/[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 426 by ph10, Wed Aug 26 15:38:32 2009 UTC revision 1046 by ph10, Tue Sep 25 16:27:58 2012 UTC
# Line 7  and semantics are as close as possible t Line 7  and semantics are as close as possible t
7  below for why this module is different).  below for why this module is different).
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2009 University of Cambridge             Copyright (c) 1997-2012 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 38  POSSIBILITY OF SUCH DAMAGE. Line 38  POSSIBILITY OF SUCH DAMAGE.
38  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
39  */  */
40    
   
41  /* This module contains the external function pcre_dfa_exec(), which is an  /* This module contains the external function pcre_dfa_exec(), which is an
42  alternative matching function that uses a sort of DFA algorithm (not a true  alternative matching function that uses a sort of DFA algorithm (not a true
43  FSM). This is NOT Perl- compatible, but it has advantages in certain  FSM). This is NOT Perl-compatible, but it has advantages in certain
44  applications. */  applications. */
45    
46    
47    /* NOTE ABOUT PERFORMANCE: A user of this function sent some code that improved
48    the performance of his patterns greatly. I could not use it as it stood, as it
49    was not thread safe, and made assumptions about pattern sizes. Also, it caused
50    test 7 to loop, and test 9 to crash with a segfault.
51    
52    The issue is the check for duplicate states, which is done by a simple linear
53    search up the state list. (Grep for "duplicate" below to find the code.) For
54    many patterns, there will never be many states active at one time, so a simple
55    linear search is fine. In patterns that have many active states, it might be a
56    bottleneck. The suggested code used an indexing scheme to remember which states
57    had previously been used for each character, and avoided the linear search when
58    it knew there was no chance of a duplicate. This was implemented when adding
59    states to the state lists.
60    
61    I wrote some thread-safe, not-limited code to try something similar at the time
62    of checking for duplicates (instead of when adding states), using index vectors
63    on the stack. It did give a 13% improvement with one specially constructed
64    pattern for certain subject strings, but on other strings and on many of the
65    simpler patterns in the test suite it did worse. The major problem, I think,
66    was the extra time to initialize the index. This had to be done for each call
67    of internal_dfa_exec(). (The supplied patch used a static vector, initialized
68    only once - I suspect this was the cause of the problems with the tests.)
69    
70    Overall, I concluded that the gains in some cases did not outweigh the losses
71    in others, so I abandoned this code. */
72    
73    
74    
75  #ifdef HAVE_CONFIG_H  #ifdef HAVE_CONFIG_H
76  #include "config.h"  #include "config.h"
77  #endif  #endif
# Line 78  never stored, so we push them well clear Line 105  never stored, so we push them well clear
105    
106    
107  /* This table identifies those opcodes that are followed immediately by a  /* This table identifies those opcodes that are followed immediately by a
108  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
109  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
110  "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
111  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
112  that follow must also be modified. */  the character is to be found. ***NOTE*** If the start of this table is
113    modified, the three tables that follow must also be modified. */
114    
115  static const uschar coptable[] = {  static const pcre_uint8 coptable[] = {
116    0,                             /* End                                    */    0,                             /* End                                    */
117    0, 0, 0, 0, 0,                 /* \A, \G, \K, \B, \b                     */    0, 0, 0, 0, 0,                 /* \A, \G, \K, \B, \b                     */
118    0, 0, 0, 0, 0, 0,              /* \D, \d, \S, \s, \W, \w                 */    0, 0, 0, 0, 0, 0,              /* \D, \d, \S, \s, \W, \w                 */
119    0, 0, 0,                       /* Any, AllAny, Anybyte                   */    0, 0, 0,                       /* Any, AllAny, Anybyte                   */
120    0, 0, 0,                       /* NOTPROP, PROP, EXTUNI                  */    0, 0,                          /* \P, \p                                 */
121    0, 0, 0, 0, 0,                 /* \R, \H, \h, \V, \v                     */    0, 0, 0, 0, 0,                 /* \R, \H, \h, \V, \v                     */
122    0, 0, 0, 0, 0,                 /* \Z, \z, Opt, ^, $                      */    0,                             /* \X                                     */
123      0, 0, 0, 0, 0, 0,              /* \Z, \z, ^, ^M, $, $M                   */
124    1,                             /* Char                                   */    1,                             /* Char                                   */
125    1,                             /* Charnc                                 */    1,                             /* Chari                                  */
126    1,                             /* not                                    */    1,                             /* not                                    */
127      1,                             /* noti                                   */
128    /* Positive single-char repeats                                          */    /* Positive single-char repeats                                          */
129    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */
130    3, 3, 3,                       /* upto, minupto, exact                   */    1+IMM2_SIZE, 1+IMM2_SIZE,      /* upto, minupto                          */
131    1, 1, 1, 3,                    /* *+, ++, ?+, upto+                      */    1+IMM2_SIZE,                   /* exact                                  */
132      1, 1, 1, 1+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */
133      1, 1, 1, 1, 1, 1,              /* *I, *?I, +I, +?I, ?I, ??I              */
134      1+IMM2_SIZE, 1+IMM2_SIZE,      /* upto I, minupto I                      */
135      1+IMM2_SIZE,                   /* exact I                                */
136      1, 1, 1, 1+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */
137    /* Negative single-char repeats - only for chars < 256                   */    /* Negative single-char repeats - only for chars < 256                   */
138    1, 1, 1, 1, 1, 1,              /* NOT *, *?, +, +?, ?, ??                */    1, 1, 1, 1, 1, 1,              /* NOT *, *?, +, +?, ?, ??                */
139    3, 3, 3,                       /* NOT upto, minupto, exact               */    1+IMM2_SIZE, 1+IMM2_SIZE,      /* NOT upto, minupto                      */
140    1, 1, 1, 3,                    /* NOT *+, ++, ?+, updo+                  */    1+IMM2_SIZE,                   /* NOT exact                              */
141      1, 1, 1, 1+IMM2_SIZE,          /* NOT *+, ++, ?+, upto+                  */
142      1, 1, 1, 1, 1, 1,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */
143      1+IMM2_SIZE, 1+IMM2_SIZE,      /* NOT upto I, minupto I                  */
144      1+IMM2_SIZE,                   /* NOT exact I                            */
145      1, 1, 1, 1+IMM2_SIZE,          /* NOT *+I, ++I, ?+I, upto+I              */
146    /* Positive type repeats                                                 */    /* Positive type repeats                                                 */
147    1, 1, 1, 1, 1, 1,              /* Type *, *?, +, +?, ?, ??               */    1, 1, 1, 1, 1, 1,              /* Type *, *?, +, +?, ?, ??               */
148    3, 3, 3,                       /* Type upto, minupto, exact              */    1+IMM2_SIZE, 1+IMM2_SIZE,      /* Type upto, minupto                     */
149    1, 1, 1, 3,                    /* Type *+, ++, ?+, upto+                 */    1+IMM2_SIZE,                   /* Type exact                             */
150      1, 1, 1, 1+IMM2_SIZE,          /* Type *+, ++, ?+, upto+                 */
151    /* Character class & ref repeats                                         */    /* Character class & ref repeats                                         */
152    0, 0, 0, 0, 0, 0,              /* *, *?, +, +?, ?, ??                    */    0, 0, 0, 0, 0, 0,              /* *, *?, +, +?, ?, ??                    */
153    0, 0,                          /* CRRANGE, CRMINRANGE                    */    0, 0,                          /* CRRANGE, CRMINRANGE                    */
# Line 114  static const uschar coptable[] = { Line 155  static const uschar coptable[] = {
155    0,                             /* NCLASS                                 */    0,                             /* NCLASS                                 */
156    0,                             /* XCLASS - variable length               */    0,                             /* XCLASS - variable length               */
157    0,                             /* REF                                    */    0,                             /* REF                                    */
158      0,                             /* REFI                                   */
159    0,                             /* RECURSE                                */    0,                             /* RECURSE                                */
160    0,                             /* CALLOUT                                */    0,                             /* CALLOUT                                */
161    0,                             /* Alt                                    */    0,                             /* Alt                                    */
162    0,                             /* Ket                                    */    0,                             /* Ket                                    */
163    0,                             /* KetRmax                                */    0,                             /* KetRmax                                */
164    0,                             /* KetRmin                                */    0,                             /* KetRmin                                */
165      0,                             /* KetRpos                                */
166      0,                             /* Reverse                                */
167    0,                             /* Assert                                 */    0,                             /* Assert                                 */
168    0,                             /* Assert not                             */    0,                             /* Assert not                             */
169    0,                             /* Assert behind                          */    0,                             /* Assert behind                          */
170    0,                             /* Assert behind not                      */    0,                             /* Assert behind not                      */
171      0, 0,                          /* ONCE, ONCE_NC                          */
172      0, 0, 0, 0, 0,                 /* BRA, BRAPOS, CBRA, CBRAPOS, COND       */
173      0, 0, 0, 0, 0,                 /* SBRA, SBRAPOS, SCBRA, SCBRAPOS, SCOND  */
174      0, 0,                          /* CREF, NCREF                            */
175      0, 0,                          /* RREF, NRREF                            */
176      0,                             /* DEF                                    */
177      0, 0, 0,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */
178      0, 0, 0,                       /* MARK, PRUNE, PRUNE_ARG                 */
179      0, 0, 0, 0,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG         */
180      0, 0, 0, 0,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */
181      0, 0                           /* CLOSE, SKIPZERO  */
182    };
183    
184    /* This table identifies those opcodes that inspect a character. It is used to
185    remember the fact that a character could have been inspected when the end of
186    the subject is reached. ***NOTE*** If the start of this table is modified, the
187    two tables that follow must also be modified. */
188    
189    static const pcre_uint8 poptable[] = {
190      0,                             /* End                                    */
191      0, 0, 0, 1, 1,                 /* \A, \G, \K, \B, \b                     */
192      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */
193      1, 1, 1,                       /* Any, AllAny, Anybyte                   */
194      1, 1,                          /* \P, \p                                 */
195      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */
196      1,                             /* \X                                     */
197      0, 0, 0, 0, 0, 0,              /* \Z, \z, ^, ^M, $, $M                   */
198      1,                             /* Char                                   */
199      1,                             /* Chari                                  */
200      1,                             /* not                                    */
201      1,                             /* noti                                   */
202      /* Positive single-char repeats                                          */
203      1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */
204      1, 1, 1,                       /* upto, minupto, exact                   */
205      1, 1, 1, 1,                    /* *+, ++, ?+, upto+                      */
206      1, 1, 1, 1, 1, 1,              /* *I, *?I, +I, +?I, ?I, ??I              */
207      1, 1, 1,                       /* upto I, minupto I, exact I             */
208      1, 1, 1, 1,                    /* *+I, ++I, ?+I, upto+I                  */
209      /* Negative single-char repeats - only for chars < 256                   */
210      1, 1, 1, 1, 1, 1,              /* NOT *, *?, +, +?, ?, ??                */
211      1, 1, 1,                       /* NOT upto, minupto, exact               */
212      1, 1, 1, 1,                    /* NOT *+, ++, ?+, upto+                  */
213      1, 1, 1, 1, 1, 1,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */
214      1, 1, 1,                       /* NOT upto I, minupto I, exact I         */
215      1, 1, 1, 1,                    /* NOT *+I, ++I, ?+I, upto+I              */
216      /* Positive type repeats                                                 */
217      1, 1, 1, 1, 1, 1,              /* Type *, *?, +, +?, ?, ??               */
218      1, 1, 1,                       /* Type upto, minupto, exact              */
219      1, 1, 1, 1,                    /* Type *+, ++, ?+, upto+                 */
220      /* Character class & ref repeats                                         */
221      1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */
222      1, 1,                          /* CRRANGE, CRMINRANGE                    */
223      1,                             /* CLASS                                  */
224      1,                             /* NCLASS                                 */
225      1,                             /* XCLASS - variable length               */
226      0,                             /* REF                                    */
227      0,                             /* REFI                                   */
228      0,                             /* RECURSE                                */
229      0,                             /* CALLOUT                                */
230      0,                             /* Alt                                    */
231      0,                             /* Ket                                    */
232      0,                             /* KetRmax                                */
233      0,                             /* KetRmin                                */
234      0,                             /* KetRpos                                */
235    0,                             /* Reverse                                */    0,                             /* Reverse                                */
236    0, 0, 0, 0,                    /* ONCE, BRA, CBRA, COND                  */    0,                             /* Assert                                 */
237    0, 0, 0,                       /* SBRA, SCBRA, SCOND                     */    0,                             /* Assert not                             */
238    0,                             /* CREF                                   */    0,                             /* Assert behind                          */
239    0,                             /* RREF                                   */    0,                             /* Assert behind not                      */
240      0, 0,                          /* ONCE, ONCE_NC                          */
241      0, 0, 0, 0, 0,                 /* BRA, BRAPOS, CBRA, CBRAPOS, COND       */
242      0, 0, 0, 0, 0,                 /* SBRA, SBRAPOS, SCBRA, SCBRAPOS, SCOND  */
243      0, 0,                          /* CREF, NCREF                            */
244      0, 0,                          /* RREF, NRREF                            */
245    0,                             /* DEF                                    */    0,                             /* DEF                                    */
246    0, 0,                          /* BRAZERO, BRAMINZERO                    */    0, 0, 0,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */
247    0, 0, 0, 0,                    /* PRUNE, SKIP, THEN, COMMIT              */    0, 0, 0,                       /* MARK, PRUNE, PRUNE_ARG                 */
248    0, 0, 0                        /* FAIL, ACCEPT, SKIPZERO                 */    0, 0, 0, 0,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG         */
249      0, 0, 0, 0,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */
250      0, 0                           /* CLOSE, SKIPZERO                        */
251  };  };
252    
253  /* 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,
254  and \w */  and \w */
255    
256  static const uschar toptable1[] = {  static const pcre_uint8 toptable1[] = {
257    0, 0, 0, 0, 0, 0,    0, 0, 0, 0, 0, 0,
258    ctype_digit, ctype_digit,    ctype_digit, ctype_digit,
259    ctype_space, ctype_space,    ctype_space, ctype_space,
# Line 146  static const uschar toptable1[] = { Line 261  static const uschar toptable1[] = {
261    0, 0                            /* OP_ANY, OP_ALLANY */    0, 0                            /* OP_ANY, OP_ALLANY */
262  };  };
263    
264  static const uschar toptable2[] = {  static const pcre_uint8 toptable2[] = {
265    0, 0, 0, 0, 0, 0,    0, 0, 0, 0, 0, 0,
266    ctype_digit, 0,    ctype_digit, 0,
267    ctype_space, 0,    ctype_space, 0,
# Line 163  these structures in, is a vector of ints Line 278  these structures in, is a vector of ints
278  typedef struct stateblock {  typedef struct stateblock {
279    int offset;                     /* Offset to opcode */    int offset;                     /* Offset to opcode */
280    int count;                      /* Count for repeats */    int count;                      /* Count for repeats */
   int ims;                        /* ims flag bits */  
281    int data;                       /* Some use extra data */    int data;                       /* Some use extra data */
282  } stateblock;  } stateblock;
283    
284  #define INTS_PER_STATEBLOCK  (sizeof(stateblock)/sizeof(int))  #define INTS_PER_STATEBLOCK  (int)(sizeof(stateblock)/sizeof(int))
285    
286    
287  #ifdef DEBUG  #ifdef PCRE_DEBUG
288  /*************************************************  /*************************************************
289  *             Print character string             *  *             Print character string             *
290  *************************************************/  *************************************************/
# Line 186  Returns:       nothing Line 300  Returns:       nothing
300  */  */
301    
302  static void  static void
303  pchars(unsigned char *p, int length, FILE *f)  pchars(const pcre_uchar *p, int length, FILE *f)
304  {  {
305  int c;  int c;
306  while (length-- > 0)  while (length-- > 0)
# Line 219  Arguments: Line 333  Arguments:
333    offsetcount       size of same    offsetcount       size of same
334    workspace         vector of workspace    workspace         vector of workspace
335    wscount           size of same    wscount           size of same
   ims               the current ims flags  
336    rlevel            function call recursion level    rlevel            function call recursion level
   recursing         regex recursive call level  
337    
338  Returns:            > 0 => number of match offset pairs placed in offsets  Returns:            > 0 => number of match offset pairs placed in offsets
339                      = 0 => offsets overflowed; longest matches are present                      = 0 => offsets overflowed; longest matches are present
# Line 236  for the current character, one for the f Line 348  for the current character, one for the f
348      { \      { \
349      next_active_state->offset = (x); \      next_active_state->offset = (x); \
350      next_active_state->count  = (y); \      next_active_state->count  = (y); \
     next_active_state->ims    = ims; \  
351      next_active_state++; \      next_active_state++; \
352      DPRINTF(("%.*sADD_ACTIVE(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \      DPRINTF(("%.*sADD_ACTIVE(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \
353      } \      } \
# Line 247  for the current character, one for the f Line 358  for the current character, one for the f
358      { \      { \
359      next_active_state->offset = (x); \      next_active_state->offset = (x); \
360      next_active_state->count  = (y); \      next_active_state->count  = (y); \
     next_active_state->ims    = ims; \  
361      next_active_state->data   = (z); \      next_active_state->data   = (z); \
362      next_active_state++; \      next_active_state++; \
363      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 259  for the current character, one for the f Line 369  for the current character, one for the f
369      { \      { \
370      next_new_state->offset = (x); \      next_new_state->offset = (x); \
371      next_new_state->count  = (y); \      next_new_state->count  = (y); \
     next_new_state->ims    = ims; \  
372      next_new_state++; \      next_new_state++; \
373      DPRINTF(("%.*sADD_NEW(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \      DPRINTF(("%.*sADD_NEW(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \
374      } \      } \
# Line 270  for the current character, one for the f Line 379  for the current character, one for the f
379      { \      { \
380      next_new_state->offset = (x); \      next_new_state->offset = (x); \
381      next_new_state->count  = (y); \      next_new_state->count  = (y); \
     next_new_state->ims    = ims; \  
382      next_new_state->data   = (z); \      next_new_state->data   = (z); \
383      next_new_state++; \      next_new_state++; \
384      DPRINTF(("%.*sADD_NEW_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \      DPRINTF(("%.*sADD_NEW_DATA(%d,%d,%d) line %d\n", rlevel*2-2, SP, \
385          (x), (y), (z), __LINE__)); \
386      } \      } \
387    else return PCRE_ERROR_DFA_WSSIZE    else return PCRE_ERROR_DFA_WSSIZE
388    
# Line 282  for the current character, one for the f Line 391  for the current character, one for the f
391  static int  static int
392  internal_dfa_exec(  internal_dfa_exec(
393    dfa_match_data *md,    dfa_match_data *md,
394    const uschar *this_start_code,    const pcre_uchar *this_start_code,
395    const uschar *current_subject,    const pcre_uchar *current_subject,
396    int start_offset,    int start_offset,
397    int *offsets,    int *offsets,
398    int offsetcount,    int offsetcount,
399    int *workspace,    int *workspace,
400    int wscount,    int wscount,
401    int ims,    int  rlevel)
   int  rlevel,  
   int  recursing)  
402  {  {
403  stateblock *active_states, *new_states, *temp_states;  stateblock *active_states, *new_states, *temp_states;
404  stateblock *next_active_state, *next_new_state;  stateblock *next_active_state, *next_new_state;
405    
406  const uschar *ctypes, *lcc, *fcc;  const pcre_uint8 *ctypes, *lcc, *fcc;
407  const uschar *ptr;  const pcre_uchar *ptr;
408  const uschar *end_code, *first_op;  const pcre_uchar *end_code, *first_op;
409    
410    dfa_recursion_info new_recursive;
411    
412  int active_count, new_count, match_count;  int active_count, new_count, match_count;
413    
414  /* Some fields in the md block are frequently referenced, so we load them into  /* Some fields in the md block are frequently referenced, so we load them into
415  independent variables in the hope that this will perform better. */  independent variables in the hope that this will perform better. */
416    
417  const uschar *start_subject = md->start_subject;  const pcre_uchar *start_subject = md->start_subject;
418  const uschar *end_subject = md->end_subject;  const pcre_uchar *end_subject = md->end_subject;
419  const uschar *start_code = md->start_code;  const pcre_uchar *start_code = md->start_code;
420    
421  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
422  BOOL utf8 = (md->poptions & PCRE_UTF8) != 0;  BOOL utf = (md->poptions & PCRE_UTF8) != 0;
423  #else  #else
424  BOOL utf8 = FALSE;  BOOL utf = FALSE;
425  #endif  #endif
426    
427    BOOL reset_could_continue = FALSE;
428    
429  rlevel++;  rlevel++;
430  offsetcount &= (-2);  offsetcount &= (-2);
431    
# Line 323  wscount = (wscount - (wscount % (INTS_PE Line 434  wscount = (wscount - (wscount % (INTS_PE
434            (2 * INTS_PER_STATEBLOCK);            (2 * INTS_PER_STATEBLOCK);
435    
436  DPRINTF(("\n%.*s---------------------\n"  DPRINTF(("\n%.*s---------------------\n"
437    "%.*sCall to internal_dfa_exec f=%d r=%d\n",    "%.*sCall to internal_dfa_exec f=%d\n",
438    rlevel*2-2, SP, rlevel*2-2, SP, rlevel, recursing));    rlevel*2-2, SP, rlevel*2-2, SP, rlevel));
439    
440  ctypes = md->tables + ctypes_offset;  ctypes = md->tables + ctypes_offset;
441  lcc = md->tables + lcc_offset;  lcc = md->tables + lcc_offset;
# Line 337  next_new_state = new_states = active_sta Line 448  next_new_state = new_states = active_sta
448  new_count = 0;  new_count = 0;
449    
450  first_op = this_start_code + 1 + LINK_SIZE +  first_op = this_start_code + 1 + LINK_SIZE +
451    ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA)? 2:0);    ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA ||
452        *this_start_code == OP_CBRAPOS || *this_start_code == OP_SCBRAPOS)
453        ? IMM2_SIZE:0);
454    
455  /* 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
456  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 365  if (*first_op == OP_REVERSE) Line 478  if (*first_op == OP_REVERSE)
478    /* If we can't go back the amount required for the longest lookbehind    /* If we can't go back the amount required for the longest lookbehind
479    pattern, go back as far as we can; some alternatives may still be viable. */    pattern, go back as far as we can; some alternatives may still be viable. */
480    
481  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
482    /* In character mode we have to step back character by character */    /* In character mode we have to step back character by character */
483    
484    if (utf8)    if (utf)
485      {      {
486      for (gone_back = 0; gone_back < max_back; gone_back++)      for (gone_back = 0; gone_back < max_back; gone_back++)
487        {        {
488        if (current_subject <= start_subject) break;        if (current_subject <= start_subject) break;
489        current_subject--;        current_subject--;
490        while (current_subject > start_subject &&        ACROSSCHAR(current_subject > start_subject, *current_subject, current_subject--);
              (*current_subject & 0xc0) == 0x80)  
         current_subject--;  
491        }        }
492      }      }
493    else    else
# Line 386  if (*first_op == OP_REVERSE) Line 497  if (*first_op == OP_REVERSE)
497    
498      {      {
499      gone_back = (current_subject - max_back < start_subject)?      gone_back = (current_subject - max_back < start_subject)?
500        current_subject - start_subject : max_back;        (int)(current_subject - start_subject) : max_back;
501      current_subject -= gone_back;      current_subject -= gone_back;
502      }      }
503    
504      /* Save the earliest consulted character */
505    
506      if (current_subject < md->start_used_ptr)
507        md->start_used_ptr = current_subject;
508    
509    /* Now we can process the individual branches. */    /* Now we can process the individual branches. */
510    
511    end_code = this_start_code;    end_code = this_start_code;
# Line 398  if (*first_op == OP_REVERSE) Line 514  if (*first_op == OP_REVERSE)
514      int back = GET(end_code, 2+LINK_SIZE);      int back = GET(end_code, 2+LINK_SIZE);
515      if (back <= gone_back)      if (back <= gone_back)
516        {        {
517        int bstate = end_code - start_code + 2 + 2*LINK_SIZE;        int bstate = (int)(end_code - start_code + 2 + 2*LINK_SIZE);
518        ADD_NEW_DATA(-bstate, 0, gone_back - back);        ADD_NEW_DATA(-bstate, 0, gone_back - back);
519        }        }
520      end_code += GET(end_code, 1);      end_code += GET(end_code, 1);
# Line 431  else Line 547  else
547    else    else
548      {      {
549      int length = 1 + LINK_SIZE +      int length = 1 + LINK_SIZE +
550        ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA)? 2:0);        ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA ||
551            *this_start_code == OP_CBRAPOS || *this_start_code == OP_SCBRAPOS)
552            ? IMM2_SIZE:0);
553      do      do
554        {        {
555        ADD_NEW(end_code - start_code + length, 0);        ADD_NEW((int)(end_code - start_code + length), 0);
556        end_code += GET(end_code, 1);        end_code += GET(end_code, 1);
557        length = 1 + LINK_SIZE;        length = 1 + LINK_SIZE;
558        }        }
# Line 444  else Line 562  else
562    
563  workspace[0] = 0;    /* Bit indicating which vector is current */  workspace[0] = 0;    /* Bit indicating which vector is current */
564    
565  DPRINTF(("%.*sEnd state = %d\n", rlevel*2-2, SP, end_code - start_code));  DPRINTF(("%.*sEnd state = %d\n", rlevel*2-2, SP, (int)(end_code - start_code)));
566    
567  /* Loop for scanning the subject */  /* Loop for scanning the subject */
568    
# Line 454  for (;;) Line 572  for (;;)
572    int i, j;    int i, j;
573    int clen, dlen;    int clen, dlen;
574    unsigned int c, d;    unsigned int c, d;
575      int forced_fail = 0;
576      BOOL partial_newline = FALSE;
577      BOOL could_continue = reset_could_continue;
578      reset_could_continue = FALSE;
579    
580    /* 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
581    new state list. */    new state list. */
# Line 467  for (;;) Line 589  for (;;)
589    workspace[0] ^= 1;              /* Remember for the restarting feature */    workspace[0] ^= 1;              /* Remember for the restarting feature */
590    workspace[1] = active_count;    workspace[1] = active_count;
591    
592  #ifdef DEBUG  #ifdef PCRE_DEBUG
593    printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP);    printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP);
594    pchars((uschar *)ptr, strlen((char *)ptr), stdout);    pchars(ptr, STRLEN_UC(ptr), stdout);
595    printf("\"\n");    printf("\"\n");
596    
597    printf("%.*sActive states: ", rlevel*2-2, SP);    printf("%.*sActive states: ", rlevel*2-2, SP);
# Line 489  for (;;) Line 611  for (;;)
611    
612    if (ptr < end_subject)    if (ptr < end_subject)
613      {      {
614      clen = 1;        /* Number of bytes in the character */      clen = 1;        /* Number of data items in the character */
615  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
616      if (utf8) { GETCHARLEN(c, ptr, clen); } else      if (utf) { GETCHARLEN(c, ptr, clen); } else
617  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
618      c = *ptr;      c = *ptr;
619      }      }
620    else    else
# Line 509  for (;;) Line 631  for (;;)
631    for (i = 0; i < active_count; i++)    for (i = 0; i < active_count; i++)
632      {      {
633      stateblock *current_state = active_states + i;      stateblock *current_state = active_states + i;
634      const uschar *code;      BOOL caseless = FALSE;
635        const pcre_uchar *code;
636      int state_offset = current_state->offset;      int state_offset = current_state->offset;
637      int count, codevalue, rrc;      int count, codevalue, rrc;
638    
639  #ifdef DEBUG  #ifdef PCRE_DEBUG
640      printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset);      printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset);
641      if (clen == 0) printf("EOL\n");      if (clen == 0) printf("EOL\n");
642        else if (c > 32 && c < 127) printf("'%c'\n", c);        else if (c > 32 && c < 127) printf("'%c'\n", c);
643          else printf("0x%02x\n", c);          else printf("0x%02x\n", c);
644  #endif  #endif
645    
     /* This variable is referred to implicity in the ADD_xxx macros. */  
   
     ims = current_state->ims;  
   
646      /* 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
647      (negated) state until the number of characters in the data field have      (negated) state until the number of characters in the data field have
648      been skipped". */      been skipped". If the could_continue flag was passed over from a previous
649        state, arrange for it to passed on. */
650    
651      if (state_offset < 0)      if (state_offset < 0)
652        {        {
# Line 535  for (;;) Line 655  for (;;)
655          DPRINTF(("%.*sSkipping this character\n", rlevel*2-2, SP));          DPRINTF(("%.*sSkipping this character\n", rlevel*2-2, SP));
656          ADD_NEW_DATA(state_offset, current_state->count,          ADD_NEW_DATA(state_offset, current_state->count,
657            current_state->data - 1);            current_state->data - 1);
658            if (could_continue) reset_could_continue = TRUE;
659          continue;          continue;
660          }          }
661        else        else
# Line 543  for (;;) Line 664  for (;;)
664          }          }
665        }        }
666    
667      /* 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.
668        See the note at the head of this module about the possibility of improving
669        performance here. */
670    
671      for (j = 0; j < i; j++)      for (j = 0; j < i; j++)
672        {        {
# Line 560  for (;;) Line 683  for (;;)
683      code = start_code + state_offset;      code = start_code + state_offset;
684      codevalue = *code;      codevalue = *code;
685    
686        /* If this opcode inspects a character, but we are at the end of the
687        subject, remember the fact for use when testing for a partial match. */
688    
689        if (clen == 0 && poptable[codevalue] != 0)
690          could_continue = TRUE;
691    
692      /* 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
693      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
694      is wrong, because sometimes zero repetitions of the subject are      is wrong, because sometimes zero repetitions of the subject are
695      permitted.      permitted.
696    
697      We also use this mechanism for opcodes such as OP_TYPEPLUS that take an      We also use this mechanism for opcodes such as OP_TYPEPLUS that take an
698      argument that is not a data character - but is always one byte long. We      argument that is not a data character - but is always one byte long because
699      have to take special action to deal with  \P, \p, \H, \h, \V, \v and \X in      the values are small. We have to take special action to deal with  \P, \p,
700      this case. To keep the other cases fast, convert these ones to new opcodes.      \H, \h, \V, \v and \X in this case. To keep the other cases fast, convert
701      */      these ones to new opcodes. */
702    
703      if (coptable[codevalue] > 0)      if (coptable[codevalue] > 0)
704        {        {
705        dlen = 1;        dlen = 1;
706  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
707        if (utf8) { GETCHARLEN(d, (code + coptable[codevalue]), dlen); } else        if (utf) { GETCHARLEN(d, (code + coptable[codevalue]), dlen); } else
708  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
709        d = code[coptable[codevalue]];        d = code[coptable[codevalue]];
710        if (codevalue >= OP_TYPESTAR)        if (codevalue >= OP_TYPESTAR)
711          {          {
# Line 606  for (;;) Line 735  for (;;)
735    
736      switch (codevalue)      switch (codevalue)
737        {        {
738    /* ========================================================================== */
739          /* These cases are never obeyed. This is a fudge that causes a compile-
740          time error if the vectors coptable or poptable, which are indexed by
741          opcode, are not the correct length. It seems to be the only way to do
742          such a check at compile time, as the sizeof() operator does not work
743          in the C preprocessor. */
744    
745          case OP_TABLE_LENGTH:
746          case OP_TABLE_LENGTH +
747            ((sizeof(coptable) == OP_TABLE_LENGTH) &&
748             (sizeof(poptable) == OP_TABLE_LENGTH)):
749          break;
750    
751  /* ========================================================================== */  /* ========================================================================== */
752        /* 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
753        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
754        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
755          subpattern, because the possessive subpattern repeats are always handled
756          using recursive calls. Thus, it never adds any new states.
757    
758          At the end of the (sub)pattern, unless we have an empty string and
759          PCRE_NOTEMPTY is set, or PCRE_NOTEMPTY_ATSTART is set and we are at the
760          start of the subject, save the match data, shifting up all previous
761        matches so we always have the longest first. */        matches so we always have the longest first. */
762    
763        case OP_KET:        case OP_KET:
764        case OP_KETRMIN:        case OP_KETRMIN:
765        case OP_KETRMAX:        case OP_KETRMAX:
766          case OP_KETRPOS:
767        if (code != end_code)        if (code != end_code)
768          {          {
769          ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0);          ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0);
# Line 624  for (;;) Line 772  for (;;)
772            ADD_ACTIVE(state_offset - GET(code, 1), 0);            ADD_ACTIVE(state_offset - GET(code, 1), 0);
773            }            }
774          }          }
775        else if (ptr > current_subject || (md->moptions & PCRE_NOTEMPTY) == 0)        else
776          {          {
777          if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0;          if (ptr > current_subject ||
778            else if (match_count > 0 && ++match_count * 2 >= offsetcount)              ((md->moptions & PCRE_NOTEMPTY) == 0 &&
779              match_count = 0;                ((md->moptions & PCRE_NOTEMPTY_ATSTART) == 0 ||
780          count = ((match_count == 0)? offsetcount : match_count * 2) - 2;                  current_subject > start_subject + md->start_offset)))
781          if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int));            {
782          if (offsetcount >= 2)            if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0;
783            {              else if (match_count > 0 && ++match_count * 2 > offsetcount)
784            offsets[0] = current_subject - start_subject;                match_count = 0;
785            offsets[1] = ptr - start_subject;            count = ((match_count == 0)? offsetcount : match_count * 2) - 2;
786            DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP,            if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int));
787              offsets[1] - offsets[0], current_subject));            if (offsetcount >= 2)
788            }              {
789          if ((md->moptions & PCRE_DFA_SHORTEST) != 0)              offsets[0] = (int)(current_subject - start_subject);
790            {              offsets[1] = (int)(ptr - start_subject);
791            DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n"              DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP,
792              "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel,                offsets[1] - offsets[0], (char *)current_subject));
793              match_count, rlevel*2-2, SP));              }
794            return match_count;            if ((md->moptions & PCRE_DFA_SHORTEST) != 0)
795                {
796                DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n"
797                  "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel,
798                  match_count, rlevel*2-2, SP));
799                return match_count;
800                }
801            }            }
802          }          }
803        break;        break;
# Line 655  for (;;) Line 809  for (;;)
809        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
810        case OP_ALT:        case OP_ALT:
811        do { code += GET(code, 1); } while (*code == OP_ALT);        do { code += GET(code, 1); } while (*code == OP_ALT);
812        ADD_ACTIVE(code - start_code, 0);        ADD_ACTIVE((int)(code - start_code), 0);
813        break;        break;
814    
815        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
# Line 663  for (;;) Line 817  for (;;)
817        case OP_SBRA:        case OP_SBRA:
818        do        do
819          {          {
820          ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0);          ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
821          code += GET(code, 1);          code += GET(code, 1);
822          }          }
823        while (*code == OP_ALT);        while (*code == OP_ALT);
# Line 672  for (;;) Line 826  for (;;)
826        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
827        case OP_CBRA:        case OP_CBRA:
828        case OP_SCBRA:        case OP_SCBRA:
829        ADD_ACTIVE(code - start_code + 3 + LINK_SIZE,  0);        ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE + IMM2_SIZE),  0);
830        code += GET(code, 1);        code += GET(code, 1);
831        while (*code == OP_ALT)        while (*code == OP_ALT)
832          {          {
833          ADD_ACTIVE(code - start_code + 1 + LINK_SIZE,  0);          ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE),  0);
834          code += GET(code, 1);          code += GET(code, 1);
835          }          }
836        break;        break;
# Line 687  for (;;) Line 841  for (;;)
841        ADD_ACTIVE(state_offset + 1, 0);        ADD_ACTIVE(state_offset + 1, 0);
842        code += 1 + GET(code, 2);        code += 1 + GET(code, 2);
843        while (*code == OP_ALT) code += GET(code, 1);        while (*code == OP_ALT) code += GET(code, 1);
844        ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0);        ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
845        break;        break;
846    
847        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
848        case OP_SKIPZERO:        case OP_SKIPZERO:
849        code += 1 + GET(code, 2);        code += 1 + GET(code, 2);
850        while (*code == OP_ALT) code += GET(code, 1);        while (*code == OP_ALT) code += GET(code, 1);
851        ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0);        ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
852        break;        break;
853    
854        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
855        case OP_CIRC:        case OP_CIRC:
856        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)))  
857          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
858        break;        break;
859    
860        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
861        case OP_EOD:        case OP_CIRCM:
862        if (ptr >= end_subject) { ADD_ACTIVE(state_offset + 1, 0); }        if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) ||
863              (ptr != end_subject && WAS_NEWLINE(ptr)))
864            { ADD_ACTIVE(state_offset + 1, 0); }
865        break;        break;
866    
867        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
868        case OP_OPT:        case OP_EOD:
869        ims = code[1];        if (ptr >= end_subject)
870        ADD_ACTIVE(state_offset + 2, 0);          {
871            if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
872              could_continue = TRUE;
873            else { ADD_ACTIVE(state_offset + 1, 0); }
874            }
875        break;        break;
876    
877        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
# Line 737  for (;;) Line 894  for (;;)
894        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
895        case OP_ANY:        case OP_ANY:
896        if (clen > 0 && !IS_NEWLINE(ptr))        if (clen > 0 && !IS_NEWLINE(ptr))
897          { ADD_NEW(state_offset + 1, 0); }          {
898            if (ptr + 1 >= md->end_subject &&
899                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
900                NLBLOCK->nltype == NLTYPE_FIXED &&
901                NLBLOCK->nllen == 2 &&
902                c == NLBLOCK->nl[0])
903              {
904              could_continue = partial_newline = TRUE;
905              }
906            else
907              {
908              ADD_NEW(state_offset + 1, 0);
909              }
910            }
911        break;        break;
912    
913        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
# Line 748  for (;;) Line 918  for (;;)
918    
919        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
920        case OP_EODN:        case OP_EODN:
921        if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen))        if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
922            could_continue = TRUE;
923          else if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen))
924          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
925        break;        break;
926    
# Line 756  for (;;) Line 928  for (;;)
928        case OP_DOLL:        case OP_DOLL:
929        if ((md->moptions & PCRE_NOTEOL) == 0)        if ((md->moptions & PCRE_NOTEOL) == 0)
930          {          {
931          if (clen == 0 ||          if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
932              could_continue = TRUE;
933            else if (clen == 0 ||
934              ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) &&              ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) &&
935                 ((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen)                 (ptr == end_subject - md->nllen)
936              ))              ))
937            { ADD_ACTIVE(state_offset + 1, 0); }            { ADD_ACTIVE(state_offset + 1, 0); }
938            else if (ptr + 1 >= md->end_subject &&
939                     (md->moptions & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) != 0 &&
940                     NLBLOCK->nltype == NLTYPE_FIXED &&
941                     NLBLOCK->nllen == 2 &&
942                     c == NLBLOCK->nl[0])
943              {
944              if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
945                {
946                reset_could_continue = TRUE;
947                ADD_NEW_DATA(-(state_offset + 1), 0, 1);
948                }
949              else could_continue = partial_newline = TRUE;
950              }
951            }
952          break;
953    
954          /*-----------------------------------------------------------------*/
955          case OP_DOLLM:
956          if ((md->moptions & PCRE_NOTEOL) == 0)
957            {
958            if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
959              could_continue = TRUE;
960            else if (clen == 0 ||
961                ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr)))
962              { ADD_ACTIVE(state_offset + 1, 0); }
963            else if (ptr + 1 >= md->end_subject &&
964                     (md->moptions & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) != 0 &&
965                     NLBLOCK->nltype == NLTYPE_FIXED &&
966                     NLBLOCK->nllen == 2 &&
967                     c == NLBLOCK->nl[0])
968              {
969              if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
970                {
971                reset_could_continue = TRUE;
972                ADD_NEW_DATA(-(state_offset + 1), 0, 1);
973                }
974              else could_continue = partial_newline = TRUE;
975              }
976          }          }
977        else if ((ims & PCRE_MULTILINE) != 0 && IS_NEWLINE(ptr))        else if (IS_NEWLINE(ptr))
978          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
979        break;        break;
980    
# Line 793  for (;;) Line 1005  for (;;)
1005    
1006          if (ptr > start_subject)          if (ptr > start_subject)
1007            {            {
1008            const uschar *temp = ptr - 1;            const pcre_uchar *temp = ptr - 1;
1009  #ifdef SUPPORT_UTF8            if (temp < md->start_used_ptr) md->start_used_ptr = temp;
1010            if (utf8) BACKCHAR(temp);  #ifdef SUPPORT_UTF
1011              if (utf) { BACKCHAR(temp); }
1012  #endif  #endif
1013            GETCHARTEST(d, temp);            GETCHARTEST(d, temp);
1014    #ifdef SUPPORT_UCP
1015              if ((md->poptions & PCRE_UCP) != 0)
1016                {
1017                if (d == '_') left_word = TRUE; else
1018                  {
1019                  int cat = UCD_CATEGORY(d);
1020                  left_word = (cat == ucp_L || cat == ucp_N);
1021                  }
1022                }
1023              else
1024    #endif
1025            left_word = d < 256 && (ctypes[d] & ctype_word) != 0;            left_word = d < 256 && (ctypes[d] & ctype_word) != 0;
1026            }            }
1027          else left_word = 0;          else left_word = FALSE;
1028    
1029          if (clen > 0) right_word = c < 256 && (ctypes[c] & ctype_word) != 0;          if (clen > 0)
1030            else right_word = 0;            {
1031    #ifdef SUPPORT_UCP
1032              if ((md->poptions & PCRE_UCP) != 0)
1033                {
1034                if (c == '_') right_word = TRUE; else
1035                  {
1036                  int cat = UCD_CATEGORY(c);
1037                  right_word = (cat == ucp_L || cat == ucp_N);
1038                  }
1039                }
1040              else
1041    #endif
1042              right_word = c < 256 && (ctypes[c] & ctype_word) != 0;
1043              }
1044            else right_word = FALSE;
1045    
1046          if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY))          if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY))
1047            { ADD_ACTIVE(state_offset + 1, 0); }            { ADD_ACTIVE(state_offset + 1, 0); }
# Line 822  for (;;) Line 1060  for (;;)
1060        if (clen > 0)        if (clen > 0)
1061          {          {
1062          BOOL OK;          BOOL OK;
1063            const pcre_uint32 *cp;
1064          const ucd_record * prop = GET_UCD(c);          const ucd_record * prop = GET_UCD(c);
1065          switch(code[1])          switch(code[1])
1066            {            {
# Line 830  for (;;) Line 1069  for (;;)
1069            break;            break;
1070    
1071            case PT_LAMP:            case PT_LAMP:
1072            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || prop->chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1073                   prop->chartype == ucp_Lt;
1074            break;            break;
1075    
1076            case PT_GC:            case PT_GC:
1077            OK = _pcre_ucp_gentype[prop->chartype] == code[2];            OK = PRIV(ucp_gentype)[prop->chartype] == code[2];
1078            break;            break;
1079    
1080            case PT_PC:            case PT_PC:
# Line 845  for (;;) Line 1085  for (;;)
1085            OK = prop->script == code[2];            OK = prop->script == code[2];
1086            break;            break;
1087    
1088              /* These are specials for combination cases. */
1089    
1090              case PT_ALNUM:
1091              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1092                   PRIV(ucp_gentype)[prop->chartype] == ucp_N;
1093              break;
1094    
1095              case PT_SPACE:    /* Perl space */
1096              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1097                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1098              break;
1099    
1100              case PT_PXSPACE:  /* POSIX space */
1101              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1102                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1103                   c == CHAR_FF || c == CHAR_CR;
1104              break;
1105    
1106              case PT_WORD:
1107              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1108                   PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
1109                   c == CHAR_UNDERSCORE;
1110              break;
1111    
1112              case PT_CLIST:
1113              cp = PRIV(ucd_caseless_sets) + prop->caseset;
1114              for (;;)
1115                {
1116                if (c < *cp) { OK = FALSE; break; }
1117                if (c == *cp++) { OK = TRUE; break; }
1118                }
1119              break;
1120    
1121            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1122    
1123            default:            default:
# Line 872  for (;;) Line 1145  for (;;)
1145        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1146        if (clen > 0)        if (clen > 0)
1147          {          {
1148          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1149                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1150                NLBLOCK->nltype == NLTYPE_FIXED &&
1151                NLBLOCK->nllen == 2 &&
1152                c == NLBLOCK->nl[0])
1153              {
1154              could_continue = partial_newline = TRUE;
1155              }
1156            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1157              (c < 256 &&              (c < 256 &&
1158                (d != OP_ANY || !IS_NEWLINE(ptr)) &&                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
1159                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
# Line 895  for (;;) Line 1176  for (;;)
1176        ADD_ACTIVE(state_offset + 2, 0);        ADD_ACTIVE(state_offset + 2, 0);
1177        if (clen > 0)        if (clen > 0)
1178          {          {
1179          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1180                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1181                NLBLOCK->nltype == NLTYPE_FIXED &&
1182                NLBLOCK->nllen == 2 &&
1183                c == NLBLOCK->nl[0])
1184              {
1185              could_continue = partial_newline = TRUE;
1186              }
1187            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1188              (c < 256 &&              (c < 256 &&
1189                (d != OP_ANY || !IS_NEWLINE(ptr)) &&                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
1190                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
# Line 917  for (;;) Line 1206  for (;;)
1206        ADD_ACTIVE(state_offset + 2, 0);        ADD_ACTIVE(state_offset + 2, 0);
1207        if (clen > 0)        if (clen > 0)
1208          {          {
1209          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1210                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1211                NLBLOCK->nltype == NLTYPE_FIXED &&
1212                NLBLOCK->nllen == 2 &&
1213                c == NLBLOCK->nl[0])
1214              {
1215              could_continue = partial_newline = TRUE;
1216              }
1217            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1218              (c < 256 &&              (c < 256 &&
1219                (d != OP_ANY || !IS_NEWLINE(ptr)) &&                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
1220                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
# Line 937  for (;;) Line 1234  for (;;)
1234        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1235        if (clen > 0)        if (clen > 0)
1236          {          {
1237          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1238                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1239                NLBLOCK->nltype == NLTYPE_FIXED &&
1240                NLBLOCK->nllen == 2 &&
1241                c == NLBLOCK->nl[0])
1242              {
1243              could_continue = partial_newline = TRUE;
1244              }
1245            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1246              (c < 256 &&              (c < 256 &&
1247                (d != OP_ANY || !IS_NEWLINE(ptr)) &&                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
1248                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1249            {            {
1250            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
1251              { ADD_NEW(state_offset + 4, 0); }              { ADD_NEW(state_offset + 1 + IMM2_SIZE + 1, 0); }
1252            else            else
1253              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
1254            }            }
# Line 954  for (;;) Line 1259  for (;;)
1259        case OP_TYPEUPTO:        case OP_TYPEUPTO:
1260        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
1261        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
1262        ADD_ACTIVE(state_offset + 4, 0);        ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0);
1263        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1264        if (clen > 0)        if (clen > 0)
1265          {          {
1266          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1267                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1268                NLBLOCK->nltype == NLTYPE_FIXED &&
1269                NLBLOCK->nllen == 2 &&
1270                c == NLBLOCK->nl[0])
1271              {
1272              could_continue = partial_newline = TRUE;
1273              }
1274            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1275              (c < 256 &&              (c < 256 &&
1276                (d != OP_ANY || !IS_NEWLINE(ptr)) &&                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
1277                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
# Line 969  for (;;) Line 1282  for (;;)
1282              next_active_state--;              next_active_state--;
1283              }              }
1284            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
1285              { ADD_NEW(state_offset + 4, 0); }              { ADD_NEW(state_offset + 2 + IMM2_SIZE, 0); }
1286            else            else
1287              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
1288            }            }
# Line 991  for (;;) Line 1304  for (;;)
1304        if (clen > 0)        if (clen > 0)
1305          {          {
1306          BOOL OK;          BOOL OK;
1307            const pcre_uint32 *cp;
1308          const ucd_record * prop = GET_UCD(c);          const ucd_record * prop = GET_UCD(c);
1309          switch(code[2])          switch(code[2])
1310            {            {
# Line 999  for (;;) Line 1313  for (;;)
1313            break;            break;
1314    
1315            case PT_LAMP:            case PT_LAMP:
1316            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || prop->chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1317                prop->chartype == ucp_Lt;
1318            break;            break;
1319    
1320            case PT_GC:            case PT_GC:
1321            OK = _pcre_ucp_gentype[prop->chartype] == code[3];            OK = PRIV(ucp_gentype)[prop->chartype] == code[3];
1322            break;            break;
1323    
1324            case PT_PC:            case PT_PC:
# Line 1014  for (;;) Line 1329  for (;;)
1329            OK = prop->script == code[3];            OK = prop->script == code[3];
1330            break;            break;
1331    
1332              /* These are specials for combination cases. */
1333    
1334              case PT_ALNUM:
1335              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1336                   PRIV(ucp_gentype)[prop->chartype] == ucp_N;
1337              break;
1338    
1339              case PT_SPACE:    /* Perl space */
1340              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1341                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1342              break;
1343    
1344              case PT_PXSPACE:  /* POSIX space */
1345              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1346                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1347                   c == CHAR_FF || c == CHAR_CR;
1348              break;
1349    
1350              case PT_WORD:
1351              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1352                   PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
1353                   c == CHAR_UNDERSCORE;
1354              break;
1355    
1356              case PT_CLIST:
1357              cp = PRIV(ucd_caseless_sets) + prop->caseset;
1358              for (;;)
1359                {
1360                if (c < *cp) { OK = FALSE; break; }
1361                if (c == *cp++) { OK = TRUE; break; }
1362                }
1363              break;
1364    
1365            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1366    
1367            default:            default:
# Line 1040  for (;;) Line 1388  for (;;)
1388        case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS:        case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS:
1389        count = current_state->count;  /* Already matched */        count = current_state->count;  /* Already matched */
1390        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1391        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)        if (clen > 0)
1392          {          {
1393          const uschar *nptr = ptr + clen;          int lgb, rgb;
1394            const pcre_uchar *nptr = ptr + clen;
1395          int ncount = 0;          int ncount = 0;
1396          if (count > 0 && codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS)          if (count > 0 && codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS)
1397            {            {
1398            active_count--;           /* Remove non-match possibility */            active_count--;           /* Remove non-match possibility */
1399            next_active_state--;            next_active_state--;
1400            }            }
1401            lgb = UCD_GRAPHBREAK(c);
1402          while (nptr < end_subject)          while (nptr < end_subject)
1403            {            {
1404            int nd;            dlen = 1;
1405            int ndlen = 1;            if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); }
1406            GETCHARLEN(nd, nptr, ndlen);            rgb = UCD_GRAPHBREAK(d);
1407            if (UCD_CATEGORY(nd) != ucp_M) break;            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
1408            ncount++;            ncount++;
1409            nptr += ndlen;            lgb = rgb;
1410              nptr += dlen;
1411            }            }
1412          count++;          count++;
1413          ADD_NEW_DATA(-state_offset, count, ncount);          ADD_NEW_DATA(-state_offset, count, ncount);
# Line 1075  for (;;) Line 1426  for (;;)
1426          int ncount = 0;          int ncount = 0;
1427          switch (c)          switch (c)
1428            {            {
1429            case 0x000b:            case CHAR_VT:
1430            case 0x000c:            case CHAR_FF:
1431            case 0x0085:            case CHAR_NEL:
1432    #ifndef EBCDIC
1433            case 0x2028:            case 0x2028:
1434            case 0x2029:            case 0x2029:
1435    #endif  /* Not EBCDIC */
1436            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1437            goto ANYNL01;            goto ANYNL01;
1438    
1439            case 0x000d:            case CHAR_CR:
1440            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            if (ptr + 1 < end_subject && ptr[1] == CHAR_LF) ncount = 1;
1441            /* Fall through */            /* Fall through */
1442    
1443            ANYNL01:            ANYNL01:
1444            case 0x000a:            case CHAR_LF:
1445            if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS)            if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS)
1446              {              {
1447              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
# Line 1115  for (;;) Line 1468  for (;;)
1468          BOOL OK;          BOOL OK;
1469          switch (c)          switch (c)
1470            {            {
1471            case 0x000a:            VSPACE_CASES:
           case 0x000b:  
           case 0x000c:  
           case 0x000d:  
           case 0x0085:  
           case 0x2028:  
           case 0x2029:  
1472            OK = TRUE;            OK = TRUE;
1473            break;            break;
1474    
# Line 1154  for (;;) Line 1501  for (;;)
1501          BOOL OK;          BOOL OK;
1502          switch (c)          switch (c)
1503            {            {
1504            case 0x09:      /* HT */            HSPACE_CASES:
           case 0x20:      /* SPACE */  
           case 0xa0:      /* NBSP */  
           case 0x1680:    /* OGHAM SPACE MARK */  
           case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
           case 0x2000:    /* EN QUAD */  
           case 0x2001:    /* EM QUAD */  
           case 0x2002:    /* EN SPACE */  
           case 0x2003:    /* EM SPACE */  
           case 0x2004:    /* THREE-PER-EM SPACE */  
           case 0x2005:    /* FOUR-PER-EM SPACE */  
           case 0x2006:    /* SIX-PER-EM SPACE */  
           case 0x2007:    /* FIGURE SPACE */  
           case 0x2008:    /* PUNCTUATION SPACE */  
           case 0x2009:    /* THIN SPACE */  
           case 0x200A:    /* HAIR SPACE */  
           case 0x202f:    /* NARROW NO-BREAK SPACE */  
           case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
           case 0x3000:    /* IDEOGRAPHIC SPACE */  
1505            OK = TRUE;            OK = TRUE;
1506            break;            break;
1507    
# Line 1213  for (;;) Line 1542  for (;;)
1542        if (clen > 0)        if (clen > 0)
1543          {          {
1544          BOOL OK;          BOOL OK;
1545            const pcre_uint32 *cp;
1546          const ucd_record * prop = GET_UCD(c);          const ucd_record * prop = GET_UCD(c);
1547          switch(code[2])          switch(code[2])
1548            {            {
# Line 1221  for (;;) Line 1551  for (;;)
1551            break;            break;
1552    
1553            case PT_LAMP:            case PT_LAMP:
1554            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || prop->chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1555                prop->chartype == ucp_Lt;
1556            break;            break;
1557    
1558            case PT_GC:            case PT_GC:
1559            OK = _pcre_ucp_gentype[prop->chartype] == code[3];            OK = PRIV(ucp_gentype)[prop->chartype] == code[3];
1560            break;            break;
1561    
1562            case PT_PC:            case PT_PC:
# Line 1236  for (;;) Line 1567  for (;;)
1567            OK = prop->script == code[3];            OK = prop->script == code[3];
1568            break;            break;
1569    
1570              /* These are specials for combination cases. */
1571    
1572              case PT_ALNUM:
1573              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1574                   PRIV(ucp_gentype)[prop->chartype] == ucp_N;
1575              break;
1576    
1577              case PT_SPACE:    /* Perl space */
1578              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1579                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1580              break;
1581    
1582              case PT_PXSPACE:  /* POSIX space */
1583              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1584                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1585                   c == CHAR_FF || c == CHAR_CR;
1586              break;
1587    
1588              case PT_WORD:
1589              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1590                   PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
1591                   c == CHAR_UNDERSCORE;
1592              break;
1593    
1594              case PT_CLIST:
1595              cp = PRIV(ucd_caseless_sets) + prop->caseset;
1596              for (;;)
1597                {
1598                if (c < *cp) { OK = FALSE; break; }
1599                if (c == *cp++) { OK = TRUE; break; }
1600                }
1601              break;
1602    
1603            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1604    
1605            default:            default:
# Line 1271  for (;;) Line 1635  for (;;)
1635        QS2:        QS2:
1636    
1637        ADD_ACTIVE(state_offset + 2, 0);        ADD_ACTIVE(state_offset + 2, 0);
1638        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)        if (clen > 0)
1639          {          {
1640          const uschar *nptr = ptr + clen;          int lgb, rgb;
1641            const pcre_uchar *nptr = ptr + clen;
1642          int ncount = 0;          int ncount = 0;
1643          if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR ||          if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR ||
1644              codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY)              codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY)
# Line 1281  for (;;) Line 1646  for (;;)
1646            active_count--;           /* Remove non-match possibility */            active_count--;           /* Remove non-match possibility */
1647            next_active_state--;            next_active_state--;
1648            }            }
1649            lgb = UCD_GRAPHBREAK(c);
1650          while (nptr < end_subject)          while (nptr < end_subject)
1651            {            {
1652            int nd;            dlen = 1;
1653            int ndlen = 1;            if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); }
1654            GETCHARLEN(nd, nptr, ndlen);            rgb = UCD_GRAPHBREAK(d);
1655            if (UCD_CATEGORY(nd) != ucp_M) break;            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
1656            ncount++;            ncount++;
1657            nptr += ndlen;            lgb = rgb;
1658              nptr += dlen;
1659            }            }
1660          ADD_NEW_DATA(-(state_offset + count), 0, ncount);          ADD_NEW_DATA(-(state_offset + count), 0, ncount);
1661          }          }
# Line 1314  for (;;) Line 1681  for (;;)
1681          int ncount = 0;          int ncount = 0;
1682          switch (c)          switch (c)
1683            {            {
1684            case 0x000b:            case CHAR_VT:
1685            case 0x000c:            case CHAR_FF:
1686            case 0x0085:            case CHAR_NEL:
1687    #ifndef EBCDIC
1688            case 0x2028:            case 0x2028:
1689            case 0x2029:            case 0x2029:
1690    #endif  /* Not EBCDIC */
1691            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1692            goto ANYNL02;            goto ANYNL02;
1693    
1694            case 0x000d:            case CHAR_CR:
1695            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            if (ptr + 1 < end_subject && ptr[1] == CHAR_LF) ncount = 1;
1696            /* Fall through */            /* Fall through */
1697    
1698            ANYNL02:            ANYNL02:
1699            case 0x000a:            case CHAR_LF:
1700            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR ||            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR ||
1701                codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY)                codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY)
1702              {              {
# Line 1362  for (;;) Line 1731  for (;;)
1731          BOOL OK;          BOOL OK;
1732          switch (c)          switch (c)
1733            {            {
1734            case 0x000a:            VSPACE_CASES:
           case 0x000b:  
           case 0x000c:  
           case 0x000d:  
           case 0x0085:  
           case 0x2028:  
           case 0x2029:  
1735            OK = TRUE;            OK = TRUE;
1736            break;            break;
1737    
# Line 1408  for (;;) Line 1771  for (;;)
1771          BOOL OK;          BOOL OK;
1772          switch (c)          switch (c)
1773            {            {
1774            case 0x09:      /* HT */            HSPACE_CASES:
           case 0x20:      /* SPACE */  
           case 0xa0:      /* NBSP */  
           case 0x1680:    /* OGHAM SPACE MARK */  
           case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
           case 0x2000:    /* EN QUAD */  
           case 0x2001:    /* EM QUAD */  
           case 0x2002:    /* EN SPACE */  
           case 0x2003:    /* EM SPACE */  
           case 0x2004:    /* THREE-PER-EM SPACE */  
           case 0x2005:    /* FOUR-PER-EM SPACE */  
           case 0x2006:    /* SIX-PER-EM SPACE */  
           case 0x2007:    /* FIGURE SPACE */  
           case 0x2008:    /* PUNCTUATION SPACE */  
           case 0x2009:    /* THIN SPACE */  
           case 0x200A:    /* HAIR SPACE */  
           case 0x202f:    /* NARROW NO-BREAK SPACE */  
           case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
           case 0x3000:    /* IDEOGRAPHIC SPACE */  
1775            OK = TRUE;            OK = TRUE;
1776            break;            break;
1777    
# Line 1455  for (;;) Line 1800  for (;;)
1800        case OP_PROP_EXTRA + OP_TYPEMINUPTO:        case OP_PROP_EXTRA + OP_TYPEMINUPTO:
1801        case OP_PROP_EXTRA + OP_TYPEPOSUPTO:        case OP_PROP_EXTRA + OP_TYPEPOSUPTO:
1802        if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT)        if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT)
1803          { ADD_ACTIVE(state_offset + 6, 0); }          { ADD_ACTIVE(state_offset + 1 + IMM2_SIZE + 3, 0); }
1804        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1805        if (clen > 0)        if (clen > 0)
1806          {          {
1807          BOOL OK;          BOOL OK;
1808            const pcre_uint32 *cp;
1809          const ucd_record * prop = GET_UCD(c);          const ucd_record * prop = GET_UCD(c);
1810          switch(code[4])          switch(code[1 + IMM2_SIZE + 1])
1811            {            {
1812            case PT_ANY:            case PT_ANY:
1813            OK = TRUE;            OK = TRUE;
1814            break;            break;
1815    
1816            case PT_LAMP:            case PT_LAMP:
1817            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || prop->chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1818                prop->chartype == ucp_Lt;
1819            break;            break;
1820    
1821            case PT_GC:            case PT_GC:
1822            OK = _pcre_ucp_gentype[prop->chartype] == code[5];            OK = PRIV(ucp_gentype)[prop->chartype] == code[1 + IMM2_SIZE + 2];
1823            break;            break;
1824    
1825            case PT_PC:            case PT_PC:
1826            OK = prop->chartype == code[5];            OK = prop->chartype == code[1 + IMM2_SIZE + 2];
1827            break;            break;
1828    
1829            case PT_SC:            case PT_SC:
1830            OK = prop->script == code[5];            OK = prop->script == code[1 + IMM2_SIZE + 2];
1831              break;
1832    
1833              /* These are specials for combination cases. */
1834    
1835              case PT_ALNUM:
1836              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1837                   PRIV(ucp_gentype)[prop->chartype] == ucp_N;
1838            break;            break;
1839    
1840              case PT_SPACE:    /* Perl space */
1841              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1842                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1843              break;
1844    
1845              case PT_PXSPACE:  /* POSIX space */
1846              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1847                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1848                   c == CHAR_FF || c == CHAR_CR;
1849              break;
1850    
1851              case PT_WORD:
1852              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1853                   PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
1854                   c == CHAR_UNDERSCORE;
1855              break;
1856    
1857              case PT_CLIST:
1858              cp = PRIV(ucd_caseless_sets) + prop->caseset;
1859              for (;;)
1860                {
1861                if (c < *cp) { OK = FALSE; break; }
1862                if (c == *cp++) { OK = TRUE; break; }
1863                }
1864              break;
1865    
1866            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1867    
1868            default:            default:
# Line 1498  for (;;) Line 1878  for (;;)
1878              next_active_state--;              next_active_state--;
1879              }              }
1880            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
1881              { ADD_NEW(state_offset + 6, 0); }              { ADD_NEW(state_offset + 1 + IMM2_SIZE + 3, 0); }
1882            else            else
1883              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
1884            }            }
# Line 1511  for (;;) Line 1891  for (;;)
1891        case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO:        case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO:
1892        case OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO:        case OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO:
1893        if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT)        if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT)
1894          { ADD_ACTIVE(state_offset + 4, 0); }          { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); }
1895        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1896        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)        if (clen > 0)
1897          {          {
1898          const uschar *nptr = ptr + clen;          int lgb, rgb;
1899            const pcre_uchar *nptr = ptr + clen;
1900          int ncount = 0;          int ncount = 0;
1901          if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO)          if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO)
1902            {            {
1903            active_count--;           /* Remove non-match possibility */            active_count--;           /* Remove non-match possibility */
1904            next_active_state--;            next_active_state--;
1905            }            }
1906            lgb = UCD_GRAPHBREAK(c);
1907          while (nptr < end_subject)          while (nptr < end_subject)
1908            {            {
1909            int nd;            dlen = 1;
1910            int ndlen = 1;            if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); }
1911            GETCHARLEN(nd, nptr, ndlen);            rgb = UCD_GRAPHBREAK(d);
1912            if (UCD_CATEGORY(nd) != ucp_M) break;            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
1913            ncount++;            ncount++;
1914            nptr += ndlen;            lgb = rgb;
1915              nptr += dlen;
1916            }            }
1917            if (nptr >= end_subject && (md->moptions & PCRE_PARTIAL_HARD) != 0)
1918                reset_could_continue = TRUE;
1919          if (++count >= GET2(code, 1))          if (++count >= GET2(code, 1))
1920            { ADD_NEW_DATA(-(state_offset + 4), 0, ncount); }            { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, ncount); }
1921          else          else
1922            { ADD_NEW_DATA(-state_offset, count, ncount); }            { ADD_NEW_DATA(-state_offset, count, ncount); }
1923          }          }
# Line 1545  for (;;) Line 1930  for (;;)
1930        case OP_ANYNL_EXTRA + OP_TYPEMINUPTO:        case OP_ANYNL_EXTRA + OP_TYPEMINUPTO:
1931        case OP_ANYNL_EXTRA + OP_TYPEPOSUPTO:        case OP_ANYNL_EXTRA + OP_TYPEPOSUPTO:
1932        if (codevalue != OP_ANYNL_EXTRA + OP_TYPEEXACT)        if (codevalue != OP_ANYNL_EXTRA + OP_TYPEEXACT)
1933          { ADD_ACTIVE(state_offset + 4, 0); }          { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); }
1934        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1935        if (clen > 0)        if (clen > 0)
1936          {          {
1937          int ncount = 0;          int ncount = 0;
1938          switch (c)          switch (c)
1939            {            {
1940            case 0x000b:            case CHAR_VT:
1941            case 0x000c:            case CHAR_FF:
1942            case 0x0085:            case CHAR_NEL:
1943    #ifndef EBCDIC
1944            case 0x2028:            case 0x2028:
1945            case 0x2029:            case 0x2029:
1946    #endif  /* Not EBCDIC */
1947            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1948            goto ANYNL03;            goto ANYNL03;
1949    
1950            case 0x000d:            case CHAR_CR:
1951            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            if (ptr + 1 < end_subject && ptr[1] == CHAR_LF) ncount = 1;
1952            /* Fall through */            /* Fall through */
1953    
1954            ANYNL03:            ANYNL03:
1955            case 0x000a:            case CHAR_LF:
1956            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO)            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO)
1957              {              {
1958              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
1959              next_active_state--;              next_active_state--;
1960              }              }
1961            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
1962              { ADD_NEW_DATA(-(state_offset + 4), 0, ncount); }              { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, ncount); }
1963            else            else
1964              { ADD_NEW_DATA(-state_offset, count, ncount); }              { ADD_NEW_DATA(-state_offset, count, ncount); }
1965            break;            break;
# Line 1589  for (;;) Line 1976  for (;;)
1976        case OP_VSPACE_EXTRA + OP_TYPEMINUPTO:        case OP_VSPACE_EXTRA + OP_TYPEMINUPTO:
1977        case OP_VSPACE_EXTRA + OP_TYPEPOSUPTO:        case OP_VSPACE_EXTRA + OP_TYPEPOSUPTO:
1978        if (codevalue != OP_VSPACE_EXTRA + OP_TYPEEXACT)        if (codevalue != OP_VSPACE_EXTRA + OP_TYPEEXACT)
1979          { ADD_ACTIVE(state_offset + 4, 0); }          { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); }
1980        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1981        if (clen > 0)        if (clen > 0)
1982          {          {
1983          BOOL OK;          BOOL OK;
1984          switch (c)          switch (c)
1985            {            {
1986            case 0x000a:            VSPACE_CASES:
           case 0x000b:  
           case 0x000c:  
           case 0x000d:  
           case 0x0085:  
           case 0x2028:  
           case 0x2029:  
1987            OK = TRUE;            OK = TRUE;
1988            break;            break;
1989    
# Line 1618  for (;;) Line 1999  for (;;)
1999              next_active_state--;              next_active_state--;
2000              }              }
2001            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
2002              { ADD_NEW_DATA(-(state_offset + 4), 0, 0); }              { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, 0); }
2003            else            else
2004              { ADD_NEW_DATA(-state_offset, count, 0); }              { ADD_NEW_DATA(-state_offset, count, 0); }
2005            }            }
# Line 1631  for (;;) Line 2012  for (;;)
2012        case OP_HSPACE_EXTRA + OP_TYPEMINUPTO:        case OP_HSPACE_EXTRA + OP_TYPEMINUPTO:
2013        case OP_HSPACE_EXTRA + OP_TYPEPOSUPTO:        case OP_HSPACE_EXTRA + OP_TYPEPOSUPTO:
2014        if (codevalue != OP_HSPACE_EXTRA + OP_TYPEEXACT)        if (codevalue != OP_HSPACE_EXTRA + OP_TYPEEXACT)
2015          { ADD_ACTIVE(state_offset + 4, 0); }          { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); }
2016        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
2017        if (clen > 0)        if (clen > 0)
2018          {          {
2019          BOOL OK;          BOOL OK;
2020          switch (c)          switch (c)
2021            {            {
2022            case 0x09:      /* HT */            HSPACE_CASES:
           case 0x20:      /* SPACE */  
           case 0xa0:      /* NBSP */  
           case 0x1680:    /* OGHAM SPACE MARK */  
           case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
           case 0x2000:    /* EN QUAD */  
           case 0x2001:    /* EM QUAD */  
           case 0x2002:    /* EN SPACE */  
           case 0x2003:    /* EM SPACE */  
           case 0x2004:    /* THREE-PER-EM SPACE */  
           case 0x2005:    /* FOUR-PER-EM SPACE */  
           case 0x2006:    /* SIX-PER-EM SPACE */  
           case 0x2007:    /* FIGURE SPACE */  
           case 0x2008:    /* PUNCTUATION SPACE */  
           case 0x2009:    /* THIN SPACE */  
           case 0x200A:    /* HAIR SPACE */  
           case 0x202f:    /* NARROW NO-BREAK SPACE */  
           case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
           case 0x3000:    /* IDEOGRAPHIC SPACE */  
2023            OK = TRUE;            OK = TRUE;
2024            break;            break;
2025    
# Line 1673  for (;;) Line 2036  for (;;)
2036              next_active_state--;              next_active_state--;
2037              }              }
2038            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
2039              { ADD_NEW_DATA(-(state_offset + 4), 0, 0); }              { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, 0); }
2040            else            else
2041              { ADD_NEW_DATA(-state_offset, count, 0); }              { ADD_NEW_DATA(-state_offset, count, 0); }
2042            }            }
# Line 1692  for (;;) Line 2055  for (;;)
2055        break;        break;
2056    
2057        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2058        case OP_CHARNC:        case OP_CHARI:
2059        if (clen == 0) break;        if (clen == 0) break;
2060    
2061  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2062        if (utf8)        if (utf)
2063          {          {
2064          if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else          if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else
2065            {            {
2066            unsigned int othercase;            unsigned int othercase;
2067            if (c < 128) othercase = fcc[c]; else            if (c < 128)
2068                othercase = fcc[c];
2069            /* If we have Unicode property support, we can use it to test the            else
2070            other case of the character. */              /* If we have Unicode property support, we can use it to test the
2071                other case of the character. */
2072  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2073            othercase = UCD_OTHERCASE(c);              othercase = UCD_OTHERCASE(c);
2074  #else  #else
2075            othercase = NOTACHAR;              othercase = NOTACHAR;
2076  #endif  #endif
2077    
2078            if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); }            if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); }
2079            }            }
2080          }          }
2081        else        else
2082  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2083          /* Not UTF mode */
       /* Non-UTF-8 mode */  
2084          {          {
2085          if (lcc[c] == lcc[d]) { ADD_NEW(state_offset + 2, 0); }          if (TABLE_GET(c, lcc, c) == TABLE_GET(d, lcc, d))
2086              { ADD_NEW(state_offset + 2, 0); }
2087          }          }
2088        break;        break;
2089    
# Line 1732  for (;;) Line 2095  for (;;)
2095        to wait for them to pass before continuing. */        to wait for them to pass before continuing. */
2096    
2097        case OP_EXTUNI:        case OP_EXTUNI:
2098        if (clen > 0 && UCD_CATEGORY(c) != ucp_M)        if (clen > 0)
2099          {          {
2100          const uschar *nptr = ptr + clen;          int lgb, rgb;
2101            const pcre_uchar *nptr = ptr + clen;
2102          int ncount = 0;          int ncount = 0;
2103            lgb = UCD_GRAPHBREAK(c);
2104          while (nptr < end_subject)          while (nptr < end_subject)
2105            {            {
2106            int nclen = 1;            dlen = 1;
2107            GETCHARLEN(c, nptr, nclen);            if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); }
2108            if (UCD_CATEGORY(c) != ucp_M) break;            rgb = UCD_GRAPHBREAK(d);
2109              if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2110            ncount++;            ncount++;
2111            nptr += nclen;            lgb = rgb;
2112              nptr += dlen;
2113            }            }
2114            if (nptr >= end_subject && (md->moptions & PCRE_PARTIAL_HARD) != 0)
2115                reset_could_continue = TRUE;
2116          ADD_NEW_DATA(-(state_offset + 1), 0, ncount);          ADD_NEW_DATA(-(state_offset + 1), 0, ncount);
2117          }          }
2118        break;        break;
# Line 1757  for (;;) Line 2126  for (;;)
2126        case OP_ANYNL:        case OP_ANYNL:
2127        if (clen > 0) switch(c)        if (clen > 0) switch(c)
2128          {          {
2129          case 0x000b:          case CHAR_VT:
2130          case 0x000c:          case CHAR_FF:
2131          case 0x0085:          case CHAR_NEL:
2132    #ifndef EBCDIC
2133          case 0x2028:          case 0x2028:
2134          case 0x2029:          case 0x2029:
2135    #endif  /* Not EBCDIC */
2136          if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;          if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
2137    
2138          case 0x000a:          case CHAR_LF:
2139          ADD_NEW(state_offset + 1, 0);          ADD_NEW(state_offset + 1, 0);
2140          break;          break;
2141    
2142          case 0x000d:          case CHAR_CR:
2143          if (ptr + 1 < end_subject && ptr[1] == 0x0a)          if (ptr + 1 >= end_subject)
2144              {
2145              ADD_NEW(state_offset + 1, 0);
2146              if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
2147                reset_could_continue = TRUE;
2148              }
2149            else if (ptr[1] == CHAR_LF)
2150            {            {
2151            ADD_NEW_DATA(-(state_offset + 1), 0, 1);            ADD_NEW_DATA(-(state_offset + 1), 0, 1);
2152            }            }
# Line 1785  for (;;) Line 2162  for (;;)
2162        case OP_NOT_VSPACE:        case OP_NOT_VSPACE:
2163        if (clen > 0) switch(c)        if (clen > 0) switch(c)
2164          {          {
2165          case 0x000a:          VSPACE_CASES:
         case 0x000b:  
         case 0x000c:  
         case 0x000d:  
         case 0x0085:  
         case 0x2028:  
         case 0x2029:  
2166          break;          break;
2167    
2168          default:          default:
# Line 1804  for (;;) Line 2175  for (;;)
2175        case OP_VSPACE:        case OP_VSPACE:
2176        if (clen > 0) switch(c)        if (clen > 0) switch(c)
2177          {          {
2178          case 0x000a:          VSPACE_CASES:
         case 0x000b:  
         case 0x000c:  
         case 0x000d:  
         case 0x0085:  
         case 0x2028:  
         case 0x2029:  
2179          ADD_NEW(state_offset + 1, 0);          ADD_NEW(state_offset + 1, 0);
2180          break;          break;
2181    
2182          default: break;          default:
2183            break;
2184          }          }
2185        break;        break;
2186    
# Line 1822  for (;;) Line 2188  for (;;)
2188        case OP_NOT_HSPACE:        case OP_NOT_HSPACE:
2189        if (clen > 0) switch(c)        if (clen > 0) switch(c)
2190          {          {
2191          case 0x09:      /* HT */          HSPACE_CASES:
         case 0x20:      /* SPACE */  
         case 0xa0:      /* NBSP */  
         case 0x1680:    /* OGHAM SPACE MARK */  
         case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
         case 0x2000:    /* EN QUAD */  
         case 0x2001:    /* EM QUAD */  
         case 0x2002:    /* EN SPACE */  
         case 0x2003:    /* EM SPACE */  
         case 0x2004:    /* THREE-PER-EM SPACE */  
         case 0x2005:    /* FOUR-PER-EM SPACE */  
         case 0x2006:    /* SIX-PER-EM SPACE */  
         case 0x2007:    /* FIGURE SPACE */  
         case 0x2008:    /* PUNCTUATION SPACE */  
         case 0x2009:    /* THIN SPACE */  
         case 0x200A:    /* HAIR SPACE */  
         case 0x202f:    /* NARROW NO-BREAK SPACE */  
         case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
         case 0x3000:    /* IDEOGRAPHIC SPACE */  
2192          break;          break;
2193    
2194          default:          default:
# Line 1853  for (;;) Line 2201  for (;;)
2201        case OP_HSPACE:        case OP_HSPACE:
2202        if (clen > 0) switch(c)        if (clen > 0) switch(c)
2203          {          {
2204          case 0x09:      /* HT */          HSPACE_CASES:
         case 0x20:      /* SPACE */  
         case 0xa0:      /* NBSP */  
         case 0x1680:    /* OGHAM SPACE MARK */  
         case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
         case 0x2000:    /* EN QUAD */  
         case 0x2001:    /* EM QUAD */  
         case 0x2002:    /* EN SPACE */  
         case 0x2003:    /* EM SPACE */  
         case 0x2004:    /* THREE-PER-EM SPACE */  
         case 0x2005:    /* FOUR-PER-EM SPACE */  
         case 0x2006:    /* SIX-PER-EM SPACE */  
         case 0x2007:    /* FIGURE SPACE */  
         case 0x2008:    /* PUNCTUATION SPACE */  
         case 0x2009:    /* THIN SPACE */  
         case 0x200A:    /* HAIR SPACE */  
         case 0x202f:    /* NARROW NO-BREAK SPACE */  
         case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
         case 0x3000:    /* IDEOGRAPHIC SPACE */  
2205          ADD_NEW(state_offset + 1, 0);          ADD_NEW(state_offset + 1, 0);
2206          break;          break;
2207    
2208            default:
2209            break;
2210          }          }
2211        break;        break;
2212    
2213        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2214        /* Match a negated single character. This is only used for one-byte        /* Match a negated single character casefully. */
       characters, that is, we know that d < 256. The character we are  
       checking (c) can be multibyte. */  
2215    
2216        case OP_NOT:        case OP_NOT:
2217          if (clen > 0 && c != d) { ADD_NEW(state_offset + dlen + 1, 0); }
2218          break;
2219    
2220          /*-----------------------------------------------------------------*/
2221          /* Match a negated single character caselessly. */
2222    
2223          case OP_NOTI:
2224        if (clen > 0)        if (clen > 0)
2225          {          {
2226          unsigned int otherd = ((ims & PCRE_CASELESS) != 0)? fcc[d] : d;          unsigned int otherd;
2227          if (c != d && c != otherd) { ADD_NEW(state_offset + dlen + 1, 0); }  #ifdef SUPPORT_UTF
2228            if (utf && d >= 128)
2229              {
2230    #ifdef SUPPORT_UCP
2231              otherd = UCD_OTHERCASE(d);
2232    #endif  /* SUPPORT_UCP */
2233              }
2234            else
2235    #endif  /* SUPPORT_UTF */
2236            otherd = TABLE_GET(d, fcc, d);
2237            if (c != d && c != otherd)
2238              { ADD_NEW(state_offset + dlen + 1, 0); }
2239          }          }
2240        break;        break;
2241    
2242        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2243          case OP_PLUSI:
2244          case OP_MINPLUSI:
2245          case OP_POSPLUSI:
2246          case OP_NOTPLUSI:
2247          case OP_NOTMINPLUSI:
2248          case OP_NOTPOSPLUSI:
2249          caseless = TRUE;
2250          codevalue -= OP_STARI - OP_STAR;
2251    
2252          /* Fall through */
2253        case OP_PLUS:        case OP_PLUS:
2254        case OP_MINPLUS:        case OP_MINPLUS:
2255        case OP_POSPLUS:        case OP_POSPLUS:
# Line 1902  for (;;) Line 2261  for (;;)
2261        if (clen > 0)        if (clen > 0)
2262          {          {
2263          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2264          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2265            {            {
2266  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2267            if (utf8 && d >= 128)            if (utf && d >= 128)
2268              {              {
2269  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2270              otherd = UCD_OTHERCASE(d);              otherd = UCD_OTHERCASE(d);
2271  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2272              }              }
2273            else            else
2274  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2275            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2276            }            }
2277          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2278            {            {
# Line 1930  for (;;) Line 2289  for (;;)
2289        break;        break;
2290    
2291        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2292          case OP_QUERYI:
2293          case OP_MINQUERYI:
2294          case OP_POSQUERYI:
2295          case OP_NOTQUERYI:
2296          case OP_NOTMINQUERYI:
2297          case OP_NOTPOSQUERYI:
2298          caseless = TRUE;
2299          codevalue -= OP_STARI - OP_STAR;
2300          /* Fall through */
2301        case OP_QUERY:        case OP_QUERY:
2302        case OP_MINQUERY:        case OP_MINQUERY:
2303        case OP_POSQUERY:        case OP_POSQUERY:
# Line 1940  for (;;) Line 2308  for (;;)
2308        if (clen > 0)        if (clen > 0)
2309          {          {
2310          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2311          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2312            {            {
2313  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2314            if (utf8 && d >= 128)            if (utf && d >= 128)
2315              {              {
2316  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2317              otherd = UCD_OTHERCASE(d);              otherd = UCD_OTHERCASE(d);
2318  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2319              }              }
2320            else            else
2321  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2322            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2323            }            }
2324          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2325            {            {
# Line 1966  for (;;) Line 2334  for (;;)
2334        break;        break;
2335    
2336        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2337          case OP_STARI:
2338          case OP_MINSTARI:
2339          case OP_POSSTARI:
2340          case OP_NOTSTARI:
2341          case OP_NOTMINSTARI:
2342          case OP_NOTPOSSTARI:
2343          caseless = TRUE;
2344          codevalue -= OP_STARI - OP_STAR;
2345          /* Fall through */
2346        case OP_STAR:        case OP_STAR:
2347        case OP_MINSTAR:        case OP_MINSTAR:
2348        case OP_POSSTAR:        case OP_POSSTAR:
# Line 1976  for (;;) Line 2353  for (;;)
2353        if (clen > 0)        if (clen > 0)
2354          {          {
2355          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2356          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2357            {            {
2358  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2359            if (utf8 && d >= 128)            if (utf && d >= 128)
2360              {              {
2361  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2362              otherd = UCD_OTHERCASE(d);              otherd = UCD_OTHERCASE(d);
2363  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2364              }              }
2365            else            else
2366  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2367            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2368            }            }
2369          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2370            {            {
# Line 2002  for (;;) Line 2379  for (;;)
2379        break;        break;
2380    
2381        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2382          case OP_EXACTI:
2383          case OP_NOTEXACTI:
2384          caseless = TRUE;
2385          codevalue -= OP_STARI - OP_STAR;
2386          /* Fall through */
2387        case OP_EXACT:        case OP_EXACT:
2388        case OP_NOTEXACT:        case OP_NOTEXACT:
2389        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
2390        if (clen > 0)        if (clen > 0)
2391          {          {
2392          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2393          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2394            {            {
2395  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2396            if (utf8 && d >= 128)            if (utf && d >= 128)
2397              {              {
2398  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2399              otherd = UCD_OTHERCASE(d);              otherd = UCD_OTHERCASE(d);
2400  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2401              }              }
2402            else            else
2403  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2404            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2405            }            }
2406          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2407            {            {
2408            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
2409              { ADD_NEW(state_offset + dlen + 3, 0); }              { ADD_NEW(state_offset + dlen + 1 + IMM2_SIZE, 0); }
2410            else            else
2411              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
2412            }            }
# Line 2032  for (;;) Line 2414  for (;;)
2414        break;        break;
2415    
2416        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2417          case OP_UPTOI:
2418          case OP_MINUPTOI:
2419          case OP_POSUPTOI:
2420          case OP_NOTUPTOI:
2421          case OP_NOTMINUPTOI:
2422          case OP_NOTPOSUPTOI:
2423          caseless = TRUE;
2424          codevalue -= OP_STARI - OP_STAR;
2425          /* Fall through */
2426        case OP_UPTO:        case OP_UPTO:
2427        case OP_MINUPTO:        case OP_MINUPTO:
2428        case OP_POSUPTO:        case OP_POSUPTO:
2429        case OP_NOTUPTO:        case OP_NOTUPTO:
2430        case OP_NOTMINUPTO:        case OP_NOTMINUPTO:
2431        case OP_NOTPOSUPTO:        case OP_NOTPOSUPTO:
2432        ADD_ACTIVE(state_offset + dlen + 3, 0);        ADD_ACTIVE(state_offset + dlen + 1 + IMM2_SIZE, 0);
2433        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
2434        if (clen > 0)        if (clen > 0)
2435          {          {
2436          unsigned int otherd = NOTACHAR;          unsigned int otherd = NOTACHAR;
2437          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2438            {            {
2439  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2440            if (utf8 && d >= 128)            if (utf && d >= 128)
2441              {              {
2442  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2443              otherd = UCD_OTHERCASE(d);              otherd = UCD_OTHERCASE(d);
2444  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2445              }              }
2446            else            else
2447  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2448            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2449            }            }
2450          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2451            {            {
# Line 2064  for (;;) Line 2455  for (;;)
2455              next_active_state--;              next_active_state--;
2456              }              }
2457            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
2458              { ADD_NEW(state_offset + dlen + 3, 0); }              { ADD_NEW(state_offset + dlen + 1 + IMM2_SIZE, 0); }
2459            else            else
2460              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
2461            }            }
# Line 2081  for (;;) Line 2472  for (;;)
2472          {          {
2473          BOOL isinclass = FALSE;          BOOL isinclass = FALSE;
2474          int next_state_offset;          int next_state_offset;
2475          const uschar *ecode;          const pcre_uchar *ecode;
2476    
2477          /* For a simple class, there is always just a 32-byte table, and we          /* For a simple class, there is always just a 32-byte table, and we
2478          can set isinclass from it. */          can set isinclass from it. */
2479    
2480          if (codevalue != OP_XCLASS)          if (codevalue != OP_XCLASS)
2481            {            {
2482            ecode = code + 33;            ecode = code + 1 + (32 / sizeof(pcre_uchar));
2483            if (clen > 0)            if (clen > 0)
2484              {              {
2485              isinclass = (c > 255)? (codevalue == OP_NCLASS) :              isinclass = (c > 255)? (codevalue == OP_NCLASS) :
2486                ((code[1 + c/8] & (1 << (c&7))) != 0);                ((((pcre_uint8 *)(code + 1))[c/8] & (1 << (c&7))) != 0);
2487              }              }
2488            }            }
2489    
# Line 2103  for (;;) Line 2494  for (;;)
2494          else          else
2495           {           {
2496           ecode = code + GET(code, 1);           ecode = code + GET(code, 1);
2497           if (clen > 0) isinclass = _pcre_xclass(c, code + 1 + LINK_SIZE);           if (clen > 0) isinclass = PRIV(xclass)(c, code + 1 + LINK_SIZE, utf);
2498           }           }
2499    
2500          /* At this point, isinclass is set for all kinds of class, and ecode          /* At this point, isinclass is set for all kinds of class, and ecode
2501          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
2502          quantifier, this is where it will be. */          quantifier, this is where it will be. */
2503    
2504          next_state_offset = ecode - start_code;          next_state_offset = (int)(ecode - start_code);
2505    
2506          switch (*ecode)          switch (*ecode)
2507            {            {
# Line 2137  for (;;) Line 2528  for (;;)
2528            case OP_CRMINRANGE:            case OP_CRMINRANGE:
2529            count = current_state->count;  /* Already matched */            count = current_state->count;  /* Already matched */
2530            if (count >= GET2(ecode, 1))            if (count >= GET2(ecode, 1))
2531              { ADD_ACTIVE(next_state_offset + 5, 0); }              { ADD_ACTIVE(next_state_offset + 1 + 2 * IMM2_SIZE, 0); }
2532            if (isinclass)            if (isinclass)
2533              {              {
2534              int max = GET2(ecode, 3);              int max = GET2(ecode, 1 + IMM2_SIZE);
2535              if (++count >= max && max != 0)   /* Max 0 => no limit */              if (++count >= max && max != 0)   /* Max 0 => no limit */
2536                { ADD_NEW(next_state_offset + 5, 0); }                { ADD_NEW(next_state_offset + 1 + 2 * IMM2_SIZE, 0); }
2537              else              else
2538                { ADD_NEW(state_offset, count); }                { ADD_NEW(state_offset, count); }
2539              }              }
# Line 2162  for (;;) Line 2553  for (;;)
2553        though the other "backtracking verbs" are not supported. */        though the other "backtracking verbs" are not supported. */
2554    
2555        case OP_FAIL:        case OP_FAIL:
2556          forced_fail++;    /* Count FAILs for multiple states */
2557        break;        break;
2558    
2559        case OP_ASSERT:        case OP_ASSERT:
# Line 2172  for (;;) Line 2564  for (;;)
2564          int rc;          int rc;
2565          int local_offsets[2];          int local_offsets[2];
2566          int local_workspace[1000];          int local_workspace[1000];
2567          const uschar *endasscode = code + GET(code, 1);          const pcre_uchar *endasscode = code + GET(code, 1);
2568    
2569          while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1);          while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1);
2570    
# Line 2180  for (;;) Line 2572  for (;;)
2572            md,                                   /* static match data */            md,                                   /* static match data */
2573            code,                                 /* this subexpression's code */            code,                                 /* this subexpression's code */
2574            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2575            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2576            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2577            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2578            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2579            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
2580            ims,                                  /* the current ims flags */            rlevel);                              /* function recursion level */
           rlevel,                               /* function recursion level */  
           recursing);                           /* pass on regex recursion */  
2581    
2582            if (rc == PCRE_ERROR_DFA_UITEM) return rc;
2583          if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK))          if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK))
2584              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }              { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); }
2585          }          }
2586        break;        break;
2587    
# Line 2210  for (;;) Line 2601  for (;;)
2601          if (code[LINK_SIZE+1] == OP_CALLOUT)          if (code[LINK_SIZE+1] == OP_CALLOUT)
2602            {            {
2603            rrc = 0;            rrc = 0;
2604            if (pcre_callout != NULL)            if (PUBL(callout) != NULL)
2605              {              {
2606              pcre_callout_block cb;              PUBL(callout_block) cb;
2607              cb.version          = 1;   /* Version 1 of the callout block */              cb.version          = 1;   /* Version 1 of the callout block */
2608              cb.callout_number   = code[LINK_SIZE+2];              cb.callout_number   = code[LINK_SIZE+2];
2609              cb.offset_vector    = offsets;              cb.offset_vector    = offsets;
2610    #ifdef COMPILE_PCRE8
2611              cb.subject          = (PCRE_SPTR)start_subject;              cb.subject          = (PCRE_SPTR)start_subject;
2612              cb.subject_length   = end_subject - start_subject;  #else
2613              cb.start_match      = current_subject - start_subject;              cb.subject          = (PCRE_SPTR16)start_subject;
2614              cb.current_position = ptr - start_subject;  #endif
2615                cb.subject_length   = (int)(end_subject - start_subject);
2616                cb.start_match      = (int)(current_subject - start_subject);
2617                cb.current_position = (int)(ptr - start_subject);
2618              cb.pattern_position = GET(code, LINK_SIZE + 3);              cb.pattern_position = GET(code, LINK_SIZE + 3);
2619              cb.next_item_length = GET(code, 3 + 2*LINK_SIZE);              cb.next_item_length = GET(code, 3 + 2*LINK_SIZE);
2620              cb.capture_top      = 1;              cb.capture_top      = 1;
2621              cb.capture_last     = -1;              cb.capture_last     = -1;
2622              cb.callout_data     = md->callout_data;              cb.callout_data     = md->callout_data;
2623              if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc;   /* Abandon */              cb.mark             = NULL;   /* No (*MARK) support */
2624                if ((rrc = (*PUBL(callout))(&cb)) < 0) return rrc;   /* Abandon */
2625              }              }
2626            if (rrc > 0) break;                      /* Fail this thread */            if (rrc > 0) break;                      /* Fail this thread */
2627            code += _pcre_OP_lengths[OP_CALLOUT];    /* Skip callout data */            code += PRIV(OP_lengths)[OP_CALLOUT];    /* Skip callout data */
2628            }            }
2629    
2630          condcode = code[LINK_SIZE+1];          condcode = code[LINK_SIZE+1];
2631    
2632          /* Back reference conditions are not supported */          /* Back reference conditions are not supported */
2633    
2634          if (condcode == OP_CREF) return PCRE_ERROR_DFA_UCOND;          if (condcode == OP_CREF || condcode == OP_NCREF)
2635              return PCRE_ERROR_DFA_UCOND;
2636    
2637          /* The DEFINE condition is always false */          /* The DEFINE condition is always false */
2638    
# Line 2246  for (;;) Line 2643  for (;;)
2643          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
2644          recursed groups. */          recursed groups. */
2645    
2646          else if (condcode == OP_RREF)          else if (condcode == OP_RREF || condcode == OP_NRREF)
2647            {            {
2648            int value = GET2(code, LINK_SIZE+2);            int value = GET2(code, LINK_SIZE + 2);
2649            if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND;            if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND;
2650            if (recursing > 0)            if (md->recursive != NULL)
2651              { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); }              { ADD_ACTIVE(state_offset + LINK_SIZE + 2 + IMM2_SIZE, 0); }
2652            else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }            else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
2653            }            }
2654    
# Line 2260  for (;;) Line 2657  for (;;)
2657          else          else
2658            {            {
2659            int rc;            int rc;
2660            const uschar *asscode = code + LINK_SIZE + 1;            const pcre_uchar *asscode = code + LINK_SIZE + 1;
2661            const uschar *endasscode = asscode + GET(asscode, 1);            const pcre_uchar *endasscode = asscode + GET(asscode, 1);
2662    
2663            while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1);            while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1);
2664    
# Line 2269  for (;;) Line 2666  for (;;)
2666              md,                                   /* fixed match data */              md,                                   /* fixed match data */
2667              asscode,                              /* this subexpression's code */              asscode,                              /* this subexpression's code */
2668              ptr,                                  /* where we currently are */              ptr,                                  /* where we currently are */
2669              ptr - start_subject,                  /* start offset */              (int)(ptr - start_subject),           /* start offset */
2670              local_offsets,                        /* offset vector */              local_offsets,                        /* offset vector */
2671              sizeof(local_offsets)/sizeof(int),    /* size of same */              sizeof(local_offsets)/sizeof(int),    /* size of same */
2672              local_workspace,                      /* workspace vector */              local_workspace,                      /* workspace vector */
2673              sizeof(local_workspace)/sizeof(int),  /* size of same */              sizeof(local_workspace)/sizeof(int),  /* size of same */
2674              ims,                                  /* the current ims flags */              rlevel);                              /* function recursion level */
             rlevel,                               /* function recursion level */  
             recursing);                           /* pass on regex recursion */  
2675    
2676              if (rc == PCRE_ERROR_DFA_UITEM) return rc;
2677            if ((rc >= 0) ==            if ((rc >= 0) ==
2678                  (condcode == OP_ASSERT || condcode == OP_ASSERTBACK))                  (condcode == OP_ASSERT || condcode == OP_ASSERTBACK))
2679              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }              { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); }
2680            else            else
2681              { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }              { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
2682            }            }
# Line 2290  for (;;) Line 2686  for (;;)
2686        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2687        case OP_RECURSE:        case OP_RECURSE:
2688          {          {
2689            dfa_recursion_info *ri;
2690          int local_offsets[1000];          int local_offsets[1000];
2691          int local_workspace[1000];          int local_workspace[1000];
2692            const pcre_uchar *callpat = start_code + GET(code, 1);
2693            int recno = (callpat == md->start_code)? 0 :
2694              GET2(callpat, 1 + LINK_SIZE);
2695          int rc;          int rc;
2696    
2697          DPRINTF(("%.*sStarting regex recursion %d\n", rlevel*2-2, SP,          DPRINTF(("%.*sStarting regex recursion\n", rlevel*2-2, SP));
2698            recursing + 1));  
2699            /* Check for repeating a recursion without advancing the subject
2700            pointer. This should catch convoluted mutual recursions. (Some simple
2701            cases are caught at compile time.) */
2702    
2703            for (ri = md->recursive; ri != NULL; ri = ri->prevrec)
2704              if (recno == ri->group_num && ptr == ri->subject_position)
2705                return PCRE_ERROR_RECURSELOOP;
2706    
2707            /* Remember this recursion and where we started it so as to
2708            catch infinite loops. */
2709    
2710            new_recursive.group_num = recno;
2711            new_recursive.subject_position = ptr;
2712            new_recursive.prevrec = md->recursive;
2713            md->recursive = &new_recursive;
2714    
2715          rc = internal_dfa_exec(          rc = internal_dfa_exec(
2716            md,                                   /* fixed match data */            md,                                   /* fixed match data */
2717            start_code + GET(code, 1),            /* this subexpression's code */            callpat,                              /* this subexpression's code */
2718            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2719            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2720            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2721            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2722            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2723            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
2724            ims,                                  /* the current ims flags */            rlevel);                              /* function recursion level */
2725            rlevel,                               /* function recursion level */  
2726            recursing + 1);                       /* regex recurse level */          md->recursive = new_recursive.prevrec;  /* Done this recursion */
2727    
2728          DPRINTF(("%.*sReturn from regex recursion %d: rc=%d\n", rlevel*2-2, SP,          DPRINTF(("%.*sReturn from regex recursion: rc=%d\n", rlevel*2-2, SP,
2729            recursing + 1, rc));            rc));
2730    
2731          /* Ran out of internal offsets */          /* Ran out of internal offsets */
2732    
# Line 2325  for (;;) Line 2740  for (;;)
2740            {            {
2741            for (rc = rc*2 - 2; rc >= 0; rc -= 2)            for (rc = rc*2 - 2; rc >= 0; rc -= 2)
2742              {              {
             const uschar *p = start_subject + local_offsets[rc];  
             const uschar *pp = start_subject + local_offsets[rc+1];  
2743              int charcount = local_offsets[rc+1] - local_offsets[rc];              int charcount = local_offsets[rc+1] - local_offsets[rc];
2744              while (p < pp) if ((*p++ & 0xc0) == 0x80) charcount--;  #ifdef SUPPORT_UTF
2745                if (utf)
2746                  {
2747                  const pcre_uchar *p = start_subject + local_offsets[rc];
2748                  const pcre_uchar *pp = start_subject + local_offsets[rc+1];
2749                  while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--;
2750                  }
2751    #endif
2752              if (charcount > 0)              if (charcount > 0)
2753                {                {
2754                ADD_NEW_DATA(-(state_offset + LINK_SIZE + 1), 0, (charcount - 1));                ADD_NEW_DATA(-(state_offset + LINK_SIZE + 1), 0, (charcount - 1));
# Line 2344  for (;;) Line 2764  for (;;)
2764        break;        break;
2765    
2766        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2767          case OP_BRAPOS:
2768          case OP_SBRAPOS:
2769          case OP_CBRAPOS:
2770          case OP_SCBRAPOS:
2771          case OP_BRAPOSZERO:
2772            {
2773            int charcount, matched_count;
2774            const pcre_uchar *local_ptr = ptr;
2775            BOOL allow_zero;
2776    
2777            if (codevalue == OP_BRAPOSZERO)
2778              {
2779              allow_zero = TRUE;
2780              codevalue = *(++code);  /* Codevalue will be one of above BRAs */
2781              }
2782            else allow_zero = FALSE;
2783    
2784            /* Loop to match the subpattern as many times as possible as if it were
2785            a complete pattern. */
2786    
2787            for (matched_count = 0;; matched_count++)
2788              {
2789              int local_offsets[2];
2790              int local_workspace[1000];
2791    
2792              int rc = internal_dfa_exec(
2793                md,                                   /* fixed match data */
2794                code,                                 /* this subexpression's code */
2795                local_ptr,                            /* where we currently are */
2796                (int)(ptr - start_subject),           /* start offset */
2797                local_offsets,                        /* offset vector */
2798                sizeof(local_offsets)/sizeof(int),    /* size of same */
2799                local_workspace,                      /* workspace vector */
2800                sizeof(local_workspace)/sizeof(int),  /* size of same */
2801                rlevel);                              /* function recursion level */
2802    
2803              /* Failed to match */
2804    
2805              if (rc < 0)
2806                {
2807                if (rc != PCRE_ERROR_NOMATCH) return rc;
2808                break;
2809                }
2810    
2811              /* Matched: break the loop if zero characters matched. */
2812    
2813              charcount = local_offsets[1] - local_offsets[0];
2814              if (charcount == 0) break;
2815              local_ptr += charcount;    /* Advance temporary position ptr */
2816              }
2817    
2818            /* At this point we have matched the subpattern matched_count
2819            times, and local_ptr is pointing to the character after the end of the
2820            last match. */
2821    
2822            if (matched_count > 0 || allow_zero)
2823              {
2824              const pcre_uchar *end_subpattern = code;
2825              int next_state_offset;
2826    
2827              do { end_subpattern += GET(end_subpattern, 1); }
2828                while (*end_subpattern == OP_ALT);
2829              next_state_offset =
2830                (int)(end_subpattern - start_code + LINK_SIZE + 1);
2831    
2832              /* Optimization: if there are no more active states, and there
2833              are no new states yet set up, then skip over the subject string
2834              right here, to save looping. Otherwise, set up the new state to swing
2835              into action when the end of the matched substring is reached. */
2836    
2837              if (i + 1 >= active_count && new_count == 0)
2838                {
2839                ptr = local_ptr;
2840                clen = 0;
2841                ADD_NEW(next_state_offset, 0);
2842                }
2843              else
2844                {
2845                const pcre_uchar *p = ptr;
2846                const pcre_uchar *pp = local_ptr;
2847                charcount = (int)(pp - p);
2848    #ifdef SUPPORT_UTF
2849                if (utf) while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--;
2850    #endif
2851                ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1));
2852                }
2853              }
2854            }
2855          break;
2856    
2857          /*-----------------------------------------------------------------*/
2858        case OP_ONCE:        case OP_ONCE:
2859          case OP_ONCE_NC:
2860          {          {
2861          int local_offsets[2];          int local_offsets[2];
2862          int local_workspace[1000];          int local_workspace[1000];
# Line 2353  for (;;) Line 2865  for (;;)
2865            md,                                   /* fixed match data */            md,                                   /* fixed match data */
2866            code,                                 /* this subexpression's code */            code,                                 /* this subexpression's code */
2867            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2868            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2869            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2870            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2871            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2872            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
2873            ims,                                  /* the current ims flags */            rlevel);                              /* function recursion level */
           rlevel,                               /* function recursion level */  
           recursing);                           /* pass on regex recursion */  
2874    
2875          if (rc >= 0)          if (rc >= 0)
2876            {            {
2877            const uschar *end_subpattern = code;            const pcre_uchar *end_subpattern = code;
2878            int charcount = local_offsets[1] - local_offsets[0];            int charcount = local_offsets[1] - local_offsets[0];
2879            int next_state_offset, repeat_state_offset;            int next_state_offset, repeat_state_offset;
2880    
2881            do { end_subpattern += GET(end_subpattern, 1); }            do { end_subpattern += GET(end_subpattern, 1); }
2882              while (*end_subpattern == OP_ALT);              while (*end_subpattern == OP_ALT);
2883            next_state_offset = end_subpattern - start_code + LINK_SIZE + 1;            next_state_offset =
2884                (int)(end_subpattern - start_code + LINK_SIZE + 1);
2885    
2886            /* If the end of this subpattern is KETRMAX or KETRMIN, we must            /* If the end of this subpattern is KETRMAX or KETRMIN, we must
2887            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 2378  for (;;) Line 2889  for (;;)
2889    
2890            repeat_state_offset = (*end_subpattern == OP_KETRMAX ||            repeat_state_offset = (*end_subpattern == OP_KETRMAX ||
2891                                   *end_subpattern == OP_KETRMIN)?                                   *end_subpattern == OP_KETRMIN)?
2892              end_subpattern - start_code - GET(end_subpattern, 1) : -1;              (int)(end_subpattern - start_code - GET(end_subpattern, 1)) : -1;
2893    
2894            /* 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
2895            current character pointer. This is important so that the duplicate            current character pointer. This is important so that the duplicate
# Line 2393  for (;;) Line 2904  for (;;)
2904            /* Optimization: if there are no more active states, and there            /* Optimization: if there are no more active states, and there
2905            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
2906            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
2907            into action when the end of the substring is reached. */            into action when the end of the matched substring is reached. */
2908    
2909            else if (i + 1 >= active_count && new_count == 0)            else if (i + 1 >= active_count && new_count == 0)
2910              {              {
# Line 2416  for (;;) Line 2927  for (;;)
2927              }              }
2928            else            else
2929              {              {
2930              const uschar *p = start_subject + local_offsets[0];  #ifdef SUPPORT_UTF
2931              const uschar *pp = start_subject + local_offsets[1];              if (utf)
2932              while (p < pp) if ((*p++ & 0xc0) == 0x80) charcount--;                {
2933                  const pcre_uchar *p = start_subject + local_offsets[0];
2934                  const pcre_uchar *pp = start_subject + local_offsets[1];
2935                  while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--;
2936                  }
2937    #endif
2938              ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1));              ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1));
2939              if (repeat_state_offset >= 0)              if (repeat_state_offset >= 0)
2940                { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); }                { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); }
2941              }              }
   
2942            }            }
2943          else if (rc != PCRE_ERROR_NOMATCH) return rc;          else if (rc != PCRE_ERROR_NOMATCH) return rc;
2944          }          }
# Line 2435  for (;;) Line 2950  for (;;)
2950    
2951        case OP_CALLOUT:        case OP_CALLOUT:
2952        rrc = 0;        rrc = 0;
2953        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
2954          {          {
2955          pcre_callout_block cb;          PUBL(callout_block) cb;
2956          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 1;   /* Version 1 of the callout block */
2957          cb.callout_number   = code[1];          cb.callout_number   = code[1];
2958          cb.offset_vector    = offsets;          cb.offset_vector    = offsets;
2959    #ifdef COMPILE_PCRE8
2960          cb.subject          = (PCRE_SPTR)start_subject;          cb.subject          = (PCRE_SPTR)start_subject;
2961          cb.subject_length   = end_subject - start_subject;  #else
2962          cb.start_match      = current_subject - start_subject;          cb.subject          = (PCRE_SPTR16)start_subject;
2963          cb.current_position = ptr - start_subject;  #endif
2964            cb.subject_length   = (int)(end_subject - start_subject);
2965            cb.start_match      = (int)(current_subject - start_subject);
2966            cb.current_position = (int)(ptr - start_subject);
2967          cb.pattern_position = GET(code, 2);          cb.pattern_position = GET(code, 2);
2968          cb.next_item_length = GET(code, 2 + LINK_SIZE);          cb.next_item_length = GET(code, 2 + LINK_SIZE);
2969          cb.capture_top      = 1;          cb.capture_top      = 1;
2970          cb.capture_last     = -1;          cb.capture_last     = -1;
2971          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
2972          if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc;   /* Abandon */          cb.mark             = NULL;   /* No (*MARK) support */
2973            if ((rrc = (*PUBL(callout))(&cb)) < 0) return rrc;   /* Abandon */
2974          }          }
2975        if (rrc == 0)        if (rrc == 0)
2976          { ADD_ACTIVE(state_offset + _pcre_OP_lengths[OP_CALLOUT], 0); }          { ADD_ACTIVE(state_offset + PRIV(OP_lengths)[OP_CALLOUT], 0); }
2977        break;        break;
2978    
2979    
# Line 2469  for (;;) Line 2989  for (;;)
2989    /* We have finished the processing at the current subject character. If no    /* We have finished the processing at the current subject character. If no
2990    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
2991    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
2992    matching has been requested, check for appropriate conditions. */    matching has been requested, check for appropriate conditions.
2993    
2994      The "forced_ fail" variable counts the number of (*F) encountered for the
2995      character. If it is equal to the original active_count (saved in
2996      workspace[1]) it means that (*F) was found on every active state. In this
2997      case we don't want to give a partial match.
2998    
2999      The "could_continue" variable is true if a state could have continued but
3000      for the fact that the end of the subject was reached. */
3001    
3002    if (new_count <= 0)    if (new_count <= 0)
3003      {      {
3004      if (match_count < 0 &&                     /* No matches found */      if (rlevel == 1 &&                               /* Top level, and */
3005          rlevel == 1 &&                         /* Top level match function */          could_continue &&                            /* Some could go on, and */
3006          (md->moptions & PCRE_PARTIAL) != 0 &&  /* Want partial matching */          forced_fail != workspace[1] &&               /* Not all forced fail & */
3007          ptr >= end_subject &&                  /* Reached end of subject */          (                                            /* either... */
3008          ptr > current_subject)                 /* Matched non-empty string */          (md->moptions & PCRE_PARTIAL_HARD) != 0      /* Hard partial */
3009            ||                                           /* or... */
3010            ((md->moptions & PCRE_PARTIAL_SOFT) != 0 &&  /* Soft partial and */
3011             match_count < 0)                            /* no matches */
3012            ) &&                                         /* And... */
3013            (
3014            partial_newline ||                           /* Either partial NL */
3015              (                                          /* or ... */
3016              ptr >= end_subject &&                /* End of subject and */
3017              ptr > md->start_used_ptr)            /* Inspected non-empty string */
3018              )
3019            )
3020        {        {
3021        if (offsetcount >= 2)        if (offsetcount >= 2)
3022          {          {
3023          offsets[0] = current_subject - start_subject;          offsets[0] = (int)(md->start_used_ptr - start_subject);
3024          offsets[1] = end_subject - start_subject;          offsets[1] = (int)(end_subject - start_subject);
3025          }          }
3026        match_count = PCRE_ERROR_PARTIAL;        match_count = PCRE_ERROR_PARTIAL;
3027        }        }
# Line 2536  Returns:          > 0 => number of match Line 3075  Returns:          > 0 => number of match
3075                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
3076  */  */
3077    
3078    #ifdef COMPILE_PCRE8
3079  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
3080  pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data,
3081    const char *subject, int length, int start_offset, int options, int *offsets,    const char *subject, int length, int start_offset, int options, int *offsets,
3082    int offsetcount, int *workspace, int wscount)    int offsetcount, int *workspace, int wscount)
3083    #else
3084    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
3085    pcre16_dfa_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
3086      PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
3087      int offsetcount, int *workspace, int wscount)
3088    #endif
3089  {  {
3090  real_pcre *re = (real_pcre *)argument_re;  REAL_PCRE *re = (REAL_PCRE *)argument_re;
3091  dfa_match_data match_block;  dfa_match_data match_block;
3092  dfa_match_data *md = &match_block;  dfa_match_data *md = &match_block;
3093  BOOL utf8, anchored, startline, firstline;  BOOL utf, anchored, startline, firstline;
3094  const uschar *current_subject, *end_subject, *lcc;  const pcre_uchar *current_subject, *end_subject;
   
 pcre_study_data internal_study;  
3095  const pcre_study_data *study = NULL;  const pcre_study_data *study = NULL;
 real_pcre internal_re;  
3096    
3097  const uschar *req_byte_ptr;  const pcre_uchar *req_char_ptr;
3098  const uschar *start_bits = NULL;  const pcre_uint8 *start_bits = NULL;
3099  BOOL first_byte_caseless = FALSE;  BOOL has_first_char = FALSE;
3100  BOOL req_byte_caseless = FALSE;  BOOL has_req_char = FALSE;
3101  int first_byte = -1;  pcre_uchar first_char = 0;
3102  int req_byte = -1;  pcre_uchar first_char2 = 0;
3103  int req_byte2 = -1;  pcre_uchar req_char = 0;
3104    pcre_uchar req_char2 = 0;
3105  int newline;  int newline;
3106    
3107  /* Plausibility checks */  /* Plausibility checks */
# Line 2567  if (re == NULL || subject == NULL || wor Line 3111  if (re == NULL || subject == NULL || wor
3111     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;
3112  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
3113  if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE;  if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE;
3114    if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
3115    
3116    /* Check that the first field in the block is the magic number. If it is not,
3117    return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to
3118    REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which
3119    means that the pattern is likely compiled with different endianness. */
3120    
3121  /* We need to find the pointer to any study data before we test for byte  if (re->magic_number != MAGIC_NUMBER)
3122  flipping, so we scan the extra_data block first. This may set two fields in the    return re->magic_number == REVERSED_MAGIC_NUMBER?
3123  match block, so we must initialize them beforehand. However, the other fields      PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC;
3124  in the match block must not be set until after the byte flipping. */  if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE;
3125    
3126    /* If restarting after a partial match, do some sanity checks on the contents
3127    of the workspace. */
3128    
3129    if ((options & PCRE_DFA_RESTART) != 0)
3130      {
3131      if ((workspace[0] & (-2)) != 0 || workspace[1] < 1 ||
3132        workspace[1] > (wscount - 2)/INTS_PER_STATEBLOCK)
3133          return PCRE_ERROR_DFA_BADRESTART;
3134      }
3135    
3136    /* Set up study, callout, and table data */
3137    
3138  md->tables = re->tables;  md->tables = re->tables;
3139  md->callout_data = NULL;  md->callout_data = NULL;
# Line 2590  if (extra_data != NULL) Line 3152  if (extra_data != NULL)
3152      md->tables = extra_data->tables;      md->tables = extra_data->tables;
3153    }    }
3154    
 /* Check that the first field in the block is the magic number. If it is not,  
 test for a regex that was compiled on a host of opposite endianness. If this is  
 the case, flipped values are put in internal_re and internal_study if there was  
 study data too. */  
   
 if (re->magic_number != MAGIC_NUMBER)  
   {  
   re = _pcre_try_flipped(re, &internal_re, study, &internal_study);  
   if (re == NULL) return PCRE_ERROR_BADMAGIC;  
   if (study != NULL) study = &internal_study;  
   }  
   
3155  /* Set some local values */  /* Set some local values */
3156    
3157  current_subject = (const unsigned char *)subject + start_offset;  current_subject = (const pcre_uchar *)subject + start_offset;
3158  end_subject = (const unsigned char *)subject + length;  end_subject = (const pcre_uchar *)subject + length;
3159  req_byte_ptr = current_subject - 1;  req_char_ptr = current_subject - 1;
3160    
3161  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3162  utf8 = (re->options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
3163    utf = (re->options & PCRE_UTF8) != 0;
3164  #else  #else
3165  utf8 = FALSE;  utf = FALSE;
3166  #endif  #endif
3167    
3168  anchored = (options & (PCRE_ANCHORED|PCRE_DFA_RESTART)) != 0 ||  anchored = (options & (PCRE_ANCHORED|PCRE_DFA_RESTART)) != 0 ||
# Line 2619  anchored = (options & (PCRE_ANCHORED|PCR Line 3170  anchored = (options & (PCRE_ANCHORED|PCR
3170    
3171  /* The remaining fixed data for passing around. */  /* The remaining fixed data for passing around. */
3172    
3173  md->start_code = (const uschar *)argument_re +  md->start_code = (const pcre_uchar *)argument_re +
3174      re->name_table_offset + re->name_count * re->name_entry_size;      re->name_table_offset + re->name_count * re->name_entry_size;
3175  md->start_subject = (const unsigned char *)subject;  md->start_subject = (const pcre_uchar *)subject;
3176  md->end_subject = end_subject;  md->end_subject = end_subject;
3177    md->start_offset = start_offset;
3178  md->moptions = options;  md->moptions = options;
3179  md->poptions = re->options;  md->poptions = re->options;
3180    
# Line 2681  else Line 3233  else
3233  /* Check a UTF-8 string if required. Unfortunately there's no way of passing  /* Check a UTF-8 string if required. Unfortunately there's no way of passing
3234  back the character offset. */  back the character offset. */
3235    
3236  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3237  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0)
3238    {    {
3239    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    int erroroffset;
3240      return PCRE_ERROR_BADUTF8;    int errorcode = PRIV(valid_utf)((pcre_uchar *)subject, length, &erroroffset);
3241    if (start_offset > 0 && start_offset < length)    if (errorcode != 0)
3242      {      {
3243      int tb = ((uschar *)subject)[start_offset];      if (offsetcount >= 2)
     if (tb > 127)  
3244        {        {
3245        tb &= 0xc0;        offsets[0] = erroroffset;
3246        if (tb != 0 && tb != 0xc0) return PCRE_ERROR_BADUTF8_OFFSET;        offsets[1] = errorcode;
3247        }        }
3248        return (errorcode <= PCRE_UTF8_ERR5 && (options & PCRE_PARTIAL_HARD) != 0)?
3249          PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
3250      }      }
3251      if (start_offset > 0 && start_offset < length &&
3252            NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
3253        return PCRE_ERROR_BADUTF8_OFFSET;
3254    }    }
3255  #endif  #endif
3256    
# Line 2702  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 3258  if (utf8 && (options & PCRE_NO_UTF8_CHEC
3258  is a feature that makes it possible to save compiled regex and re-use them  is a feature that makes it possible to save compiled regex and re-use them
3259  in other programs later. */  in other programs later. */
3260    
3261  if (md->tables == NULL) md->tables = _pcre_default_tables;  if (md->tables == NULL) md->tables = PRIV(default_tables);
3262    
3263  /* The lower casing table and the "must be at the start of a line" flag are  /* The "must be at the start of a line" flags are used in a loop when finding
3264  used in a loop when finding where to start. */  where to start. */
3265    
 lcc = md->tables + lcc_offset;  
3266  startline = (re->flags & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
3267  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
3268    
# Line 2721  if (!anchored) Line 3276  if (!anchored)
3276    {    {
3277    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
3278      {      {
3279      first_byte = re->first_byte & 255;      has_first_char = TRUE;
3280      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      first_char = first_char2 = (pcre_uchar)(re->first_char);
3281        first_byte = lcc[first_byte];      if ((re->flags & PCRE_FCH_CASELESS) != 0)
3282          {
3283          first_char2 = TABLE_GET(first_char, md->tables + fcc_offset, first_char);
3284    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
3285          if (utf && first_char > 127)
3286            first_char2 = UCD_OTHERCASE(first_char);
3287    #endif
3288          }
3289      }      }
3290    else    else
3291      {      {
3292      if (startline && study != NULL &&      if (!startline && study != NULL &&
3293           (study->options & PCRE_STUDY_MAPPED) != 0)           (study->flags & PCRE_STUDY_MAPPED) != 0)
3294        start_bits = study->start_bits;        start_bits = study->start_bits;
3295      }      }
3296    }    }
# Line 2738  character" set. */ Line 3300  character" set. */
3300    
3301  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
3302    {    {
3303    req_byte = re->req_byte & 255;    has_req_char = TRUE;
3304    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_char = req_char2 = (pcre_uchar)(re->req_char);
3305    req_byte2 = (md->tables + fcc_offset)[req_byte];  /* case flipped */    if ((re->flags & PCRE_RCH_CASELESS) != 0)
3306        {
3307        req_char2 = TABLE_GET(req_char, md->tables + fcc_offset, req_char);
3308    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
3309        if (utf && req_char > 127)
3310          req_char2 = UCD_OTHERCASE(req_char);
3311    #endif
3312        }
3313    }    }
3314    
3315  /* 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
# Line 2753  for (;;) Line 3322  for (;;)
3322    
3323    if ((options & PCRE_DFA_RESTART) == 0)    if ((options & PCRE_DFA_RESTART) == 0)
3324      {      {
3325      const uschar *save_end_subject = end_subject;      const pcre_uchar *save_end_subject = end_subject;
3326    
3327      /* If firstline is TRUE, the start of the match is constrained to the first      /* If firstline is TRUE, the start of the match is constrained to the first
3328      line of a multiline string. Implement this by temporarily adjusting      line of a multiline string. Implement this by temporarily adjusting
# Line 2762  for (;;) Line 3331  for (;;)
3331    
3332      if (firstline)      if (firstline)
3333        {        {
3334        USPTR t = current_subject;        PCRE_PUCHAR t = current_subject;
3335  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3336        if (utf8)        if (utf)
3337          {          {
3338          while (t < md->end_subject && !IS_NEWLINE(t))          while (t < md->end_subject && !IS_NEWLINE(t))
3339            {            {
3340            t++;            t++;
3341            while (t < end_subject && (*t & 0xc0) == 0x80) t++;            ACROSSCHAR(t < end_subject, *t, t++);
3342            }            }
3343          }          }
3344        else        else
# Line 2779  for (;;) Line 3348  for (;;)
3348        }        }
3349    
3350      /* There are some optimizations that avoid running the match if a known      /* There are some optimizations that avoid running the match if a known
3351      starting point is not found, or if a known later character is not present.      starting point is not found. However, there is an option that disables
3352      However, there is an option that disables these, for testing and for      these, for testing and for ensuring that all callouts do actually occur.
3353      ensuring that all callouts do actually occur. */      The option can be set in the regex by (*NO_START_OPT) or passed in
3354        match-time options. */
3355    
3356      if ((options & PCRE_NO_START_OPTIMIZE) == 0)      if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)
3357        {        {
3358          /* Advance to a known first char. */
3359    
3360        /* Advance to a known first byte. */        if (has_first_char)
   
       if (first_byte >= 0)  
3361          {          {
3362          if (first_byte_caseless)          if (first_char != first_char2)
3363            while (current_subject < end_subject &&            while (current_subject < end_subject &&
3364                   lcc[*current_subject] != first_byte)                *current_subject != first_char && *current_subject != first_char2)
3365              current_subject++;              current_subject++;
3366          else          else
3367            while (current_subject < end_subject &&            while (current_subject < end_subject &&
3368                   *current_subject != first_byte)                   *current_subject != first_char)
3369              current_subject++;              current_subject++;
3370          }          }
3371    
# Line 2806  for (;;) Line 3375  for (;;)
3375          {          {
3376          if (current_subject > md->start_subject + start_offset)          if (current_subject > md->start_subject + start_offset)
3377            {            {
3378  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3379            if (utf8)            if (utf)
3380              {              {
3381              while (current_subject < end_subject &&              while (current_subject < end_subject &&
3382                     !WAS_NEWLINE(current_subject))                     !WAS_NEWLINE(current_subject))
3383                {                {
3384                current_subject++;                current_subject++;
3385                while(current_subject < end_subject &&                ACROSSCHAR(current_subject < end_subject, *current_subject,
3386                      (*current_subject & 0xc0) == 0x80)                  current_subject++);
                 current_subject++;  
3387                }                }
3388              }              }
3389            else            else
# Line 2842  for (;;) Line 3410  for (;;)
3410          while (current_subject < end_subject)          while (current_subject < end_subject)
3411            {            {
3412            register unsigned int c = *current_subject;            register unsigned int c = *current_subject;
3413            if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++;  #ifndef COMPILE_PCRE8
3414              else break;            if (c > 255) c = 255;
3415    #endif
3416              if ((start_bits[c/8] & (1 << (c&7))) == 0)
3417                {
3418                current_subject++;
3419    #if defined SUPPORT_UTF && defined COMPILE_PCRE8
3420                /* In non 8-bit mode, the iteration will stop for
3421                characters > 255 at the beginning or not stop at all. */
3422                if (utf)
3423                  ACROSSCHAR(current_subject < end_subject, *current_subject,
3424                    current_subject++);
3425    #endif
3426                }
3427              else break;
3428            }            }
3429          }          }
3430        }        }
# Line 2851  for (;;) Line 3432  for (;;)
3432      /* Restore fudged end_subject */      /* Restore fudged end_subject */
3433    
3434      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, and can  
   also be explicitly deactivated. */  
   
   if ((options & PCRE_NO_START_OPTIMIZE) == 0 &&  
       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);  
3435    
3436      /* 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
3437      place we found it at last time. */      disabling is explicitly requested (and of course, by the test above, this
3438        code is not obeyed when restarting after a partial match). */
3439    
3440      if (p > req_byte_ptr)      if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0 &&
3441            (options & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) == 0)
3442        {        {
3443        if (req_byte_caseless)        /* If the pattern was studied, a minimum subject length may be set. This
3444          {        is a lower bound; no actual string of that length may actually match the
3445          while (p < end_subject)        pattern. Although the value is, strictly, in characters, we treat it as
3446            {        bytes to avoid spending too much time in this optimization. */
3447            register int pp = *p++;  
3448            if (pp == req_byte || pp == req_byte2) { p--; break; }        if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 &&
3449            }            (pcre_uint32)(end_subject - current_subject) < study->minlength)
3450          }          return PCRE_ERROR_NOMATCH;
3451        else  
3452          /* If req_char is set, we know that that character must appear in the
3453          subject for the match to succeed. If the first character is set, req_char
3454          must be later in the subject; otherwise the test starts at the match
3455          point. This optimization can save a huge amount of work in patterns with
3456          nested unlimited repeats that aren't going to match. Writing separate
3457          code for cased/caseless versions makes it go faster, as does using an
3458          autoincrement and backing off on a match.
3459    
3460          HOWEVER: when the subject string is very, very long, searching to its end
3461          can take a long time, and give bad performance on quite ordinary
3462          patterns. This showed up when somebody was matching /^C/ on a 32-megabyte
3463          string... so we don't do this when the string is sufficiently long. */
3464    
3465          if (has_req_char && end_subject - current_subject < REQ_BYTE_MAX)
3466          {          {
3467          while (p < end_subject)          register PCRE_PUCHAR p = current_subject + (has_first_char? 1:0);
3468    
3469            /* We don't need to repeat the search if we haven't yet reached the
3470            place we found it at last time. */
3471    
3472            if (p > req_char_ptr)
3473            {            {
3474            if (*p++ == req_byte) { p--; break; }            if (req_char != req_char2)
3475            }              {
3476          }              while (p < end_subject)
3477                  {
3478                  register int pp = *p++;
3479                  if (pp == req_char || pp == req_char2) { p--; break; }
3480                  }
3481                }
3482              else
3483                {
3484                while (p < end_subject)
3485                  {
3486                  if (*p++ == req_char) { p--; break; }
3487                  }
3488                }
3489    
3490        /* If we can't find the required character, break the matching loop,            /* If we can't find the required character, break the matching loop,
3491        which will cause a return or PCRE_ERROR_NOMATCH. */            which will cause a return or PCRE_ERROR_NOMATCH. */
3492    
3493        if (p >= end_subject) break;            if (p >= end_subject) break;
3494    
3495        /* If we have found the required character, save the point where we            /* If we have found the required character, save the point where we
3496        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
3497        the start hasn't passed this character yet. */            the start hasn't passed this character yet. */
3498    
3499        req_byte_ptr = p;            req_char_ptr = p;
3500              }
3501            }
3502        }        }
3503      }      }   /* End of optimizations that are done when not restarting */
3504    
3505    /* OK, now we can do the business */    /* OK, now we can do the business */
3506    
3507      md->start_used_ptr = current_subject;
3508      md->recursive = NULL;
3509    
3510    rc = internal_dfa_exec(    rc = internal_dfa_exec(
3511      md,                                /* fixed match data */      md,                                /* fixed match data */
3512      md->start_code,                    /* this subexpression's code */      md->start_code,                    /* this subexpression's code */
# Line 2921  for (;;) Line 3516  for (;;)
3516      offsetcount,                       /* size of same */      offsetcount,                       /* size of same */
3517      workspace,                         /* workspace vector */      workspace,                         /* workspace vector */
3518      wscount,                           /* size of same */      wscount,                           /* size of same */
3519      re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL), /* ims flags */      0);                                /* function recurse level */
     0,                                 /* function recurse level */  
     0);                                /* regex recurse level */  
3520    
3521    /* Anything other than "no match" means we are done, always; otherwise, carry    /* Anything other than "no match" means we are done, always; otherwise, carry
3522    on only if not anchored. */    on only if not anchored. */
# Line 2935  for (;;) Line 3528  for (;;)
3528    
3529    if (firstline && IS_NEWLINE(current_subject)) break;    if (firstline && IS_NEWLINE(current_subject)) break;
3530    current_subject++;    current_subject++;
3531    if (utf8)  #ifdef SUPPORT_UTF
3532      if (utf)
3533      {      {
3534      while (current_subject < end_subject && (*current_subject & 0xc0) == 0x80)      ACROSSCHAR(current_subject < end_subject, *current_subject,
3535        current_subject++;        current_subject++);
3536      }      }
3537    #endif
3538    if (current_subject > end_subject) break;    if (current_subject > end_subject) break;
3539    
3540    /* If we have just passed a CR and we are now at a LF, and the pattern does    /* If we have just passed a CR and we are now at a LF, and the pattern does

Legend:
Removed from v.426  
changed lines
  Added in v.1046

  ViewVC Help
Powered by ViewVC 1.1.5