;
2726                }
2727    
2728              /* Matched: break the loop if zero characters matched. */
2729    
2730              charcount = local_offsets[1] - local_offsets[0];
2731              if (charcount == 0) break;
2732              local_ptr += charcount;    /* Advance temporary position ptr */
2733              }
2734    
2735            /* At this point we have matched the subpattern matched_count
2736            times, and local_ptr is pointing to the character after the end of the
2737            last match. */
2738    
2739            if (matched_count > 0 || allow_zero)
2740              {
2741              const uschar *end_subpattern = code;
2742              int next_state_offset;
2743    
2744              do { end_subpattern += GET(end_subpattern, 1); }
2745                while (*end_subpattern == OP_ALT);
2746              next_state_offset =
2747                (int)(end_subpattern - start_code + LINK_SIZE + 1);
2748    
2749              /* Optimization: if there are no more active states, and there
2750              are no new states yet set up, then skip over the subject string
2751              right here, to save looping. Otherwise, set up the new state to swing
2752              into action when the end of the matched substring is reached. */
2753    
2754              if (i + 1 >= active_count && new_count == 0)
2755                {
2756                ptr = local_ptr;
2757                clen = 0;
2758                ADD_NEW(next_state_offset, 0);
2759                }
2760              else
2761                {
2762                const uschar *p = ptr;
2763                const uschar *pp = local_ptr;
2764                charcount = pp - p;
2765                while (p < pp) if ((*p++ & 0xc0) == 0x80) charcount--;
2766                ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1));
2767                }
2768              }
2769            }
2770          break;
2771    
2772          /*-----------------------------------------------------------------*/
2773        case OP_ONCE:        case OP_ONCE:
2774          {          {
2775          int local_offsets[2];          int local_offsets[2];
# Line 2300  for (;;) Line 2779  for (;;)
2779            md,                                   /* fixed match data */            md,                                   /* fixed match data */
2780            code,                                 /* this subexpression's code */            code,                                 /* this subexpression's code */
2781            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2782            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2783            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2784            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2785            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2786            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
           ims,                                  /* the current ims flags */  
2787            rlevel,                               /* function recursion level */            rlevel,                               /* function recursion level */
2788            recursing);                           /* pass on regex recursion */            recursing);                           /* pass on regex recursion */
2789    
# Line 2317  for (;;) Line 2795  for (;;)
2795    
2796            do { end_subpattern += GET(end_subpattern, 1); }            do { end_subpattern += GET(end_subpattern, 1); }
2797              while (*end_subpattern == OP_ALT);              while (*end_subpattern == OP_ALT);
2798            next_state_offset = end_subpattern - start_code + LINK_SIZE + 1;            next_state_offset =
2799                (int)(end_subpattern - start_code + LINK_SIZE + 1);
2800    
2801            /* If the end of this subpattern is KETRMAX or KETRMIN, we must            /* If the end of this subpattern is KETRMAX or KETRMIN, we must
2802            arrange for the repeat state also to be added to the relevant list.            arrange for the repeat state also to be added to the relevant list.
# Line 2325  for (;;) Line 2804  for (;;)
2804    
2805            repeat_state_offset = (*end_subpattern == OP_KETRMAX ||            repeat_state_offset = (*end_subpattern == OP_KETRMAX ||
2806                                   *end_subpattern == OP_KETRMIN)?                                   *end_subpattern == OP_KETRMIN)?
2807              end_subpattern - start_code - GET(end_subpattern, 1) : -1;              (int)(end_subpattern - start_code - GET(end_subpattern, 1)) : -1;
2808    
2809            /* If we have matched an empty string, add the next state at the            /* If we have matched an empty string, add the next state at the
2810            current character pointer. This is important so that the duplicate            current character pointer. This is important so that the duplicate
# Line 2340  for (;;) Line 2819  for (;;)
2819            /* Optimization: if there are no more active states, and there            /* Optimization: if there are no more active states, and there
2820            are no new states yet set up, then skip over the subject string            are no new states yet set up, then skip over the subject string
2821            right here, to save looping. Otherwise, set up the new state to swing            right here, to save looping. Otherwise, set up the new state to swing
2822            into action when the end of the substring is reached. */            into action when the end of the matched substring is reached. */
2823    
2824            else if (i + 1 >= active_count && new_count == 0)            else if (i + 1 >= active_count && new_count == 0)
2825              {              {
# Line 2370  for (;;) Line 2849  for (;;)
2849              if (repeat_state_offset >= 0)              if (repeat_state_offset >= 0)
2850                { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); }                { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); }
2851              }              }
   
2852            }            }
2853          else if (rc != PCRE_ERROR_NOMATCH) return rc;          else if (rc != PCRE_ERROR_NOMATCH) return rc;
2854          }          }
# Line 2381  for (;;) Line 2859  for (;;)
2859        /* Handle callouts */        /* Handle callouts */
2860    
2861        case OP_CALLOUT:        case OP_CALLOUT:
2862          rrc = 0;
2863        if (pcre_callout != NULL)        if (pcre_callout != NULL)
2864          {          {
         int rrc;  
2865          pcre_callout_block cb;          pcre_callout_block cb;
2866          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 1;   /* Version 1 of the callout block */
2867          cb.callout_number   = code[1];          cb.callout_number   = code[1];
2868          cb.offset_vector    = offsets;          cb.offset_vector    = offsets;
2869          cb.subject          = (PCRE_SPTR)start_subject;          cb.subject          = (PCRE_SPTR)start_subject;
2870          cb.subject_length   = end_subject - start_subject;          cb.subject_length   = (int)(end_subject - start_subject);
2871          cb.start_match      = current_subject - start_subject;          cb.start_match      = (int)(current_subject - start_subject);
2872          cb.current_position = ptr - start_subject;          cb.current_position = (int)(ptr - start_subject);
2873          cb.pattern_position = GET(code, 2);          cb.pattern_position = GET(code, 2);
2874          cb.next_item_length = GET(code, 2 + LINK_SIZE);          cb.next_item_length = GET(code, 2 + LINK_SIZE);
2875          cb.capture_top      = 1;          cb.capture_top      = 1;
2876          cb.capture_last     = -1;          cb.capture_last     = -1;
2877          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
2878          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); }  
2879          }          }
2880          if (rrc == 0)
2881            { ADD_ACTIVE(state_offset + _pcre_OP_lengths[OP_CALLOUT], 0); }
2882        break;        break;
2883    
2884    
# Line 2415  for (;;) Line 2894  for (;;)
2894    /* We have finished the processing at the current subject character. If no    /* We have finished the processing at the current subject character. If no
2895    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
2896    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
2897    matching has been requested, check for appropriate conditions. */    matching has been requested, check for appropriate conditions.
2898    
2899      The "forced_ fail" variable counts the number of (*F) encountered for the
2900      character. If it is equal to the original active_count (saved in
2901      workspace[1]) it means that (*F) was found on every active state. In this
2902      case we don't want to give a partial match.
2903    
2904      The "could_continue" variable is true if a state could have continued but
2905      for the fact that the end of the subject was reached. */
2906    
2907    if (new_count <= 0)    if (new_count <= 0)
2908      {      {
2909      if (match_count < 0 &&                     /* No matches found */      if (rlevel == 1 &&                               /* Top level, and */
2910          rlevel == 1 &&                         /* Top level match function */          could_continue &&                            /* Some could go on */
2911          (md->moptions & PCRE_PARTIAL) != 0 &&  /* Want partial matching */          forced_fail != workspace[1] &&               /* Not all forced fail & */
2912            (                                            /* either... */
2913            (md->moptions & PCRE_PARTIAL_HARD) != 0      /* Hard partial */
2914            ||                                           /* or... */
2915            ((md->moptions & PCRE_PARTIAL_SOFT) != 0 &&  /* Soft partial and */
2916             match_count < 0)                            /* no matches */
2917            ) &&                                         /* And... */
2918          ptr >= end_subject &&                  /* Reached end of subject */          ptr >= end_subject &&                  /* Reached end of subject */
2919          ptr > current_subject)                 /* Matched non-empty string */          ptr > md->start_used_ptr)              /* Inspected non-empty string */
2920        {        {
2921        if (offsetcount >= 2)        if (offsetcount >= 2)
2922          {          {
2923          offsets[0] = current_subject - start_subject;          offsets[0] = (int)(md->start_used_ptr - start_subject);
2924          offsets[1] = end_subject - start_subject;          offsets[1] = (int)(end_subject - start_subject);
2925          }          }
2926        match_count = PCRE_ERROR_PARTIAL;        match_count = PCRE_ERROR_PARTIAL;
2927        }        }
# Line 2482  Returns:          > 0 => number of match Line 2975  Returns:          > 0 => number of match
2975                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
2976  */  */
2977    
2978  PCRE_EXP_DEFN int  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
2979  pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data,
2980    const char *subject, int length, int start_offset, int options, int *offsets,    const char *subject, int length, int start_offset, int options, int *offsets,
2981    int offsetcount, int *workspace, int wscount)    int offsetcount, int *workspace, int wscount)
# Line 2513  if (re == NULL || subject == NULL || wor Line 3006  if (re == NULL || subject == NULL || wor
3006     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;
3007  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
3008  if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE;  if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE;
3009    if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
3010    
3011  /* We need to find the pointer to any study data before we test for byte  /* We need to find the pointer to any study data before we test for byte
3012  flipping, so we scan the extra_data block first. This may set two fields in the  flipping, so we scan the extra_data block first. This may set two fields in the
# Line 2569  md->start_code = (const uschar *)argumen Line 3063  md->start_code = (const uschar *)argumen
3063      re->name_table_offset + re->name_count * re->name_entry_size;      re->name_table_offset + re->name_count * re->name_entry_size;
3064  md->start_subject = (const unsigned char *)subject;  md->start_subject = (const unsigned char *)subject;
3065  md->end_subject = end_subject;  md->end_subject = end_subject;
3066    md->start_offset = start_offset;
3067  md->moptions = options;  md->moptions = options;
3068  md->poptions = re->options;  md->poptions = re->options;
3069    
3070    /* If the BSR option is not set at match time, copy what was set
3071    at compile time. */
3072    
3073    if ((md->moptions & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == 0)
3074      {
3075      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
3076        md->moptions |= re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE);
3077    #ifdef BSR_ANYCRLF
3078      else md->moptions |= PCRE_BSR_ANYCRLF;
3079    #endif
3080      }
3081    
3082  /* Handle different types of newline. The three bits give eight cases. If  /* Handle different types of newline. The three bits give eight cases. If
3083  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. */
3084    
# Line 2579  switch ((((options & PCRE_NEWLINE_BITS) Line 3086  switch ((((options & PCRE_NEWLINE_BITS)
3086           PCRE_NEWLINE_BITS)           PCRE_NEWLINE_BITS)
3087    {    {
3088    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
3089    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = CHAR_CR; break;
3090    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = CHAR_NL; break;
3091    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
3092         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break;
3093    case PCRE_NEWLINE_ANY: newline = -1; break;    case PCRE_NEWLINE_ANY: newline = -1; break;
3094    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
3095    default: return PCRE_ERROR_BADNEWLINE;    default: return PCRE_ERROR_BADNEWLINE;
# Line 2618  back the character offset. */ Line 3125  back the character offset. */
3125  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3126  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
3127    {    {
3128    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    int erroroffset;
3129      return PCRE_ERROR_BADUTF8;    int errorcode = _pcre_valid_utf8((uschar *)subject, length, &erroroffset);
3130    if (start_offset > 0 && start_offset < length)    if (errorcode != 0)
3131      {      {
3132      int tb = ((uschar *)subject)[start_offset];      if (offsetcount >= 2)
     if (tb > 127)  
3133        {        {
3134        tb &= 0xc0;        offsets[0] = erroroffset;
3135        if (tb != 0 && tb != 0xc0) return PCRE_ERROR_BADUTF8_OFFSET;        offsets[1] = errorcode;
3136        }        }
3137      }      return (errorcode <= PCRE_UTF8_ERR5 && (options & PCRE_PARTIAL_HARD) != 0)?
3138          PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
3139        }
3140      if (start_offset > 0 && start_offset < length &&
3141            (((USPTR)subject)[start_offset] & 0xc0) == 0x80)
3142        return PCRE_ERROR_BADUTF8_OFFSET;
3143    }    }
3144  #endif  #endif
3145    
# Line 2642  if (md->tables == NULL) md->tables = _pc Line 3153  if (md->tables == NULL) md->tables = _pc
3153  used in a loop when finding where to start. */  used in a loop when finding where to start. */
3154    
3155  lcc = md->tables + lcc_offset;  lcc = md->tables + lcc_offset;
3156  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
3157  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
3158    
3159  /* 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 2653  studied, there may be a bitmap of possib Line 3164  studied, there may be a bitmap of possib
3164    
3165  if (!anchored)  if (!anchored)
3166    {    {
3167    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
3168      {      {
3169      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
3170      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
# Line 2661  if (!anchored) Line 3172  if (!anchored)
3172      }      }
3173    else    else
3174      {      {
3175      if (startline && study != NULL &&      if (!startline && study != NULL &&
3176           (study->options & PCRE_STUDY_MAPPED) != 0)           (study->flags & PCRE_STUDY_MAPPED) != 0)
3177        start_bits = study->start_bits;        start_bits = study->start_bits;
3178      }      }
3179    }    }
# Line 2670  if (!anchored) Line 3181  if (!anchored)
3181  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
3182  character" set. */  character" set. */
3183    
3184  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
3185    {    {
3186    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
3187    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
# Line 2678  if ((re->options & PCRE_REQCHSET) != 0) Line 3189  if ((re->options & PCRE_REQCHSET) != 0)
3189    }    }
3190    
3191  /* 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
3192  failed match. Unless restarting, optimize by moving to the first match  failed match. If not restarting, perform certain optimizations at the start of
3193  character if possible, when not anchored. Then unless wanting a partial match,  a match. */
 check for a required later character. */  
3194    
3195  for (;;)  for (;;)
3196    {    {
# Line 2690  for (;;) Line 3200  for (;;)
3200      {      {
3201      const uschar *save_end_subject = end_subject;      const uschar *save_end_subject = end_subject;
3202    
3203      /* 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
3204      start of the match is constrained to the first line of a multiline string.      line of a multiline string. Implement this by temporarily adjusting
3205      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
3206      scanning at a newline. If the match fails at the newline, later code breaks      the newline, later code breaks this loop. */
     this loop. */  
3207    
3208      if (firstline)      if (firstline)
3209        {        {
3210        const uschar *t = current_subject;        USPTR t = current_subject;
3211    #ifdef SUPPORT_UTF8
3212          if (utf8)
3213            {
3214            while (t < md->end_subject && !IS_NEWLINE(t))
3215              {
3216              t++;
3217              while (t < end_subject && (*t & 0xc0) == 0x80) t++;
3218              }
3219            }
3220          else
3221    #endif
3222        while (t < md->end_subject && !IS_NEWLINE(t)) t++;        while (t < md->end_subject && !IS_NEWLINE(t)) t++;
3223        end_subject = t;        end_subject = t;
3224        }        }
3225    
3226      if (first_byte >= 0)      /* There are some optimizations that avoid running the match if a known
3227        starting point is not found. However, there is an option that disables
3228        these, for testing and for ensuring that all callouts do actually occur.
3229        The option can be set in the regex by (*NO_START_OPT) or passed in
3230        match-time options. */
3231    
3232        if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)
3233        {        {
3234        if (first_byte_caseless)        /* Advance to a known first byte. */
3235          while (current_subject < end_subject &&  
3236                 lcc[*current_subject] != first_byte)        if (first_byte >= 0)
3237            current_subject++;          {
3238        else          if (first_byte_caseless)
3239          while (current_subject < end_subject && *current_subject != first_byte)            while (current_subject < end_subject &&
3240            current_subject++;                   lcc[*current_subject] != first_byte)
3241        }              current_subject++;
3242            else
3243              while (current_subject < end_subject &&
3244                     *current_subject != first_byte)
3245                current_subject++;
3246            }
3247    
3248      /* Or to just after a linebreak for a multiline match if possible */        /* Or to just after a linebreak for a multiline match if possible */
3249    
3250      else if (startline)        else if (startline)
       {  
       if (current_subject > md->start_subject + start_offset)  
3251          {          {
3252          while (current_subject <= end_subject && !WAS_NEWLINE(current_subject))          if (current_subject > md->start_subject + start_offset)
3253            current_subject++;            {
3254    #ifdef SUPPORT_UTF8
3255              if (utf8)
3256                {
3257                while (current_subject < end_subject &&
3258                       !WAS_NEWLINE(current_subject))
3259                  {
3260                  current_subject++;
3261                  while(current_subject < end_subject &&
3262                        (*current_subject & 0xc0) == 0x80)
3263                    current_subject++;
3264                  }
3265                }
3266              else
3267    #endif
3268              while (current_subject < end_subject && !WAS_NEWLINE(current_subject))
3269                current_subject++;
3270    
3271              /* If we have just passed a CR and the newline option is ANY or
3272              ANYCRLF, and we are now at a LF, advance the match position by one
3273              more character. */
3274    
3275          /* If we have just passed a CR and the newline option is ANY or            if (current_subject[-1] == CHAR_CR &&
3276          ANYCRLF, and we are now at a LF, advance the match position by one more                 (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
3277          character. */                 current_subject < end_subject &&
3278                   *current_subject == CHAR_NL)
3279          if (current_subject[-1] == '\r' &&              current_subject++;
3280               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&            }
              current_subject < end_subject &&  
              *current_subject == '\n')  
           current_subject++;  
3281          }          }
       }  
3282    
3283      /* Or to a non-unique first char after study */        /* Or to a non-unique first char after study */
3284    
3285      else if (start_bits != NULL)        else if (start_bits != NULL)
       {  
       while (current_subject < end_subject)  
3286          {          {
3287          register unsigned int c = *current_subject;          while (current_subject < end_subject)
3288          if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++;            {
3289              register unsigned int c = *current_subject;
3290              if ((start_bits[c/8] & (1 << (c&7))) == 0)
3291                {
3292                current_subject++;
3293    #ifdef SUPPORT_UTF8
3294                if (utf8)
3295                  while(current_subject < end_subject &&
3296                        (*current_subject & 0xc0) == 0x80) current_subject++;
3297    #endif
3298                }
3299            else break;            else break;
3300              }
3301          }          }
3302        }        }
3303    
3304      /* Restore fudged end_subject */      /* Restore fudged end_subject */
3305    
3306      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);  
3307    
3308      /* 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
3309      place we found it at last time. */      disabling is explicitly requested (and of course, by the test above, this
3310        code is not obeyed when restarting after a partial match). */
3311    
3312      if (p > req_byte_ptr)      if ((options & PCRE_NO_START_OPTIMIZE) == 0 &&
3313            (options & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) == 0)
3314        {        {
3315        if (req_byte_caseless)        /* If the pattern was studied, a minimum subject length may be set. This
3316          {        is a lower bound; no actual string of that length may actually match the
3317          while (p < end_subject)        pattern. Although the value is, strictly, in characters, we treat it as
3318            {        bytes to avoid spending too much time in this optimization. */
3319            register int pp = *p++;  
3320            if (pp == req_byte || pp == req_byte2) { p--; break; }        if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 &&
3321            }            (pcre_uint32)(end_subject - current_subject) < study->minlength)
3322          }          return PCRE_ERROR_NOMATCH;
3323        else  
3324          /* If req_byte is set, we know that that character must appear in the
3325          subject for the match to succeed. If the first character is set, req_byte
3326          must be later in the subject; otherwise the test starts at the match
3327          point. This optimization can save a huge amount of work in patterns with
3328          nested unlimited repeats that aren't going to match. Writing separate
3329          code for cased/caseless versions makes it go faster, as does using an
3330          autoincrement and backing off on a match.
3331    
3332          HOWEVER: when the subject string is very, very long, searching to its end
3333          can take a long time, and give bad performance on quite ordinary
3334          patterns. This showed up when somebody was matching /^C/ on a 32-megabyte
3335          string... so we don't do this when the string is sufficiently long. */
3336    
3337          if (req_byte >= 0 && end_subject - current_subject < REQ_BYTE_MAX)
3338          {          {
3339          while (p < end_subject)          register const uschar *p = current_subject + ((first_byte >= 0)? 1 : 0);
3340    
3341            /* We don't need to repeat the search if we haven't yet reached the
3342            place we found it at last time. */
3343    
3344            if (p > req_byte_ptr)
3345            {            {
3346            if (*p++ == req_byte) { p--; break; }            if (req_byte_caseless)
3347            }              {
3348          }              while (p < end_subject)
3349                  {
3350                  register int pp = *p++;
3351                  if (pp == req_byte || pp == req_byte2) { p--; break; }
3352                  }
3353                }
3354              else
3355                {
3356                while (p < end_subject)
3357                  {
3358                  if (*p++ == req_byte) { p--; break; }
3359                  }
3360                }
3361    
3362        /* If we can't find the required character, break the matching loop,            /* If we can't find the required character, break the matching loop,
3363        which will cause a return or PCRE_ERROR_NOMATCH. */            which will cause a return or PCRE_ERROR_NOMATCH. */
3364    
3365        if (p >= end_subject) break;            if (p >= end_subject) break;
3366    
3367        /* If we have found the required character, save the point where we            /* If we have found the required character, save the point where we
3368        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
3369        the start hasn't passed this character yet. */            the start hasn't passed this character yet. */
3370    
3371        req_byte_ptr = p;            req_byte_ptr = p;
3372              }
3373            }
3374        }        }
3375      }      }   /* End of optimizations that are done when not restarting */
3376    
3377    /* OK, now we can do the business */    /* OK, now we can do the business */
3378    
3379      md->start_used_ptr = current_subject;
3380    
3381    rc = internal_dfa_exec(    rc = internal_dfa_exec(
3382      md,                                /* fixed match data */      md,                                /* fixed match data */
3383      md->start_code,                    /* this subexpression's code */      md->start_code,                    /* this subexpression's code */
# Line 2819  for (;;) Line 3387  for (;;)
3387      offsetcount,                       /* size of same */      offsetcount,                       /* size of same */
3388      workspace,                         /* workspace vector */      workspace,                         /* workspace vector */
3389      wscount,                           /* size of same */      wscount,                           /* size of same */
     re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL), /* ims flags */  
3390      0,                                 /* function recurse level */      0,                                 /* function recurse level */
3391      0);                                /* regex recurse level */      0);                                /* regex recurse level */
3392    
# Line 2840  for (;;) Line 3407  for (;;)
3407      }      }
3408    if (current_subject > end_subject) break;    if (current_subject > end_subject) break;
3409    
3410    /* 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
3411    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
3412    character. */    or ANY or ANYCRLF, advance the match position by one more character. */
3413    
3414    if (current_subject[-1] == '\r' &&    if (current_subject[-1] == CHAR_CR &&
3415         (md->nltype == NLTYPE_ANY ||        current_subject < end_subject &&
3416          md->nltype == NLTYPE_ANYCRLF ||        *current_subject == CHAR_NL &&
3417          md->nllen == 2) &&        (re->flags & PCRE_HASCRORLF) == 0 &&
3418         current_subject < end_subject &&          (md->nltype == NLTYPE_ANY ||
3419         *current_subject == '\n')           md->nltype == NLTYPE_ANYCRLF ||
3420             md->nllen == 2))
3421      current_subject++;      current_subject++;
3422    
3423    }   /* "Bumpalong" loop */    }   /* "Bumpalong" loop */

Legend:
Removed from v.200  
changed lines
  Added in v.613

  ViewVC Help
Powered by ViewVC 1.1.5