/[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 150 by ph10, Tue Apr 17 08:22:40 2007 UTC revision 1100 by chpe, Tue Oct 16 15:56:26 2012 UTC
# Line 3  Line 3 
3  *************************************************/  *************************************************/
4    
5  /* PCRE is a library of functions to support regular expressions whose syntax  /* PCRE is a library of functions to support regular expressions whose syntax
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language (but see
7    below for why this module is different).
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2007 University of Cambridge             Copyright (c) 1997-2012 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 37  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
76    #include "config.h"
77    #endif
78    
79  #define NLBLOCK md             /* Block containing newline information */  #define NLBLOCK md             /* Block containing newline information */
80  #define PSSTART start_subject  /* Field containing processed string start */  #define PSSTART start_subject  /* Field containing processed string start */
81  #define PSEND   end_subject    /* Field containing processed string end */  #define PSEND   end_subject    /* Field containing processed string end */
# Line 56  applications. */ Line 88  applications. */
88  #define SP "                   "  #define SP "                   "
89    
90    
   
91  /*************************************************  /*************************************************
92  *      Code parameters and static tables         *  *      Code parameters and static tables         *
93  *************************************************/  *************************************************/
94    
95  /* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes  /* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes
96  into others, under special conditions. A gap of 20 between the blocks should be  into others, under special conditions. A gap of 20 between the blocks should be
97  enough. */  enough. The resulting opcodes don't have to be less than 256 because they are
98    never stored, so we push them well clear of the normal opcodes. */
99    
100  #define OP_PROP_EXTRA 100  #define OP_PROP_EXTRA       300
101  #define OP_EXTUNI_EXTRA 120  #define OP_EXTUNI_EXTRA     320
102  #define OP_ANYNL_EXTRA 140  #define OP_ANYNL_EXTRA      340
103    #define OP_HSPACE_EXTRA     360
104    #define OP_VSPACE_EXTRA     380
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. */  small value. Non-zero values in the table are the offsets from the opcode where
112    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 uschar coptable[] = {  static const pcre_uint8 coptable[] = {
116    0,                             /* End                                    */    0,                             /* End                                    */
117    0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /* \A, \G, \B, \b, \D, \d, \S, \s, \W, \w */    0, 0, 0, 0, 0,                 /* \A, \G, \K, \B, \b                     */
118    0, 0,                          /* Any, Anybyte                           */    0, 0, 0, 0, 0, 0,              /* \D, \d, \S, \s, \W, \w                 */
119    0, 0, 0, 0,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */    0, 0, 0,                       /* Any, AllAny, Anybyte                   */
120    0, 0, 0, 0, 0,                 /* \Z, \z, Opt, ^, $                      */    0, 0,                          /* \P, \p                                 */
121      0, 0, 0, 0, 0,                 /* \R, \H, \h, \V, \v                     */
122      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 104  static uschar coptable[] = { Line 155  static 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,                       /* MARK, PRUNE, PRUNE_ARG                 */
248      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 uschar toptable1[] = {  static const pcre_uint8 toptable1[] = {
257    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,
260    ctype_word,  ctype_word,    ctype_word,  ctype_word,
261    0                               /* OP_ANY */    0, 0                            /* OP_ANY, OP_ALLANY */
262  };  };
263    
264  static uschar toptable2[] = {  static const pcre_uint8 toptable2[] = {
265    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,
268    ctype_word,  0,    ctype_word,  0,
269    1                               /* OP_ANY */    1, 1                            /* OP_ANY, OP_ALLANY */
270  };  };
271    
272    
# Line 151  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 174  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;  pcre_uint32 c;
306  while (length-- > 0)  while (length-- > 0)
307    {    {
308    if (isprint(c = *(p++)))    if (isprint(c = *(p++)))
309      fprintf(f, "%c", c);      fprintf(f, "%c", c);
310    else    else
311      fprintf(f, "\\x%02x", c);      fprintf(f, "\\x{%02x}", c);
312    }    }
313  }  }
314  #endif  #endif
# Line 207  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 =>  Returns:            > 0 => number of match offset pairs placed in offsets
339                      = 0 =>                      = 0 => offsets overflowed; longest matches are present
340                       -1 => failed to match                       -1 => failed to match
341                     < -1 => some kind of unexpected problem                     < -1 => some kind of unexpected problem
342    
# Line 224  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 235  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 247  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 258  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 270  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 311  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 325  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 353  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 374  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 386  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 419  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 432  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 441  for (;;) Line 571  for (;;)
571    {    {
572    int i, j;    int i, j;
573    int clen, dlen;    int clen, dlen;
574    unsigned int c, d;    pcre_uint32 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 455  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 477  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      GETCHARLENTEST(c, ptr, clen);
617  #endif  /* SUPPORT_UTF8 */  #else
618      c = *ptr;      c = *ptr;
619    #endif  /* SUPPORT_UTF */
620      }      }
621    else    else
622      {      {
# Line 497  for (;;) Line 632  for (;;)
632    for (i = 0; i < active_count; i++)    for (i = 0; i < active_count; i++)
633      {      {
634      stateblock *current_state = active_states + i;      stateblock *current_state = active_states + i;
635      const uschar *code;      BOOL caseless = FALSE;
636        const pcre_uchar *code;
637      int state_offset = current_state->offset;      int state_offset = current_state->offset;
638      int count, codevalue;      int count, codevalue, rrc;
     int chartype, script;  
639    
640  #ifdef DEBUG  #ifdef PCRE_DEBUG
641      printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset);      printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset);
642      if (clen == 0) printf("EOL\n");      if (clen == 0) printf("EOL\n");
643        else if (c > 32 && c < 127) printf("'%c'\n", c);        else if (c > 32 && c < 127) printf("'%c'\n", c);
644          else printf("0x%02x\n", c);          else printf("0x%02x\n", c);
645  #endif  #endif
646    
     /* This variable is referred to implicity in the ADD_xxx macros. */  
   
     ims = current_state->ims;  
   
647      /* 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
648      (negated) state until the number of characters in the data field have      (negated) state until the number of characters in the data field have
649      been skipped". */      been skipped". If the could_continue flag was passed over from a previous
650        state, arrange for it to passed on. */
651    
652      if (state_offset < 0)      if (state_offset < 0)
653        {        {
# Line 524  for (;;) Line 656  for (;;)
656          DPRINTF(("%.*sSkipping this character\n", rlevel*2-2, SP));          DPRINTF(("%.*sSkipping this character\n", rlevel*2-2, SP));
657          ADD_NEW_DATA(state_offset, current_state->count,          ADD_NEW_DATA(state_offset, current_state->count,
658            current_state->data - 1);            current_state->data - 1);
659            if (could_continue) reset_could_continue = TRUE;
660          continue;          continue;
661          }          }
662        else        else
# Line 532  for (;;) Line 665  for (;;)
665          }          }
666        }        }
667    
668      /* 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.
669        See the note at the head of this module about the possibility of improving
670        performance here. */
671    
672      for (j = 0; j < i; j++)      for (j = 0; j < i; j++)
673        {        {
# Line 549  for (;;) Line 684  for (;;)
684      code = start_code + state_offset;      code = start_code + state_offset;
685      codevalue = *code;      codevalue = *code;
686    
687        /* If this opcode inspects a character, but we are at the end of the
688        subject, remember the fact for use when testing for a partial match. */
689    
690        if (clen == 0 && poptable[codevalue] != 0)
691          could_continue = TRUE;
692    
693      /* 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
694      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
695      is wrong, because sometimes zero repetitions of the subject are      is wrong, because sometimes zero repetitions of the subject are
696      permitted.      permitted.
697    
698      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
699      argument that is not a data character - but is always one byte long.      argument that is not a data character - but is always one byte long because
700      Unfortunately, we have to take special action to deal with  \P, \p, and      the values are small. We have to take special action to deal with  \P, \p,
701      \X in this case. To keep the other cases fast, convert these ones to new      \H, \h, \V, \v and \X in this case. To keep the other cases fast, convert
702      opcodes. */      these ones to new opcodes. */
703    
704      if (coptable[codevalue] > 0)      if (coptable[codevalue] > 0)
705        {        {
706        dlen = 1;        dlen = 1;
707  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
708        if (utf8) { GETCHARLEN(d, (code + coptable[codevalue]), dlen); } else        if (utf) { GETCHARLEN(d, (code + coptable[codevalue]), dlen); } else
709  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
710        d = code[coptable[codevalue]];        d = code[coptable[codevalue]];
711        if (codevalue >= OP_TYPESTAR)        if (codevalue >= OP_TYPESTAR)
712          {          {
# Line 576  for (;;) Line 717  for (;;)
717            case OP_PROP: codevalue += OP_PROP_EXTRA; break;            case OP_PROP: codevalue += OP_PROP_EXTRA; break;
718            case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break;            case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break;
719            case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break;            case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break;
720              case OP_NOT_HSPACE:
721              case OP_HSPACE: codevalue += OP_HSPACE_EXTRA; break;
722              case OP_NOT_VSPACE:
723              case OP_VSPACE: codevalue += OP_VSPACE_EXTRA; break;
724            default: break;            default: break;
725            }            }
726          }          }
# Line 591  for (;;) Line 736  for (;;)
736    
737      switch (codevalue)      switch (codevalue)
738        {        {
739    /* ========================================================================== */
740          /* These cases are never obeyed. This is a fudge that causes a compile-
741          time error if the vectors coptable or poptable, which are indexed by
742          opcode, are not the correct length. It seems to be the only way to do
743          such a check at compile time, as the sizeof() operator does not work
744          in the C preprocessor. */
745    
746          case OP_TABLE_LENGTH:
747          case OP_TABLE_LENGTH +
748            ((sizeof(coptable) == OP_TABLE_LENGTH) &&
749             (sizeof(poptable) == OP_TABLE_LENGTH)):
750          break;
751    
752  /* ========================================================================== */  /* ========================================================================== */
753        /* 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
754        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
755        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
756          subpattern, because the possessive subpattern repeats are always handled
757          using recursive calls. Thus, it never adds any new states.
758    
759          At the end of the (sub)pattern, unless we have an empty string and
760          PCRE_NOTEMPTY is set, or PCRE_NOTEMPTY_ATSTART is set and we are at the
761          start of the subject, save the match data, shifting up all previous
762        matches so we always have the longest first. */        matches so we always have the longest first. */
763    
764        case OP_KET:        case OP_KET:
765        case OP_KETRMIN:        case OP_KETRMIN:
766        case OP_KETRMAX:        case OP_KETRMAX:
767          case OP_KETRPOS:
768        if (code != end_code)        if (code != end_code)
769          {          {
770          ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0);          ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0);
# Line 609  for (;;) Line 773  for (;;)
773            ADD_ACTIVE(state_offset - GET(code, 1), 0);            ADD_ACTIVE(state_offset - GET(code, 1), 0);
774            }            }
775          }          }
776        else if (ptr > current_subject || (md->moptions & PCRE_NOTEMPTY) == 0)        else
777          {          {
778          if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0;          if (ptr > current_subject ||
779            else if (match_count > 0 && ++match_count * 2 >= offsetcount)              ((md->moptions & PCRE_NOTEMPTY) == 0 &&
780              match_count = 0;                ((md->moptions & PCRE_NOTEMPTY_ATSTART) == 0 ||
781          count = ((match_count == 0)? offsetcount : match_count * 2) - 2;                  current_subject > start_subject + md->start_offset)))
782          if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int));            {
783          if (offsetcount >= 2)            if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0;
784            {              else if (match_count > 0 && ++match_count * 2 > offsetcount)
785            offsets[0] = current_subject - start_subject;                match_count = 0;
786            offsets[1] = ptr - start_subject;            count = ((match_count == 0)? offsetcount : match_count * 2) - 2;
787            DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP,            if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int));
788              offsets[1] - offsets[0], current_subject));            if (offsetcount >= 2)
789            }              {
790          if ((md->moptions & PCRE_DFA_SHORTEST) != 0)              offsets[0] = (int)(current_subject - start_subject);
791            {              offsets[1] = (int)(ptr - start_subject);
792            DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n"              DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP,
793              "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel,                offsets[1] - offsets[0], (char *)current_subject));
794              match_count, rlevel*2-2, SP));              }
795            return match_count;            if ((md->moptions & PCRE_DFA_SHORTEST) != 0)
796                {
797                DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n"
798                  "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel,
799                  match_count, rlevel*2-2, SP));
800                return match_count;
801                }
802            }            }
803          }          }
804        break;        break;
# Line 640  for (;;) Line 810  for (;;)
810        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
811        case OP_ALT:        case OP_ALT:
812        do { code += GET(code, 1); } while (*code == OP_ALT);        do { code += GET(code, 1); } while (*code == OP_ALT);
813        ADD_ACTIVE(code - start_code, 0);        ADD_ACTIVE((int)(code - start_code), 0);
814        break;        break;
815    
816        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
# Line 648  for (;;) Line 818  for (;;)
818        case OP_SBRA:        case OP_SBRA:
819        do        do
820          {          {
821          ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0);          ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
822          code += GET(code, 1);          code += GET(code, 1);
823          }          }
824        while (*code == OP_ALT);        while (*code == OP_ALT);
# Line 657  for (;;) Line 827  for (;;)
827        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
828        case OP_CBRA:        case OP_CBRA:
829        case OP_SCBRA:        case OP_SCBRA:
830        ADD_ACTIVE(code - start_code + 3 + LINK_SIZE,  0);        ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE + IMM2_SIZE),  0);
831        code += GET(code, 1);        code += GET(code, 1);
832        while (*code == OP_ALT)        while (*code == OP_ALT)
833          {          {
834          ADD_ACTIVE(code - start_code + 1 + LINK_SIZE,  0);          ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE),  0);
835          code += GET(code, 1);          code += GET(code, 1);
836          }          }
837        break;        break;
# Line 672  for (;;) Line 842  for (;;)
842        ADD_ACTIVE(state_offset + 1, 0);        ADD_ACTIVE(state_offset + 1, 0);
843        code += 1 + GET(code, 2);        code += 1 + GET(code, 2);
844        while (*code == OP_ALT) code += GET(code, 1);        while (*code == OP_ALT) code += GET(code, 1);
845        ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0);        ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
846          break;
847    
848          /*-----------------------------------------------------------------*/
849          case OP_SKIPZERO:
850          code += 1 + GET(code, 2);
851          while (*code == OP_ALT) code += GET(code, 1);
852          ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0);
853        break;        break;
854    
855        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
856        case OP_CIRC:        case OP_CIRC:
857        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)))  
858          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
859        break;        break;
860    
861        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
862        case OP_EOD:        case OP_CIRCM:
863        if (ptr >= end_subject) { ADD_ACTIVE(state_offset + 1, 0); }        if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) ||
864              (ptr != end_subject && WAS_NEWLINE(ptr)))
865            { ADD_ACTIVE(state_offset + 1, 0); }
866        break;        break;
867    
868        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
869        case OP_OPT:        case OP_EOD:
870        ims = code[1];        if (ptr >= end_subject)
871        ADD_ACTIVE(state_offset + 2, 0);          {
872            if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
873              could_continue = TRUE;
874            else { ADD_ACTIVE(state_offset + 1, 0); }
875            }
876        break;        break;
877    
878        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
# Line 714  for (;;) Line 894  for (;;)
894    
895        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
896        case OP_ANY:        case OP_ANY:
897        if (clen > 0 && ((ims & PCRE_DOTALL) != 0 || !IS_NEWLINE(ptr)))        if (clen > 0 && !IS_NEWLINE(ptr))
898            {
899            if (ptr + 1 >= md->end_subject &&
900                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
901                NLBLOCK->nltype == NLTYPE_FIXED &&
902                NLBLOCK->nllen == 2 &&
903                c == NLBLOCK->nl[0])
904              {
905              could_continue = partial_newline = TRUE;
906              }
907            else
908              {
909              ADD_NEW(state_offset + 1, 0);
910              }
911            }
912          break;
913    
914          /*-----------------------------------------------------------------*/
915          case OP_ALLANY:
916          if (clen > 0)
917          { ADD_NEW(state_offset + 1, 0); }          { ADD_NEW(state_offset + 1, 0); }
918        break;        break;
919    
920        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
921        case OP_EODN:        case OP_EODN:
922        if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen))        if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
923            could_continue = TRUE;
924          else if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen))
925          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
926        break;        break;
927    
# Line 728  for (;;) Line 929  for (;;)
929        case OP_DOLL:        case OP_DOLL:
930        if ((md->moptions & PCRE_NOTEOL) == 0)        if ((md->moptions & PCRE_NOTEOL) == 0)
931          {          {
932          if (clen == 0 ||          if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
933              (IS_NEWLINE(ptr) &&            could_continue = TRUE;
934                 ((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen)          else if (clen == 0 ||
935                ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) &&
936                   (ptr == end_subject - md->nllen)
937              ))              ))
938            { ADD_ACTIVE(state_offset + 1, 0); }            { ADD_ACTIVE(state_offset + 1, 0); }
939            else if (ptr + 1 >= md->end_subject &&
940                     (md->moptions & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) != 0 &&
941                     NLBLOCK->nltype == NLTYPE_FIXED &&
942                     NLBLOCK->nllen == 2 &&
943                     c == NLBLOCK->nl[0])
944              {
945              if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
946                {
947                reset_could_continue = TRUE;
948                ADD_NEW_DATA(-(state_offset + 1), 0, 1);
949                }
950              else could_continue = partial_newline = TRUE;
951              }
952            }
953          break;
954    
955          /*-----------------------------------------------------------------*/
956          case OP_DOLLM:
957          if ((md->moptions & PCRE_NOTEOL) == 0)
958            {
959            if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0)
960              could_continue = TRUE;
961            else if (clen == 0 ||
962                ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr)))
963              { ADD_ACTIVE(state_offset + 1, 0); }
964            else if (ptr + 1 >= md->end_subject &&
965                     (md->moptions & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) != 0 &&
966                     NLBLOCK->nltype == NLTYPE_FIXED &&
967                     NLBLOCK->nllen == 2 &&
968                     c == NLBLOCK->nl[0])
969              {
970              if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
971                {
972                reset_could_continue = TRUE;
973                ADD_NEW_DATA(-(state_offset + 1), 0, 1);
974                }
975              else could_continue = partial_newline = TRUE;
976              }
977          }          }
978        else if ((ims & PCRE_MULTILINE) != 0 && IS_NEWLINE(ptr))        else if (IS_NEWLINE(ptr))
979          { ADD_ACTIVE(state_offset + 1, 0); }          { ADD_ACTIVE(state_offset + 1, 0); }
980        break;        break;
981    
# Line 765  for (;;) Line 1006  for (;;)
1006    
1007          if (ptr > start_subject)          if (ptr > start_subject)
1008            {            {
1009            const uschar *temp = ptr - 1;            const pcre_uchar *temp = ptr - 1;
1010  #ifdef SUPPORT_UTF8            if (temp < md->start_used_ptr) md->start_used_ptr = temp;
1011            if (utf8) BACKCHAR(temp);  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
1012              if (utf) { BACKCHAR(temp); }
1013  #endif  #endif
1014            GETCHARTEST(d, temp);            GETCHARTEST(d, temp);
1015    #ifdef SUPPORT_UCP
1016              if ((md->poptions & PCRE_UCP) != 0)
1017                {
1018                if (d == '_') left_word = TRUE; else
1019                  {
1020                  int cat = UCD_CATEGORY(d);
1021                  left_word = (cat == ucp_L || cat == ucp_N);
1022                  }
1023                }
1024              else
1025    #endif
1026            left_word = d < 256 && (ctypes[d] & ctype_word) != 0;            left_word = d < 256 && (ctypes[d] & ctype_word) != 0;
1027            }            }
1028          else left_word = 0;          else left_word = FALSE;
1029    
1030          if (clen > 0) right_word = c < 256 && (ctypes[c] & ctype_word) != 0;          if (clen > 0)
1031            else right_word = 0;            {
1032    #ifdef SUPPORT_UCP
1033              if ((md->poptions & PCRE_UCP) != 0)
1034                {
1035                if (c == '_') right_word = TRUE; else
1036                  {
1037                  int cat = UCD_CATEGORY(c);
1038                  right_word = (cat == ucp_L || cat == ucp_N);
1039                  }
1040                }
1041              else
1042    #endif
1043              right_word = c < 256 && (ctypes[c] & ctype_word) != 0;
1044              }
1045            else right_word = FALSE;
1046    
1047          if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY))          if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY))
1048            { ADD_ACTIVE(state_offset + 1, 0); }            { ADD_ACTIVE(state_offset + 1, 0); }
# Line 783  for (;;) Line 1050  for (;;)
1050        break;        break;
1051    
1052    
 #ifdef SUPPORT_UCP  
   
1053        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
1054        /* Check the next character by Unicode property. We will get here only        /* Check the next character by Unicode property. We will get here only
1055        if the support is in the binary; otherwise a compile-time error occurs.        if the support is in the binary; otherwise a compile-time error occurs.
1056        */        */
1057    
1058    #ifdef SUPPORT_UCP
1059        case OP_PROP:        case OP_PROP:
1060        case OP_NOTPROP:        case OP_NOTPROP:
1061        if (clen > 0)        if (clen > 0)
1062          {          {
1063          BOOL OK;          BOOL OK;
1064          int category = _pcre_ucp_findprop(c, &chartype, &script);          const pcre_uint32 *cp;
1065            const ucd_record * prop = GET_UCD(c);
1066          switch(code[1])          switch(code[1])
1067            {            {
1068            case PT_ANY:            case PT_ANY:
# Line 803  for (;;) Line 1070  for (;;)
1070            break;            break;
1071    
1072            case PT_LAMP:            case PT_LAMP:
1073            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1074                   prop->chartype == ucp_Lt;
1075            break;            break;
1076    
1077            case PT_GC:            case PT_GC:
1078            OK = category == code[2];            OK = PRIV(ucp_gentype)[prop->chartype] == code[2];
1079            break;            break;
1080    
1081            case PT_PC:            case PT_PC:
1082            OK = chartype == code[2];            OK = prop->chartype == code[2];
1083            break;            break;
1084    
1085            case PT_SC:            case PT_SC:
1086            OK = script == code[2];            OK = prop->script == code[2];
1087              break;
1088    
1089              /* These are specials for combination cases. */
1090    
1091              case PT_ALNUM:
1092              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1093                   PRIV(ucp_gentype)[prop->chartype] == ucp_N;
1094              break;
1095    
1096              case PT_SPACE:    /* Perl space */
1097              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1098                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1099            break;            break;
1100    
1101              case PT_PXSPACE:  /* POSIX space */
1102              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1103                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1104                   c == CHAR_FF || c == CHAR_CR;
1105              break;
1106    
1107              case PT_WORD:
1108              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1109                   PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
1110                   c == CHAR_UNDERSCORE;
1111              break;
1112    
1113              case PT_CLIST:
1114              cp = PRIV(ucd_caseless_sets) + prop->caseset;
1115              for (;;)
1116                {
1117                if (c < *cp) { OK = FALSE; break; }
1118                if (c == *cp++) { OK = TRUE; break; }
1119                }
1120              break;
1121    
1122            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1123    
1124            default:            default:
# Line 835  for (;;) Line 1136  for (;;)
1136  /* ========================================================================== */  /* ========================================================================== */
1137        /* These opcodes likewise inspect the subject character, but have an        /* These opcodes likewise inspect the subject character, but have an
1138        argument that is not a data character. It is one of these opcodes:        argument that is not a data character. It is one of these opcodes:
1139        OP_ANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE, OP_WORDCHAR,        OP_ANY, OP_ALLANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE,
1140        OP_NOT_WORDCHAR. The value is loaded into d. */        OP_WORDCHAR, OP_NOT_WORDCHAR. The value is loaded into d. */
1141    
1142        case OP_TYPEPLUS:        case OP_TYPEPLUS:
1143        case OP_TYPEMINPLUS:        case OP_TYPEMINPLUS:
# Line 845  for (;;) Line 1146  for (;;)
1146        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1147        if (clen > 0)        if (clen > 0)
1148          {          {
1149          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1150                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1151                NLBLOCK->nltype == NLTYPE_FIXED &&
1152                NLBLOCK->nllen == 2 &&
1153                c == NLBLOCK->nl[0])
1154              {
1155              could_continue = partial_newline = TRUE;
1156              }
1157            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1158              (c < 256 &&              (c < 256 &&
1159                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1160                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1161            {            {
1162            if (count > 0 && codevalue == OP_TYPEPOSPLUS)            if (count > 0 && codevalue == OP_TYPEPOSPLUS)
# Line 871  for (;;) Line 1177  for (;;)
1177        ADD_ACTIVE(state_offset + 2, 0);        ADD_ACTIVE(state_offset + 2, 0);
1178        if (clen > 0)        if (clen > 0)
1179          {          {
1180          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1181                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1182                NLBLOCK->nltype == NLTYPE_FIXED &&
1183                NLBLOCK->nllen == 2 &&
1184                c == NLBLOCK->nl[0])
1185              {
1186              could_continue = partial_newline = TRUE;
1187              }
1188            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1189              (c < 256 &&              (c < 256 &&
1190                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1191                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1192            {            {
1193            if (codevalue == OP_TYPEPOSQUERY)            if (codevalue == OP_TYPEPOSQUERY)
# Line 896  for (;;) Line 1207  for (;;)
1207        ADD_ACTIVE(state_offset + 2, 0);        ADD_ACTIVE(state_offset + 2, 0);
1208        if (clen > 0)        if (clen > 0)
1209          {          {
1210          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1211                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1212                NLBLOCK->nltype == NLTYPE_FIXED &&
1213                NLBLOCK->nllen == 2 &&
1214                c == NLBLOCK->nl[0])
1215              {
1216              could_continue = partial_newline = TRUE;
1217              }
1218            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1219              (c < 256 &&              (c < 256 &&
1220                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1221                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1222            {            {
1223            if (codevalue == OP_TYPEPOSSTAR)            if (codevalue == OP_TYPEPOSSTAR)
# Line 919  for (;;) Line 1235  for (;;)
1235        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1236        if (clen > 0)        if (clen > 0)
1237          {          {
1238          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1239                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1240                NLBLOCK->nltype == NLTYPE_FIXED &&
1241                NLBLOCK->nllen == 2 &&
1242                c == NLBLOCK->nl[0])
1243              {
1244              could_continue = partial_newline = TRUE;
1245              }
1246            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1247              (c < 256 &&              (c < 256 &&
1248                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1249                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1250            {            {
1251            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
1252              { ADD_NEW(state_offset + 4, 0); }              { ADD_NEW(state_offset + 1 + IMM2_SIZE + 1, 0); }
1253            else            else
1254              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
1255            }            }
# Line 939  for (;;) Line 1260  for (;;)
1260        case OP_TYPEUPTO:        case OP_TYPEUPTO:
1261        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
1262        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
1263        ADD_ACTIVE(state_offset + 4, 0);        ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0);
1264        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1265        if (clen > 0)        if (clen > 0)
1266          {          {
1267          if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||          if (d == OP_ANY && ptr + 1 >= md->end_subject &&
1268                (md->moptions & (PCRE_PARTIAL_HARD)) != 0 &&
1269                NLBLOCK->nltype == NLTYPE_FIXED &&
1270                NLBLOCK->nllen == 2 &&
1271                c == NLBLOCK->nl[0])
1272              {
1273              could_continue = partial_newline = TRUE;
1274              }
1275            else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) ||
1276              (c < 256 &&              (c < 256 &&
1277                (d != OP_ANY ||                (d != OP_ANY || !IS_NEWLINE(ptr)) &&
                (ims & PCRE_DOTALL) != 0 ||  
                !IS_NEWLINE(ptr)  
               ) &&  
1278                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))                ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0))
1279            {            {
1280            if (codevalue == OP_TYPEPOSUPTO)            if (codevalue == OP_TYPEPOSUPTO)
# Line 957  for (;;) Line 1283  for (;;)
1283              next_active_state--;              next_active_state--;
1284              }              }
1285            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
1286              { ADD_NEW(state_offset + 4, 0); }              { ADD_NEW(state_offset + 2 + IMM2_SIZE, 0); }
1287            else            else
1288              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
1289            }            }
# Line 970  for (;;) Line 1296  for (;;)
1296        argument. It keeps the code above fast for the other cases. The argument        argument. It keeps the code above fast for the other cases. The argument
1297        is in the d variable. */        is in the d variable. */
1298    
1299    #ifdef SUPPORT_UCP
1300        case OP_PROP_EXTRA + OP_TYPEPLUS:        case OP_PROP_EXTRA + OP_TYPEPLUS:
1301        case OP_PROP_EXTRA + OP_TYPEMINPLUS:        case OP_PROP_EXTRA + OP_TYPEMINPLUS:
1302        case OP_PROP_EXTRA + OP_TYPEPOSPLUS:        case OP_PROP_EXTRA + OP_TYPEPOSPLUS:
# Line 978  for (;;) Line 1305  for (;;)
1305        if (clen > 0)        if (clen > 0)
1306          {          {
1307          BOOL OK;          BOOL OK;
1308          int category = _pcre_ucp_findprop(c, &chartype, &script);          const pcre_uint32 *cp;
1309            const ucd_record * prop = GET_UCD(c);
1310          switch(code[2])          switch(code[2])
1311            {            {
1312            case PT_ANY:            case PT_ANY:
# Line 986  for (;;) Line 1314  for (;;)
1314            break;            break;
1315    
1316            case PT_LAMP:            case PT_LAMP:
1317            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1318                prop->chartype == ucp_Lt;
1319            break;            break;
1320    
1321            case PT_GC:            case PT_GC:
1322            OK = category == code[3];            OK = PRIV(ucp_gentype)[prop->chartype] == code[3];
1323            break;            break;
1324    
1325            case PT_PC:            case PT_PC:
1326            OK = chartype == code[3];            OK = prop->chartype == code[3];
1327            break;            break;
1328    
1329            case PT_SC:            case PT_SC:
1330            OK = script == code[3];            OK = prop->script == code[3];
1331              break;
1332    
1333              /* These are specials for combination cases. */
1334    
1335              case PT_ALNUM:
1336              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1337                   PRIV(ucp_gentype)[prop->chartype] == ucp_N;
1338              break;
1339    
1340              case PT_SPACE:    /* Perl space */
1341              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1342                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1343              break;
1344    
1345              case PT_PXSPACE:  /* POSIX space */
1346              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1347                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1348                   c == CHAR_FF || c == CHAR_CR;
1349            break;            break;
1350    
1351              case PT_WORD:
1352              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1353                   PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
1354                   c == CHAR_UNDERSCORE;
1355              break;
1356    
1357              case PT_CLIST:
1358              cp = PRIV(ucd_caseless_sets) + prop->caseset;
1359              for (;;)
1360                {
1361                if (c < *cp) { OK = FALSE; break; }
1362                if (c == *cp++) { OK = TRUE; break; }
1363                }
1364              break;
1365    
1366            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1367    
1368            default:            default:
# Line 1027  for (;;) Line 1389  for (;;)
1389        case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS:        case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS:
1390        count = current_state->count;  /* Already matched */        count = current_state->count;  /* Already matched */
1391        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }        if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1392        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0)
1393          {          {
1394          const uschar *nptr = ptr + clen;          int lgb, rgb;
1395            const pcre_uchar *nptr = ptr + clen;
1396          int ncount = 0;          int ncount = 0;
1397          if (count > 0 && codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS)          if (count > 0 && codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS)
1398            {            {
1399            active_count--;           /* Remove non-match possibility */            active_count--;           /* Remove non-match possibility */
1400            next_active_state--;            next_active_state--;
1401            }            }
1402            lgb = UCD_GRAPHBREAK(c);
1403          while (nptr < end_subject)          while (nptr < end_subject)
1404            {            {
1405            int nd;            dlen = 1;
1406            int ndlen = 1;            if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); }
1407            GETCHARLEN(nd, nptr, ndlen);            rgb = UCD_GRAPHBREAK(d);
1408            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
1409            ncount++;            ncount++;
1410            nptr += ndlen;            lgb = rgb;
1411              nptr += dlen;
1412            }            }
1413          count++;          count++;
1414          ADD_NEW_DATA(-state_offset, count, ncount);          ADD_NEW_DATA(-state_offset, count, ncount);
1415          }          }
1416        break;        break;
1417    #endif
1418    
1419        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
1420        case OP_ANYNL_EXTRA + OP_TYPEPLUS:        case OP_ANYNL_EXTRA + OP_TYPEPLUS:
# Line 1061  for (;;) Line 1427  for (;;)
1427          int ncount = 0;          int ncount = 0;
1428          switch (c)          switch (c)
1429            {            {
1430            case 0x000d:            case CHAR_VT:
1431            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            case CHAR_FF:
1432            /* Fall through */            case CHAR_NEL:
1433            case 0x000a:  #ifndef EBCDIC
           case 0x000b:  
           case 0x000c:  
           case 0x0085:  
1434            case 0x2028:            case 0x2028:
1435            case 0x2029:            case 0x2029:
1436    #endif  /* Not EBCDIC */
1437              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1438              goto ANYNL01;
1439    
1440              case CHAR_CR:
1441              if (ptr + 1 < end_subject && RAWUCHARTEST(ptr + 1) == CHAR_LF) ncount = 1;
1442              /* Fall through */
1443    
1444              ANYNL01:
1445              case CHAR_LF:
1446            if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS)            if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS)
1447              {              {
1448              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
# Line 1078  for (;;) Line 1451  for (;;)
1451            count++;            count++;
1452            ADD_NEW_DATA(-state_offset, count, ncount);            ADD_NEW_DATA(-state_offset, count, ncount);
1453            break;            break;
1454    
1455              default:
1456              break;
1457              }
1458            }
1459          break;
1460    
1461          /*-----------------------------------------------------------------*/
1462          case OP_VSPACE_EXTRA + OP_TYPEPLUS:
1463          case OP_VSPACE_EXTRA + OP_TYPEMINPLUS:
1464          case OP_VSPACE_EXTRA + OP_TYPEPOSPLUS:
1465          count = current_state->count;  /* Already matched */
1466          if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1467          if (clen > 0)
1468            {
1469            BOOL OK;
1470            switch (c)
1471              {
1472              VSPACE_CASES:
1473              OK = TRUE;
1474              break;
1475    
1476              default:
1477              OK = FALSE;
1478              break;
1479              }
1480    
1481            if (OK == (d == OP_VSPACE))
1482              {
1483              if (count > 0 && codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSPLUS)
1484                {
1485                active_count--;           /* Remove non-match possibility */
1486                next_active_state--;
1487                }
1488              count++;
1489              ADD_NEW_DATA(-state_offset, count, 0);
1490              }
1491            }
1492          break;
1493    
1494          /*-----------------------------------------------------------------*/
1495          case OP_HSPACE_EXTRA + OP_TYPEPLUS:
1496          case OP_HSPACE_EXTRA + OP_TYPEMINPLUS:
1497          case OP_HSPACE_EXTRA + OP_TYPEPOSPLUS:
1498          count = current_state->count;  /* Already matched */
1499          if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); }
1500          if (clen > 0)
1501            {
1502            BOOL OK;
1503            switch (c)
1504              {
1505              HSPACE_CASES:
1506              OK = TRUE;
1507              break;
1508    
1509            default:            default:
1510              OK = FALSE;
1511            break;            break;
1512            }            }
1513    
1514            if (OK == (d == OP_HSPACE))
1515              {
1516              if (count > 0 && codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSPLUS)
1517                {
1518                active_count--;           /* Remove non-match possibility */
1519                next_active_state--;
1520                }
1521              count++;
1522              ADD_NEW_DATA(-state_offset, count, 0);
1523              }
1524          }          }
1525        break;        break;
1526    
1527        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
1528    #ifdef SUPPORT_UCP
1529        case OP_PROP_EXTRA + OP_TYPEQUERY:        case OP_PROP_EXTRA + OP_TYPEQUERY:
1530        case OP_PROP_EXTRA + OP_TYPEMINQUERY:        case OP_PROP_EXTRA + OP_TYPEMINQUERY:
1531        case OP_PROP_EXTRA + OP_TYPEPOSQUERY:        case OP_PROP_EXTRA + OP_TYPEPOSQUERY:
# Line 1102  for (;;) Line 1543  for (;;)
1543        if (clen > 0)        if (clen > 0)
1544          {          {
1545          BOOL OK;          BOOL OK;
1546          int category = _pcre_ucp_findprop(c, &chartype, &script);          const pcre_uint32 *cp;
1547            const ucd_record * prop = GET_UCD(c);
1548          switch(code[2])          switch(code[2])
1549            {            {
1550            case PT_ANY:            case PT_ANY:
# Line 1110  for (;;) Line 1552  for (;;)
1552            break;            break;
1553    
1554            case PT_LAMP:            case PT_LAMP:
1555            OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;            OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1556                prop->chartype == ucp_Lt;
1557            break;            break;
1558    
1559            case PT_GC:            case PT_GC:
1560            OK = category == code[3];            OK = PRIV(ucp_gentype)[prop->chartype] == code[3];
1561            break;            break;
1562    
1563            case PT_PC:            case PT_PC:
1564            OK = chartype == code[3];            OK = prop->chartype == code[3];
1565            break;            break;
1566    
1567            case PT_SC:            case PT_SC:
1568            OK = script == code[3];            OK = prop->script == code[3];
1569              break;
1570    
1571              /* These are specials for combination cases. */
1572    
1573              case PT_ALNUM:
1574              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1575                   PRIV(ucp_gentype)[prop->chartype] == ucp_N;
1576              break;
1577    
1578              case PT_SPACE:    /* Perl space */
1579              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1580                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1581              break;
1582    
1583              case PT_PXSPACE:  /* POSIX space */
1584              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1585                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1586                   c == CHAR_FF || c == CHAR_CR;
1587              break;
1588    
1589              case PT_WORD:
1590              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1591                   PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
1592                   c == CHAR_UNDERSCORE;
1593            break;            break;
1594    
1595              case PT_CLIST:
1596              cp = PRIV(ucd_caseless_sets) + prop->caseset;
1597              for (;;)
1598                {
1599                if (c < *cp) { OK = FALSE; break; }
1600                if (c == *cp++) { OK = TRUE; break; }
1601                }
1602              break;
1603    
1604            /* Should never occur, but keep compilers from grumbling. */            /* Should never occur, but keep compilers from grumbling. */
1605    
1606            default:            default:
# Line 1160  for (;;) Line 1636  for (;;)
1636        QS2:        QS2:
1637    
1638        ADD_ACTIVE(state_offset + 2, 0);        ADD_ACTIVE(state_offset + 2, 0);
1639        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0)
1640          {          {
1641          const uschar *nptr = ptr + clen;          int lgb, rgb;
1642            const pcre_uchar *nptr = ptr + clen;
1643          int ncount = 0;          int ncount = 0;
1644          if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR ||          if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR ||
1645              codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY)              codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY)
# Line 1170  for (;;) Line 1647  for (;;)
1647            active_count--;           /* Remove non-match possibility */            active_count--;           /* Remove non-match possibility */
1648            next_active_state--;            next_active_state--;
1649            }            }
1650            lgb = UCD_GRAPHBREAK(c);
1651          while (nptr < end_subject)          while (nptr < end_subject)
1652            {            {
1653            int nd;            dlen = 1;
1654            int ndlen = 1;            if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); }
1655            GETCHARLEN(nd, nptr, ndlen);            rgb = UCD_GRAPHBREAK(d);
1656            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
1657            ncount++;            ncount++;
1658            nptr += ndlen;            lgb = rgb;
1659              nptr += dlen;
1660            }            }
1661          ADD_NEW_DATA(-(state_offset + count), 0, ncount);          ADD_NEW_DATA(-(state_offset + count), 0, ncount);
1662          }          }
1663        break;        break;
1664    #endif
1665    
1666        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
1667        case OP_ANYNL_EXTRA + OP_TYPEQUERY:        case OP_ANYNL_EXTRA + OP_TYPEQUERY:
# Line 1202  for (;;) Line 1682  for (;;)
1682          int ncount = 0;          int ncount = 0;
1683          switch (c)          switch (c)
1684            {            {
1685            case 0x000d:            case CHAR_VT:
1686            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            case CHAR_FF:
1687            /* Fall through */            case CHAR_NEL:
1688            case 0x000a:  #ifndef EBCDIC
           case 0x000b:  
           case 0x000c:  
           case 0x0085:  
1689            case 0x2028:            case 0x2028:
1690            case 0x2029:            case 0x2029:
1691    #endif  /* Not EBCDIC */
1692              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1693              goto ANYNL02;
1694    
1695              case CHAR_CR:
1696              if (ptr + 1 < end_subject && RAWUCHARTEST(ptr + 1) == CHAR_LF) ncount = 1;
1697              /* Fall through */
1698    
1699              ANYNL02:
1700              case CHAR_LF:
1701            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR ||            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR ||
1702                codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY)                codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY)
1703              {              {
# Line 1219  for (;;) Line 1706  for (;;)
1706              }              }
1707            ADD_NEW_DATA(-(state_offset + count), 0, ncount);            ADD_NEW_DATA(-(state_offset + count), 0, ncount);
1708            break;            break;
1709    
1710            default:            default:
1711            break;            break;
1712            }            }
# Line 1226  for (;;) Line 1714  for (;;)
1714        break;        break;
1715    
1716        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
1717        case OP_PROP_EXTRA + OP_TYPEEXACT:        case OP_VSPACE_EXTRA + OP_TYPEQUERY:
1718        case OP_PROP_EXTRA + OP_TYPEUPTO:        case OP_VSPACE_EXTRA + OP_TYPEMINQUERY:
1719        case OP_PROP_EXTRA + OP_TYPEMINUPTO:        case OP_VSPACE_EXTRA + OP_TYPEPOSQUERY:
1720        case OP_PROP_EXTRA + OP_TYPEPOSUPTO:        count = 2;
1721        if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT)        goto QS4;
1722          { ADD_ACTIVE(state_offset + 6, 0); }  
1723        count = current_state->count;  /* Number already matched */        case OP_VSPACE_EXTRA + OP_TYPESTAR:
1724          case OP_VSPACE_EXTRA + OP_TYPEMINSTAR:
1725          case OP_VSPACE_EXTRA + OP_TYPEPOSSTAR:
1726          count = 0;
1727    
1728          QS4:
1729          ADD_ACTIVE(state_offset + 2, 0);
1730        if (clen > 0)        if (clen > 0)
1731          {          {
1732          BOOL OK;          BOOL OK;
1733          int category = _pcre_ucp_findprop(c, &chartype, &script);          switch (c)
         switch(code[4])  
1734            {            {
1735            case PT_ANY:            VSPACE_CASES:
1736            OK = TRUE;            OK = TRUE;
1737            break;            break;
1738    
           case PT_LAMP:  
           OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt;  
           break;  
   
           case PT_GC:  
           OK = category == code[5];  
           break;  
   
           case PT_PC:  
           OK = chartype == code[5];  
           break;  
   
           case PT_SC:  
           OK = script == code[5];  
           break;  
   
           /* Should never occur, but keep compilers from grumbling. */  
   
1739            default:            default:
1740            OK = codevalue != OP_PROP;            OK = FALSE;
1741            break;            break;
1742            }            }
1743            if (OK == (d == OP_VSPACE))
         if (OK == (d == OP_PROP))  
1744            {            {
1745            if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSUPTO)            if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSSTAR ||
1746                  codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSQUERY)
1747              {              {
1748              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
1749              next_active_state--;              next_active_state--;
1750              }              }
1751            if (++count >= GET2(code, 1))            ADD_NEW_DATA(-(state_offset + count), 0, 0);
             { ADD_NEW(state_offset + 6, 0); }  
           else  
             { ADD_NEW(state_offset, count); }  
1752            }            }
1753          }          }
1754        break;        break;
1755    
1756        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
1757        case OP_EXTUNI_EXTRA + OP_TYPEEXACT:        case OP_HSPACE_EXTRA + OP_TYPEQUERY:
1758        case OP_EXTUNI_EXTRA + OP_TYPEUPTO:        case OP_HSPACE_EXTRA + OP_TYPEMINQUERY:
1759        case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO:        case OP_HSPACE_EXTRA + OP_TYPEPOSQUERY:
1760        case OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO:        count = 2;
1761        if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT)        goto QS5;
1762          { ADD_ACTIVE(state_offset + 4, 0); }  
1763        count = current_state->count;  /* Number already matched */        case OP_HSPACE_EXTRA + OP_TYPESTAR:
1764        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        case OP_HSPACE_EXTRA + OP_TYPEMINSTAR:
1765          case OP_HSPACE_EXTRA + OP_TYPEPOSSTAR:
1766          count = 0;
1767    
1768          QS5:
1769          ADD_ACTIVE(state_offset + 2, 0);
1770          if (clen > 0)
1771          {          {
1772          const uschar *nptr = ptr + clen;          BOOL OK;
1773            switch (c)
1774              {
1775              HSPACE_CASES:
1776              OK = TRUE;
1777              break;
1778    
1779              default:
1780              OK = FALSE;
1781              break;
1782              }
1783    
1784            if (OK == (d == OP_HSPACE))
1785              {
1786              if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSSTAR ||
1787                  codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSQUERY)
1788                {
1789                active_count--;           /* Remove non-match possibility */
1790                next_active_state--;
1791                }
1792              ADD_NEW_DATA(-(state_offset + count), 0, 0);
1793              }
1794            }
1795          break;
1796    
1797          /*-----------------------------------------------------------------*/
1798    #ifdef SUPPORT_UCP
1799          case OP_PROP_EXTRA + OP_TYPEEXACT:
1800          case OP_PROP_EXTRA + OP_TYPEUPTO:
1801          case OP_PROP_EXTRA + OP_TYPEMINUPTO:
1802          case OP_PROP_EXTRA + OP_TYPEPOSUPTO:
1803          if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT)
1804            { ADD_ACTIVE(state_offset + 1 + IMM2_SIZE + 3, 0); }
1805          count = current_state->count;  /* Number already matched */
1806          if (clen > 0)
1807            {
1808            BOOL OK;
1809            const pcre_uint32 *cp;
1810            const ucd_record * prop = GET_UCD(c);
1811            switch(code[1 + IMM2_SIZE + 1])
1812              {
1813              case PT_ANY:
1814              OK = TRUE;
1815              break;
1816    
1817              case PT_LAMP:
1818              OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll ||
1819                prop->chartype == ucp_Lt;
1820              break;
1821    
1822              case PT_GC:
1823              OK = PRIV(ucp_gentype)[prop->chartype] == code[1 + IMM2_SIZE + 2];
1824              break;
1825    
1826              case PT_PC:
1827              OK = prop->chartype == code[1 + IMM2_SIZE + 2];
1828              break;
1829    
1830              case PT_SC:
1831              OK = prop->script == code[1 + IMM2_SIZE + 2];
1832              break;
1833    
1834              /* These are specials for combination cases. */
1835    
1836              case PT_ALNUM:
1837              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1838                   PRIV(ucp_gentype)[prop->chartype] == ucp_N;
1839              break;
1840    
1841              case PT_SPACE:    /* Perl space */
1842              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1843                   c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR;
1844              break;
1845    
1846              case PT_PXSPACE:  /* POSIX space */
1847              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
1848                   c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
1849                   c == CHAR_FF || c == CHAR_CR;
1850              break;
1851    
1852              case PT_WORD:
1853              OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
1854                   PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
1855                   c == CHAR_UNDERSCORE;
1856              break;
1857    
1858              case PT_CLIST:
1859              cp = PRIV(ucd_caseless_sets) + prop->caseset;
1860              for (;;)
1861                {
1862                if (c < *cp) { OK = FALSE; break; }
1863                if (c == *cp++) { OK = TRUE; break; }
1864                }
1865              break;
1866    
1867              /* Should never occur, but keep compilers from grumbling. */
1868    
1869              default:
1870              OK = codevalue != OP_PROP;
1871              break;
1872              }
1873    
1874            if (OK == (d == OP_PROP))
1875              {
1876              if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSUPTO)
1877                {
1878                active_count--;           /* Remove non-match possibility */
1879                next_active_state--;
1880                }
1881              if (++count >= GET2(code, 1))
1882                { ADD_NEW(state_offset + 1 + IMM2_SIZE + 3, 0); }
1883              else
1884                { ADD_NEW(state_offset, count); }
1885              }
1886            }
1887          break;
1888    
1889          /*-----------------------------------------------------------------*/
1890          case OP_EXTUNI_EXTRA + OP_TYPEEXACT:
1891          case OP_EXTUNI_EXTRA + OP_TYPEUPTO:
1892          case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO:
1893          case OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO:
1894          if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT)
1895            { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); }
1896          count = current_state->count;  /* Number already matched */
1897          if (clen > 0)
1898            {
1899            int lgb, rgb;
1900            const pcre_uchar *nptr = ptr + clen;
1901          int ncount = 0;          int ncount = 0;
1902          if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO)          if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO)
1903            {            {
1904            active_count--;           /* Remove non-match possibility */            active_count--;           /* Remove non-match possibility */
1905            next_active_state--;            next_active_state--;
1906            }            }
1907            lgb = UCD_GRAPHBREAK(c);
1908          while (nptr < end_subject)          while (nptr < end_subject)
1909            {            {
1910            int nd;            dlen = 1;
1911            int ndlen = 1;            if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); }
1912            GETCHARLEN(nd, nptr, ndlen);            rgb = UCD_GRAPHBREAK(d);
1913            if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break;            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
1914            ncount++;            ncount++;
1915            nptr += ndlen;            lgb = rgb;
1916              nptr += dlen;
1917            }            }
1918            if (nptr >= end_subject && (md->moptions & PCRE_PARTIAL_HARD) != 0)
1919                reset_could_continue = TRUE;
1920          if (++count >= GET2(code, 1))          if (++count >= GET2(code, 1))
1921            { ADD_NEW_DATA(-(state_offset + 4), 0, ncount); }            { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, ncount); }
1922          else          else
1923            { ADD_NEW_DATA(-state_offset, count, ncount); }            { ADD_NEW_DATA(-state_offset, count, ncount); }
1924          }          }
1925        break;        break;
1926    #endif
1927    
1928        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
1929        case OP_ANYNL_EXTRA + OP_TYPEEXACT:        case OP_ANYNL_EXTRA + OP_TYPEEXACT:
# Line 1320  for (;;) Line 1931  for (;;)
1931        case OP_ANYNL_EXTRA + OP_TYPEMINUPTO:        case OP_ANYNL_EXTRA + OP_TYPEMINUPTO:
1932        case OP_ANYNL_EXTRA + OP_TYPEPOSUPTO:        case OP_ANYNL_EXTRA + OP_TYPEPOSUPTO:
1933        if (codevalue != OP_ANYNL_EXTRA + OP_TYPEEXACT)        if (codevalue != OP_ANYNL_EXTRA + OP_TYPEEXACT)
1934          { ADD_ACTIVE(state_offset + 4, 0); }          { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); }
1935        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
1936        if (clen > 0)        if (clen > 0)
1937          {          {
1938          int ncount = 0;          int ncount = 0;
1939          switch (c)          switch (c)
1940            {            {
1941            case 0x000d:            case CHAR_VT:
1942            if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1;            case CHAR_FF:
1943            /* Fall through */            case CHAR_NEL:
1944            case 0x000a:  #ifndef EBCDIC
           case 0x000b:  
           case 0x000c:  
           case 0x0085:  
1945            case 0x2028:            case 0x2028:
1946            case 0x2029:            case 0x2029:
1947    #endif  /* Not EBCDIC */
1948              if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
1949              goto ANYNL03;
1950    
1951              case CHAR_CR:
1952              if (ptr + 1 < end_subject && RAWUCHARTEST(ptr + 1) == CHAR_LF) ncount = 1;
1953              /* Fall through */
1954    
1955              ANYNL03:
1956              case CHAR_LF:
1957            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO)            if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO)
1958              {              {
1959              active_count--;           /* Remove non-match possibility */              active_count--;           /* Remove non-match possibility */
1960              next_active_state--;              next_active_state--;
1961              }              }
1962            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
1963              { ADD_NEW_DATA(-(state_offset + 4), 0, ncount); }              { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, ncount); }
1964            else            else
1965              { ADD_NEW_DATA(-state_offset, count, ncount); }              { ADD_NEW_DATA(-state_offset, count, ncount); }
1966            break;            break;
1967    
1968              default:
1969              break;
1970              }
1971            }
1972          break;
1973    
1974          /*-----------------------------------------------------------------*/
1975          case OP_VSPACE_EXTRA + OP_TYPEEXACT:
1976          case OP_VSPACE_EXTRA + OP_TYPEUPTO:
1977          case OP_VSPACE_EXTRA + OP_TYPEMINUPTO:
1978          case OP_VSPACE_EXTRA + OP_TYPEPOSUPTO:
1979          if (codevalue != OP_VSPACE_EXTRA + OP_TYPEEXACT)
1980            { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); }
1981          count = current_state->count;  /* Number already matched */
1982          if (clen > 0)
1983            {
1984            BOOL OK;
1985            switch (c)
1986              {
1987              VSPACE_CASES:
1988              OK = TRUE;
1989              break;
1990    
1991              default:
1992              OK = FALSE;
1993              }
1994    
1995            if (OK == (d == OP_VSPACE))
1996              {
1997              if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSUPTO)
1998                {
1999                active_count--;           /* Remove non-match possibility */
2000                next_active_state--;
2001                }
2002              if (++count >= GET2(code, 1))
2003                { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, 0); }
2004              else
2005                { ADD_NEW_DATA(-state_offset, count, 0); }
2006              }
2007            }
2008          break;
2009    
2010          /*-----------------------------------------------------------------*/
2011          case OP_HSPACE_EXTRA + OP_TYPEEXACT:
2012          case OP_HSPACE_EXTRA + OP_TYPEUPTO:
2013          case OP_HSPACE_EXTRA + OP_TYPEMINUPTO:
2014          case OP_HSPACE_EXTRA + OP_TYPEPOSUPTO:
2015          if (codevalue != OP_HSPACE_EXTRA + OP_TYPEEXACT)
2016            { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); }
2017          count = current_state->count;  /* Number already matched */
2018          if (clen > 0)
2019            {
2020            BOOL OK;
2021            switch (c)
2022              {
2023              HSPACE_CASES:
2024              OK = TRUE;
2025              break;
2026    
2027            default:            default:
2028              OK = FALSE;
2029            break;            break;
2030            }            }
2031    
2032            if (OK == (d == OP_HSPACE))
2033              {
2034              if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSUPTO)
2035                {
2036                active_count--;           /* Remove non-match possibility */
2037                next_active_state--;
2038                }
2039              if (++count >= GET2(code, 1))
2040                { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, 0); }
2041              else
2042                { ADD_NEW_DATA(-state_offset, count, 0); }
2043              }
2044          }          }
2045        break;        break;
2046    
# Line 1364  for (;;) Line 2056  for (;;)
2056        break;        break;
2057    
2058        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2059        case OP_CHARNC:        case OP_CHARI:
2060        if (clen == 0) break;        if (clen == 0) break;
2061    
2062  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2063        if (utf8)        if (utf)
2064          {          {
2065          if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else          if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else
2066            {            {
2067            unsigned int othercase;            unsigned int othercase;
2068            if (c < 128) othercase = fcc[c]; else            if (c < 128)
2069                othercase = fcc[c];
2070            /* If we have Unicode property support, we can use it to test the            else
2071            other case of the character. */              /* If we have Unicode property support, we can use it to test the
2072                other case of the character. */
2073  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2074            othercase = _pcre_ucp_othercase(c);              othercase = UCD_OTHERCASE(c);
2075  #else  #else
2076            othercase = NOTACHAR;              othercase = NOTACHAR;
2077  #endif  #endif
2078    
2079            if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); }            if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); }
2080            }            }
2081          }          }
2082        else        else
2083  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2084          /* Not UTF mode */
       /* Non-UTF-8 mode */  
2085          {          {
2086          if (lcc[c] == lcc[d]) { ADD_NEW(state_offset + 2, 0); }          if (TABLE_GET(c, lcc, c) == TABLE_GET(d, lcc, d))
2087              { ADD_NEW(state_offset + 2, 0); }
2088          }          }
2089        break;        break;
2090    
# Line 1404  for (;;) Line 2096  for (;;)
2096        to wait for them to pass before continuing. */        to wait for them to pass before continuing. */
2097    
2098        case OP_EXTUNI:        case OP_EXTUNI:
2099        if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M)        if (clen > 0)
2100          {          {
2101          const uschar *nptr = ptr + clen;          int lgb, rgb;
2102            const pcre_uchar *nptr = ptr + clen;
2103          int ncount = 0;          int ncount = 0;
2104            lgb = UCD_GRAPHBREAK(c);
2105          while (nptr < end_subject)          while (nptr < end_subject)
2106            {            {
2107            int nclen = 1;            dlen = 1;
2108            GETCHARLEN(c, nptr, nclen);            if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); }
2109            if (_pcre_ucp_findprop(c, &chartype, &script) != ucp_M) break;            rgb = UCD_GRAPHBREAK(d);
2110              if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2111            ncount++;            ncount++;
2112            nptr += nclen;            lgb = rgb;
2113              nptr += dlen;
2114            }            }
2115            if (nptr >= end_subject && (md->moptions & PCRE_PARTIAL_HARD) != 0)
2116                reset_could_continue = TRUE;
2117          ADD_NEW_DATA(-(state_offset + 1), 0, ncount);          ADD_NEW_DATA(-(state_offset + 1), 0, ncount);
2118          }          }
2119        break;        break;
# Line 1429  for (;;) Line 2127  for (;;)
2127        case OP_ANYNL:        case OP_ANYNL:
2128        if (clen > 0) switch(c)        if (clen > 0) switch(c)
2129          {          {
2130          case 0x000a:          case CHAR_VT:
2131          case 0x000b:          case CHAR_FF:
2132          case 0x000c:          case CHAR_NEL:
2133          case 0x0085:  #ifndef EBCDIC
2134          case 0x2028:          case 0x2028:
2135          case 0x2029:          case 0x2029:
2136    #endif  /* Not EBCDIC */
2137            if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break;
2138    
2139            case CHAR_LF:
2140          ADD_NEW(state_offset + 1, 0);          ADD_NEW(state_offset + 1, 0);
2141          break;          break;
2142          case 0x000d:  
2143          if (ptr + 1 < end_subject && ptr[1] == 0x0a)          case CHAR_CR:
2144            if (ptr + 1 >= end_subject)
2145              {
2146              ADD_NEW(state_offset + 1, 0);
2147              if ((md->moptions & PCRE_PARTIAL_HARD) != 0)
2148                reset_could_continue = TRUE;
2149              }
2150            else if (RAWUCHARTEST(ptr + 1) == CHAR_LF)
2151            {            {
2152            ADD_NEW_DATA(-(state_offset + 1), 0, 1);            ADD_NEW_DATA(-(state_offset + 1), 0, 1);
2153            }            }
# Line 1451  for (;;) Line 2160  for (;;)
2160        break;        break;
2161    
2162        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2163        /* Match a negated single character. This is only used for one-byte        case OP_NOT_VSPACE:
2164        characters, that is, we know that d < 256. The character we are        if (clen > 0) switch(c)
2165        checking (c) can be multibyte. */          {
2166            VSPACE_CASES:
2167            break;
2168    
2169            default:
2170            ADD_NEW(state_offset + 1, 0);
2171            break;
2172            }
2173          break;
2174    
2175          /*-----------------------------------------------------------------*/
2176          case OP_VSPACE:
2177          if (clen > 0) switch(c)
2178            {
2179            VSPACE_CASES:
2180            ADD_NEW(state_offset + 1, 0);
2181            break;
2182    
2183            default:
2184            break;
2185            }
2186          break;
2187    
2188          /*-----------------------------------------------------------------*/
2189          case OP_NOT_HSPACE:
2190          if (clen > 0) switch(c)
2191            {
2192            HSPACE_CASES:
2193            break;
2194    
2195            default:
2196            ADD_NEW(state_offset + 1, 0);
2197            break;
2198            }
2199          break;
2200    
2201          /*-----------------------------------------------------------------*/
2202          case OP_HSPACE:
2203          if (clen > 0) switch(c)
2204            {
2205            HSPACE_CASES:
2206            ADD_NEW(state_offset + 1, 0);
2207            break;
2208    
2209            default:
2210            break;
2211            }
2212          break;
2213    
2214          /*-----------------------------------------------------------------*/
2215          /* Match a negated single character casefully. */
2216    
2217        case OP_NOT:        case OP_NOT:
2218          if (clen > 0 && c != d) { ADD_NEW(state_offset + dlen + 1, 0); }
2219          break;
2220    
2221          /*-----------------------------------------------------------------*/
2222          /* Match a negated single character caselessly. */
2223    
2224          case OP_NOTI:
2225        if (clen > 0)        if (clen > 0)
2226          {          {
2227          unsigned int otherd = ((ims & PCRE_CASELESS) != 0)? fcc[d] : d;          unsigned int otherd;
2228          if (c != d && c != otherd) { ADD_NEW(state_offset + dlen + 1, 0); }  #ifdef SUPPORT_UTF
2229            if (utf && d >= 128)
2230              {
2231    #ifdef SUPPORT_UCP
2232              otherd = UCD_OTHERCASE(d);
2233    #endif  /* SUPPORT_UCP */
2234              }
2235            else
2236    #endif  /* SUPPORT_UTF */
2237            otherd = TABLE_GET(d, fcc, d);
2238            if (c != d && c != otherd)
2239              { ADD_NEW(state_offset + dlen + 1, 0); }
2240          }          }
2241        break;        break;
2242    
2243        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2244          case OP_PLUSI:
2245          case OP_MINPLUSI:
2246          case OP_POSPLUSI:
2247          case OP_NOTPLUSI:
2248          case OP_NOTMINPLUSI:
2249          case OP_NOTPOSPLUSI:
2250          caseless = TRUE;
2251          codevalue -= OP_STARI - OP_STAR;
2252    
2253          /* Fall through */
2254        case OP_PLUS:        case OP_PLUS:
2255        case OP_MINPLUS:        case OP_MINPLUS:
2256        case OP_POSPLUS:        case OP_POSPLUS:
# Line 1474  for (;;) Line 2261  for (;;)
2261        if (count > 0) { ADD_ACTIVE(state_offset + dlen + 1, 0); }        if (count > 0) { ADD_ACTIVE(state_offset + dlen + 1, 0); }
2262        if (clen > 0)        if (clen > 0)
2263          {          {
2264          unsigned int otherd = NOTACHAR;          pcre_uint32 otherd = NOTACHAR;
2265          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2266            {            {
2267  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2268            if (utf8 && d >= 128)            if (utf && d >= 128)
2269              {              {
2270  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2271              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2272  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2273              }              }
2274            else            else
2275  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2276            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2277            }            }
2278          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2279            {            {
# Line 1503  for (;;) Line 2290  for (;;)
2290        break;        break;
2291    
2292        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2293          case OP_QUERYI:
2294          case OP_MINQUERYI:
2295          case OP_POSQUERYI:
2296          case OP_NOTQUERYI:
2297          case OP_NOTMINQUERYI:
2298          case OP_NOTPOSQUERYI:
2299          caseless = TRUE;
2300          codevalue -= OP_STARI - OP_STAR;
2301          /* Fall through */
2302        case OP_QUERY:        case OP_QUERY:
2303        case OP_MINQUERY:        case OP_MINQUERY:
2304        case OP_POSQUERY:        case OP_POSQUERY:
# Line 1512  for (;;) Line 2308  for (;;)
2308        ADD_ACTIVE(state_offset + dlen + 1, 0);        ADD_ACTIVE(state_offset + dlen + 1, 0);
2309        if (clen > 0)        if (clen > 0)
2310          {          {
2311          unsigned int otherd = NOTACHAR;          pcre_uint32 otherd = NOTACHAR;
2312          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2313            {            {
2314  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2315            if (utf8 && d >= 128)            if (utf && d >= 128)
2316              {              {
2317  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2318              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2319  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2320              }              }
2321            else            else
2322  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2323            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2324            }            }
2325          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2326            {            {
# Line 1539  for (;;) Line 2335  for (;;)
2335        break;        break;
2336    
2337        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2338          case OP_STARI:
2339          case OP_MINSTARI:
2340          case OP_POSSTARI:
2341          case OP_NOTSTARI:
2342          case OP_NOTMINSTARI:
2343          case OP_NOTPOSSTARI:
2344          caseless = TRUE;
2345          codevalue -= OP_STARI - OP_STAR;
2346          /* Fall through */
2347        case OP_STAR:        case OP_STAR:
2348        case OP_MINSTAR:        case OP_MINSTAR:
2349        case OP_POSSTAR:        case OP_POSSTAR:
# Line 1548  for (;;) Line 2353  for (;;)
2353        ADD_ACTIVE(state_offset + dlen + 1, 0);        ADD_ACTIVE(state_offset + dlen + 1, 0);
2354        if (clen > 0)        if (clen > 0)
2355          {          {
2356          unsigned int otherd = NOTACHAR;          pcre_uint32 otherd = NOTACHAR;
2357          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2358            {            {
2359  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2360            if (utf8 && d >= 128)            if (utf && d >= 128)
2361              {              {
2362  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2363              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2364  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2365              }              }
2366            else            else
2367  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2368            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2369            }            }
2370          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2371            {            {
# Line 1575  for (;;) Line 2380  for (;;)
2380        break;        break;
2381    
2382        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2383          case OP_EXACTI:
2384          case OP_NOTEXACTI:
2385          caseless = TRUE;
2386          codevalue -= OP_STARI - OP_STAR;
2387          /* Fall through */
2388        case OP_EXACT:        case OP_EXACT:
2389        case OP_NOTEXACT:        case OP_NOTEXACT:
2390        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
2391        if (clen > 0)        if (clen > 0)
2392          {          {
2393          unsigned int otherd = NOTACHAR;          pcre_uint32 otherd = NOTACHAR;
2394          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2395            {            {
2396  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2397            if (utf8 && d >= 128)            if (utf && d >= 128)
2398              {              {
2399  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2400              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2401  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2402              }              }
2403            else            else
2404  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2405            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2406            }            }
2407          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2408            {            {
2409            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
2410              { ADD_NEW(state_offset + dlen + 3, 0); }              { ADD_NEW(state_offset + dlen + 1 + IMM2_SIZE, 0); }
2411            else            else
2412              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
2413            }            }
# Line 1605  for (;;) Line 2415  for (;;)
2415        break;        break;
2416    
2417        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2418          case OP_UPTOI:
2419          case OP_MINUPTOI:
2420          case OP_POSUPTOI:
2421          case OP_NOTUPTOI:
2422          case OP_NOTMINUPTOI:
2423          case OP_NOTPOSUPTOI:
2424          caseless = TRUE;
2425          codevalue -= OP_STARI - OP_STAR;
2426          /* Fall through */
2427        case OP_UPTO:        case OP_UPTO:
2428        case OP_MINUPTO:        case OP_MINUPTO:
2429        case OP_POSUPTO:        case OP_POSUPTO:
2430        case OP_NOTUPTO:        case OP_NOTUPTO:
2431        case OP_NOTMINUPTO:        case OP_NOTMINUPTO:
2432        case OP_NOTPOSUPTO:        case OP_NOTPOSUPTO:
2433        ADD_ACTIVE(state_offset + dlen + 3, 0);        ADD_ACTIVE(state_offset + dlen + 1 + IMM2_SIZE, 0);
2434        count = current_state->count;  /* Number already matched */        count = current_state->count;  /* Number already matched */
2435        if (clen > 0)        if (clen > 0)
2436          {          {
2437          unsigned int otherd = NOTACHAR;          pcre_uint32 otherd = NOTACHAR;
2438          if ((ims & PCRE_CASELESS) != 0)          if (caseless)
2439            {            {
2440  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2441            if (utf8 && d >= 128)            if (utf && d >= 128)
2442              {              {
2443  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2444              otherd = _pcre_ucp_othercase(d);              otherd = UCD_OTHERCASE(d);
2445  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2446              }              }
2447            else            else
2448  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
2449            otherd = fcc[d];            otherd = TABLE_GET(d, fcc, d);
2450            }            }
2451          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))          if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR))
2452            {            {
# Line 1637  for (;;) Line 2456  for (;;)
2456              next_active_state--;              next_active_state--;
2457              }              }
2458            if (++count >= GET2(code, 1))            if (++count >= GET2(code, 1))
2459              { ADD_NEW(state_offset + dlen + 3, 0); }              { ADD_NEW(state_offset + dlen + 1 + IMM2_SIZE, 0); }
2460            else            else
2461              { ADD_NEW(state_offset, count); }              { ADD_NEW(state_offset, count); }
2462            }            }
# Line 1654  for (;;) Line 2473  for (;;)
2473          {          {
2474          BOOL isinclass = FALSE;          BOOL isinclass = FALSE;
2475          int next_state_offset;          int next_state_offset;
2476          const uschar *ecode;          const pcre_uchar *ecode;
2477    
2478          /* 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
2479          can set isinclass from it. */          can set isinclass from it. */
2480    
2481          if (codevalue != OP_XCLASS)          if (codevalue != OP_XCLASS)
2482            {            {
2483            ecode = code + 33;            ecode = code + 1 + (32 / sizeof(pcre_uchar));
2484            if (clen > 0)            if (clen > 0)
2485              {              {
2486              isinclass = (c > 255)? (codevalue == OP_NCLASS) :              isinclass = (c > 255)? (codevalue == OP_NCLASS) :
2487                ((code[1 + c/8] & (1 << (c&7))) != 0);                ((((pcre_uint8 *)(code + 1))[c/8] & (1 << (c&7))) != 0);
2488              }              }
2489            }            }
2490    
# Line 1676  for (;;) Line 2495  for (;;)
2495          else          else
2496           {           {
2497           ecode = code + GET(code, 1);           ecode = code + GET(code, 1);
2498           if (clen > 0) isinclass = _pcre_xclass(c, code + 1 + LINK_SIZE);           if (clen > 0) isinclass = PRIV(xclass)(c, code + 1 + LINK_SIZE, utf);
2499           }           }
2500    
2501          /* 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
2502          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
2503          quantifier, this is where it will be. */          quantifier, this is where it will be. */
2504    
2505          next_state_offset = ecode - start_code;          next_state_offset = (int)(ecode - start_code);
2506    
2507          switch (*ecode)          switch (*ecode)
2508            {            {
# Line 1710  for (;;) Line 2529  for (;;)
2529            case OP_CRMINRANGE:            case OP_CRMINRANGE:
2530            count = current_state->count;  /* Already matched */            count = current_state->count;  /* Already matched */
2531            if (count >= GET2(ecode, 1))            if (count >= GET2(ecode, 1))
2532              { ADD_ACTIVE(next_state_offset + 5, 0); }              { ADD_ACTIVE(next_state_offset + 1 + 2 * IMM2_SIZE, 0); }
2533            if (isinclass)            if (isinclass)
2534              {              {
2535              int max = GET2(ecode, 3);              int max = GET2(ecode, 1 + IMM2_SIZE);
2536              if (++count >= max && max != 0)   /* Max 0 => no limit */              if (++count >= max && max != 0)   /* Max 0 => no limit */
2537                { ADD_NEW(next_state_offset + 5, 0); }                { ADD_NEW(next_state_offset + 1 + 2 * IMM2_SIZE, 0); }
2538              else              else
2539                { ADD_NEW(state_offset, count); }                { ADD_NEW(state_offset, count); }
2540              }              }
# Line 1730  for (;;) Line 2549  for (;;)
2549    
2550  /* ========================================================================== */  /* ========================================================================== */
2551        /* These are the opcodes for fancy brackets of various kinds. We have        /* These are the opcodes for fancy brackets of various kinds. We have
2552        to use recursion in order to handle them. */        to use recursion in order to handle them. The "always failing" assertion
2553          (?!) is optimised to OP_FAIL when compiling, so we have to support that,
2554          though the other "backtracking verbs" are not supported. */
2555    
2556          case OP_FAIL:
2557          forced_fail++;    /* Count FAILs for multiple states */
2558          break;
2559    
2560        case OP_ASSERT:        case OP_ASSERT:
2561        case OP_ASSERT_NOT:        case OP_ASSERT_NOT:
# Line 1740  for (;;) Line 2565  for (;;)
2565          int rc;          int rc;
2566          int local_offsets[2];          int local_offsets[2];
2567          int local_workspace[1000];          int local_workspace[1000];
2568          const uschar *endasscode = code + GET(code, 1);          const pcre_uchar *endasscode = code + GET(code, 1);
2569    
2570          while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1);          while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1);
2571    
# Line 1748  for (;;) Line 2573  for (;;)
2573            md,                                   /* static match data */            md,                                   /* static match data */
2574            code,                                 /* this subexpression's code */            code,                                 /* this subexpression's code */
2575            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2576            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2577            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2578            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2579            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2580            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
2581            ims,                                  /* the current ims flags */            rlevel);                              /* function recursion level */
           rlevel,                               /* function recursion level */  
           recursing);                           /* pass on regex recursion */  
2582    
2583            if (rc == PCRE_ERROR_DFA_UITEM) return rc;
2584          if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK))          if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK))
2585              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }              { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); }
2586          }          }
2587        break;        break;
2588    
# Line 1768  for (;;) Line 2592  for (;;)
2592          {          {
2593          int local_offsets[1000];          int local_offsets[1000];
2594          int local_workspace[1000];          int local_workspace[1000];
2595          int condcode = code[LINK_SIZE+1];          int codelink = GET(code, 1);
2596            int condcode;
2597    
2598            /* Because of the way auto-callout works during compile, a callout item
2599            is inserted between OP_COND and an assertion condition. This does not
2600            happen for the other conditions. */
2601    
2602            if (code[LINK_SIZE+1] == OP_CALLOUT)
2603              {
2604              rrc = 0;
2605              if (PUBL(callout) != NULL)
2606                {
2607                PUBL(callout_block) cb;
2608                cb.version          = 1;   /* Version 1 of the callout block */
2609                cb.callout_number   = code[LINK_SIZE+2];
2610                cb.offset_vector    = offsets;
2611    #if defined COMPILE_PCRE8
2612                cb.subject          = (PCRE_SPTR)start_subject;
2613    #elif defined COMPILE_PCRE16
2614                cb.subject          = (PCRE_SPTR16)start_subject;
2615    #elif defined COMPILE_PCRE32
2616                cb.subject          = (PCRE_SPTR32)start_subject;
2617    #endif
2618                cb.subject_length   = (int)(end_subject - start_subject);
2619                cb.start_match      = (int)(current_subject - start_subject);
2620                cb.current_position = (int)(ptr - start_subject);
2621                cb.pattern_position = GET(code, LINK_SIZE + 3);
2622                cb.next_item_length = GET(code, 3 + 2*LINK_SIZE);
2623                cb.capture_top      = 1;
2624                cb.capture_last     = -1;
2625                cb.callout_data     = md->callout_data;
2626                cb.mark             = NULL;   /* No (*MARK) support */
2627                if ((rrc = (*PUBL(callout))(&cb)) < 0) return rrc;   /* Abandon */
2628                }
2629              if (rrc > 0) break;                      /* Fail this thread */
2630              code += PRIV(OP_lengths)[OP_CALLOUT];    /* Skip callout data */
2631              }
2632    
2633            condcode = code[LINK_SIZE+1];
2634    
2635          /* Back reference conditions are not supported */          /* Back reference conditions are not supported */
2636    
2637          if (condcode == OP_CREF) return PCRE_ERROR_DFA_UCOND;          if (condcode == OP_CREF || condcode == OP_NCREF)
2638              return PCRE_ERROR_DFA_UCOND;
2639    
2640          /* The DEFINE condition is always false */          /* The DEFINE condition is always false */
2641    
2642          if (condcode == OP_DEF)          if (condcode == OP_DEF)
2643            {            { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
           ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0);  
           }  
2644    
2645          /* The only supported version of OP_RREF is for the value RREF_ANY,          /* The only supported version of OP_RREF is for the value RREF_ANY,
2646          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
2647          recursed groups. */          recursed groups. */
2648    
2649          else if (condcode == OP_RREF)          else if (condcode == OP_RREF || condcode == OP_NRREF)
2650            {            {
2651            int value = GET2(code, LINK_SIZE+2);            int value = GET2(code, LINK_SIZE + 2);
2652            if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND;            if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND;
2653            if (recursing > 0) { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); }            if (md->recursive != NULL)
2654              else { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); }              { ADD_ACTIVE(state_offset + LINK_SIZE + 2 + IMM2_SIZE, 0); }
2655              else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
2656            }            }
2657    
2658          /* Otherwise, the condition is an assertion */          /* Otherwise, the condition is an assertion */
# Line 1798  for (;;) Line 2660  for (;;)
2660          else          else
2661            {            {
2662            int rc;            int rc;
2663            const uschar *asscode = code + LINK_SIZE + 1;            const pcre_uchar *asscode = code + LINK_SIZE + 1;
2664            const uschar *endasscode = asscode + GET(asscode, 1);            const pcre_uchar *endasscode = asscode + GET(asscode, 1);
2665    
2666            while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1);            while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1);
2667    
# Line 1807  for (;;) Line 2669  for (;;)
2669              md,                                   /* fixed match data */              md,                                   /* fixed match data */
2670              asscode,                              /* this subexpression's code */              asscode,                              /* this subexpression's code */
2671              ptr,                                  /* where we currently are */              ptr,                                  /* where we currently are */
2672              ptr - start_subject,                  /* start offset */              (int)(ptr - start_subject),           /* start offset */
2673              local_offsets,                        /* offset vector */              local_offsets,                        /* offset vector */
2674              sizeof(local_offsets)/sizeof(int),    /* size of same */              sizeof(local_offsets)/sizeof(int),    /* size of same */
2675              local_workspace,                      /* workspace vector */              local_workspace,                      /* workspace vector */
2676              sizeof(local_workspace)/sizeof(int),  /* size of same */              sizeof(local_workspace)/sizeof(int),  /* size of same */
2677              ims,                                  /* the current ims flags */              rlevel);                              /* function recursion level */
             rlevel,                               /* function recursion level */  
             recursing);                           /* pass on regex recursion */  
2678    
2679              if (rc == PCRE_ERROR_DFA_UITEM) return rc;
2680            if ((rc >= 0) ==            if ((rc >= 0) ==
2681                  (condcode == OP_ASSERT || condcode == OP_ASSERTBACK))                  (condcode == OP_ASSERT || condcode == OP_ASSERTBACK))
2682              { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); }              { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); }
2683            else            else
2684              { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); }              { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); }
2685            }            }
2686          }          }
2687        break;        break;
# Line 1828  for (;;) Line 2689  for (;;)
2689        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2690        case OP_RECURSE:        case OP_RECURSE:
2691          {          {
2692            dfa_recursion_info *ri;
2693          int local_offsets[1000];          int local_offsets[1000];
2694          int local_workspace[1000];          int local_workspace[1000];
2695            const pcre_uchar *callpat = start_code + GET(code, 1);
2696            int recno = (callpat == md->start_code)? 0 :
2697              GET2(callpat, 1 + LINK_SIZE);
2698          int rc;          int rc;
2699    
2700          DPRINTF(("%.*sStarting regex recursion %d\n", rlevel*2-2, SP,          DPRINTF(("%.*sStarting regex recursion\n", rlevel*2-2, SP));
2701            recursing + 1));  
2702            /* Check for repeating a recursion without advancing the subject
2703            pointer. This should catch convoluted mutual recursions. (Some simple
2704            cases are caught at compile time.) */
2705    
2706            for (ri = md->recursive; ri != NULL; ri = ri->prevrec)
2707              if (recno == ri->group_num && ptr == ri->subject_position)
2708                return PCRE_ERROR_RECURSELOOP;
2709    
2710            /* Remember this recursion and where we started it so as to
2711            catch infinite loops. */
2712    
2713            new_recursive.group_num = recno;
2714            new_recursive.subject_position = ptr;
2715            new_recursive.prevrec = md->recursive;
2716            md->recursive = &new_recursive;
2717    
2718          rc = internal_dfa_exec(          rc = internal_dfa_exec(
2719            md,                                   /* fixed match data */            md,                                   /* fixed match data */
2720            start_code + GET(code, 1),            /* this subexpression's code */            callpat,                              /* this subexpression's code */
2721            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2722            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2723            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2724            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2725            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2726            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
2727            ims,                                  /* the current ims flags */            rlevel);                              /* function recursion level */
           rlevel,                               /* function recursion level */  
           recursing + 1);                       /* regex recurse level */  
2728    
2729          DPRINTF(("%.*sReturn from regex recursion %d: rc=%d\n", rlevel*2-2, SP,          md->recursive = new_recursive.prevrec;  /* Done this recursion */
2730            recursing + 1, rc));  
2731            DPRINTF(("%.*sReturn from regex recursion: rc=%d\n", rlevel*2-2, SP,
2732              rc));
2733    
2734          /* Ran out of internal offsets */          /* Ran out of internal offsets */
2735    
# Line 1863  for (;;) Line 2743  for (;;)
2743            {            {
2744            for (rc = rc*2 - 2; rc >= 0; rc -= 2)            for (rc = rc*2 - 2; rc >= 0; rc -= 2)
2745              {              {
             const uschar *p = start_subject + local_offsets[rc];  
             const uschar *pp = start_subject + local_offsets[rc+1];  
2746              int charcount = local_offsets[rc+1] - local_offsets[rc];              int charcount = local_offsets[rc+1] - local_offsets[rc];
2747              while (p < pp) if ((*p++ & 0xc0) == 0x80) charcount--;  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2748                if (utf)
2749                  {
2750                  const pcre_uchar *p = start_subject + local_offsets[rc];
2751                  const pcre_uchar *pp = start_subject + local_offsets[rc+1];
2752                  while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--;
2753                  }
2754    #endif
2755              if (charcount > 0)              if (charcount > 0)
2756                {                {
2757                ADD_NEW_DATA(-(state_offset + LINK_SIZE + 1), 0, (charcount - 1));                ADD_NEW_DATA(-(state_offset + LINK_SIZE + 1), 0, (charcount - 1));
# Line 1882  for (;;) Line 2767  for (;;)
2767        break;        break;
2768    
2769        /*-----------------------------------------------------------------*/        /*-----------------------------------------------------------------*/
2770          case OP_BRAPOS:
2771          case OP_SBRAPOS:
2772          case OP_CBRAPOS:
2773          case OP_SCBRAPOS:
2774          case OP_BRAPOSZERO:
2775            {
2776            int charcount, matched_count;
2777            const pcre_uchar *local_ptr = ptr;
2778            BOOL allow_zero;
2779    
2780            if (codevalue == OP_BRAPOSZERO)
2781              {
2782              allow_zero = TRUE;
2783              codevalue = *(++code);  /* Codevalue will be one of above BRAs */
2784              }
2785            else allow_zero = FALSE;
2786    
2787            /* Loop to match the subpattern as many times as possible as if it were
2788            a complete pattern. */
2789    
2790            for (matched_count = 0;; matched_count++)
2791              {
2792              int local_offsets[2];
2793              int local_workspace[1000];
2794    
2795              int rc = internal_dfa_exec(
2796                md,                                   /* fixed match data */
2797                code,                                 /* this subexpression's code */
2798                local_ptr,                            /* where we currently are */
2799                (int)(ptr - start_subject),           /* start offset */
2800                local_offsets,                        /* offset vector */
2801                sizeof(local_offsets)/sizeof(int),    /* size of same */
2802                local_workspace,                      /* workspace vector */
2803                sizeof(local_workspace)/sizeof(int),  /* size of same */
2804                rlevel);                              /* function recursion level */
2805    
2806              /* Failed to match */
2807    
2808              if (rc < 0)
2809                {
2810                if (rc != PCRE_ERROR_NOMATCH) return rc;
2811                break;
2812                }
2813    
2814              /* Matched: break the loop if zero characters matched. */
2815    
2816              charcount = local_offsets[1] - local_offsets[0];
2817              if (charcount == 0) break;
2818              local_ptr += charcount;    /* Advance temporary position ptr */
2819              }
2820    
2821            /* At this point we have matched the subpattern matched_count
2822            times, and local_ptr is pointing to the character after the end of the
2823            last match. */
2824    
2825            if (matched_count > 0 || allow_zero)
2826              {
2827              const pcre_uchar *end_subpattern = code;
2828              int next_state_offset;
2829    
2830              do { end_subpattern += GET(end_subpattern, 1); }
2831                while (*end_subpattern == OP_ALT);
2832              next_state_offset =
2833                (int)(end_subpattern - start_code + LINK_SIZE + 1);
2834    
2835              /* Optimization: if there are no more active states, and there
2836              are no new states yet set up, then skip over the subject string
2837              right here, to save looping. Otherwise, set up the new state to swing
2838              into action when the end of the matched substring is reached. */
2839    
2840              if (i + 1 >= active_count && new_count == 0)
2841                {
2842                ptr = local_ptr;
2843                clen = 0;
2844                ADD_NEW(next_state_offset, 0);
2845                }
2846              else
2847                {
2848                const pcre_uchar *p = ptr;
2849                const pcre_uchar *pp = local_ptr;
2850                charcount = (int)(pp - p);
2851    #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2852                if (utf) while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--;
2853    #endif
2854                ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1));
2855                }
2856              }
2857            }
2858          break;
2859    
2860          /*-----------------------------------------------------------------*/
2861        case OP_ONCE:        case OP_ONCE:
2862          case OP_ONCE_NC:
2863          {          {
2864          int local_offsets[2];          int local_offsets[2];
2865          int local_workspace[1000];          int local_workspace[1000];
# Line 1891  for (;;) Line 2868  for (;;)
2868            md,                                   /* fixed match data */            md,                                   /* fixed match data */
2869            code,                                 /* this subexpression's code */            code,                                 /* this subexpression's code */
2870            ptr,                                  /* where we currently are */            ptr,                                  /* where we currently are */
2871            ptr - start_subject,                  /* start offset */            (int)(ptr - start_subject),           /* start offset */
2872            local_offsets,                        /* offset vector */            local_offsets,                        /* offset vector */
2873            sizeof(local_offsets)/sizeof(int),    /* size of same */            sizeof(local_offsets)/sizeof(int),    /* size of same */
2874            local_workspace,                      /* workspace vector */            local_workspace,                      /* workspace vector */
2875            sizeof(local_workspace)/sizeof(int),  /* size of same */            sizeof(local_workspace)/sizeof(int),  /* size of same */
2876            ims,                                  /* the current ims flags */            rlevel);                              /* function recursion level */
           rlevel,                               /* function recursion level */  
           recursing);                           /* pass on regex recursion */  
2877    
2878          if (rc >= 0)          if (rc >= 0)
2879            {            {
2880            const uschar *end_subpattern = code;            const pcre_uchar *end_subpattern = code;
2881            int charcount = local_offsets[1] - local_offsets[0];            int charcount = local_offsets[1] - local_offsets[0];
2882            int next_state_offset, repeat_state_offset;            int next_state_offset, repeat_state_offset;
2883    
2884            do { end_subpattern += GET(end_subpattern, 1); }            do { end_subpattern += GET(end_subpattern, 1); }
2885              while (*end_subpattern == OP_ALT);              while (*end_subpattern == OP_ALT);
2886            next_state_offset = end_subpattern - start_code + LINK_SIZE + 1;            next_state_offset =
2887                (int)(end_subpattern - start_code + LINK_SIZE + 1);
2888    
2889            /* If the end of this subpattern is KETRMAX or KETRMIN, we must            /* If the end of this subpattern is KETRMAX or KETRMIN, we must
2890            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 1916  for (;;) Line 2892  for (;;)
2892    
2893            repeat_state_offset = (*end_subpattern == OP_KETRMAX ||            repeat_state_offset = (*end_subpattern == OP_KETRMAX ||
2894                                   *end_subpattern == OP_KETRMIN)?                                   *end_subpattern == OP_KETRMIN)?
2895              end_subpattern - start_code - GET(end_subpattern, 1) : -1;              (int)(end_subpattern - start_code - GET(end_subpattern, 1)) : -1;
2896    
2897            /* 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
2898            current character pointer. This is important so that the duplicate            current character pointer. This is important so that the duplicate
# Line 1931  for (;;) Line 2907  for (;;)
2907            /* Optimization: if there are no more active states, and there            /* Optimization: if there are no more active states, and there
2908            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
2909            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
2910            into action when the end of the substring is reached. */            into action when the end of the matched substring is reached. */
2911    
2912            else if (i + 1 >= active_count && new_count == 0)            else if (i + 1 >= active_count && new_count == 0)
2913              {              {
# Line 1954  for (;;) Line 2930  for (;;)
2930              }              }
2931            else            else
2932              {              {
2933              const uschar *p = start_subject + local_offsets[0];  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2934              const uschar *pp = start_subject + local_offsets[1];              if (utf)
2935              while (p < pp) if ((*p++ & 0xc0) == 0x80) charcount--;                {
2936                  const pcre_uchar *p = start_subject + local_offsets[0];
2937                  const pcre_uchar *pp = start_subject + local_offsets[1];
2938                  while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--;
2939                  }
2940    #endif
2941              ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1));              ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1));
2942              if (repeat_state_offset >= 0)              if (repeat_state_offset >= 0)
2943                { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); }                { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); }
2944              }              }
   
2945            }            }
2946          else if (rc != PCRE_ERROR_NOMATCH) return rc;          else if (rc != PCRE_ERROR_NOMATCH) return rc;
2947          }          }
# Line 1972  for (;;) Line 2952  for (;;)
2952        /* Handle callouts */        /* Handle callouts */
2953    
2954        case OP_CALLOUT:        case OP_CALLOUT:
2955        if (pcre_callout != NULL)        rrc = 0;
2956          if (PUBL(callout) != NULL)
2957          {          {
2958          int rrc;          PUBL(callout_block) cb;
         pcre_callout_block cb;  
2959          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 1;   /* Version 1 of the callout block */
2960          cb.callout_number   = code[1];          cb.callout_number   = code[1];
2961          cb.offset_vector    = offsets;          cb.offset_vector    = offsets;
2962    #if defined COMPILE_PCRE8
2963          cb.subject          = (PCRE_SPTR)start_subject;          cb.subject          = (PCRE_SPTR)start_subject;
2964          cb.subject_length   = end_subject - start_subject;  #elif defined COMPILE_PCRE16
2965          cb.start_match      = current_subject - start_subject;          cb.subject          = (PCRE_SPTR16)start_subject;
2966          cb.current_position = ptr - start_subject;  #elif defined COMPILE_PCRE32
2967            cb.subject          = (PCRE_SPTR32)start_subject;
2968    #endif
2969            cb.subject_length   = (int)(end_subject - start_subject);
2970            cb.start_match      = (int)(current_subject - start_subject);
2971            cb.current_position = (int)(ptr - start_subject);
2972          cb.pattern_position = GET(code, 2);          cb.pattern_position = GET(code, 2);
2973          cb.next_item_length = GET(code, 2 + LINK_SIZE);          cb.next_item_length = GET(code, 2 + LINK_SIZE);
2974          cb.capture_top      = 1;          cb.capture_top      = 1;
2975          cb.capture_last     = -1;          cb.capture_last     = -1;
2976          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
2977          if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc;   /* Abandon */          cb.mark             = NULL;   /* No (*MARK) support */
2978          if (rrc == 0) { ADD_ACTIVE(state_offset + 2 + 2*LINK_SIZE, 0); }          if ((rrc = (*PUBL(callout))(&cb)) < 0) return rrc;   /* Abandon */
2979          }          }
2980          if (rrc == 0)
2981            { ADD_ACTIVE(state_offset + PRIV(OP_lengths)[OP_CALLOUT], 0); }
2982        break;        break;
2983    
2984    
# Line 2006  for (;;) Line 2994  for (;;)
2994    /* We have finished the processing at the current subject character. If no    /* We have finished the processing at the current subject character. If no
2995    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
2996    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
2997    matching has been requested, check for appropriate conditions. */    matching has been requested, check for appropriate conditions.
2998    
2999      The "forced_ fail" variable counts the number of (*F) encountered for the
3000      character. If it is equal to the original active_count (saved in
3001      workspace[1]) it means that (*F) was found on every active state. In this
3002      case we don't want to give a partial match.
3003    
3004      The "could_continue" variable is true if a state could have continued but
3005      for the fact that the end of the subject was reached. */
3006    
3007    if (new_count <= 0)    if (new_count <= 0)
3008      {      {
3009      if (match_count < 0 &&                     /* No matches found */      if (rlevel == 1 &&                               /* Top level, and */
3010          rlevel == 1 &&                         /* Top level match function */          could_continue &&                            /* Some could go on, and */
3011          (md->moptions & PCRE_PARTIAL) != 0 &&  /* Want partial matching */          forced_fail != workspace[1] &&               /* Not all forced fail & */
3012          ptr >= end_subject &&                  /* Reached end of subject */          (                                            /* either... */
3013          ptr > current_subject)                 /* Matched non-empty string */          (md->moptions & PCRE_PARTIAL_HARD) != 0      /* Hard partial */
3014            ||                                           /* or... */
3015            ((md->moptions & PCRE_PARTIAL_SOFT) != 0 &&  /* Soft partial and */
3016             match_count < 0)                            /* no matches */
3017            ) &&                                         /* And... */
3018            (
3019            partial_newline ||                           /* Either partial NL */
3020              (                                          /* or ... */
3021              ptr >= end_subject &&                /* End of subject and */
3022              ptr > md->start_used_ptr)            /* Inspected non-empty string */
3023              )
3024            )
3025        {        {
3026        if (offsetcount >= 2)        if (offsetcount >= 2)
3027          {          {
3028          offsets[0] = current_subject - start_subject;          offsets[0] = (int)(md->start_used_ptr - start_subject);
3029          offsets[1] = end_subject - start_subject;          offsets[1] = (int)(end_subject - start_subject);
3030          }          }
3031        match_count = PCRE_ERROR_PARTIAL;        match_count = PCRE_ERROR_PARTIAL;
3032        }        }
# Line 2073  Returns:          > 0 => number of match Line 3080  Returns:          > 0 => number of match
3080                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
3081  */  */
3082    
3083  PCRE_EXP_DEFN int  #if defined COMPILE_PCRE8
3084    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
3085  pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data,
3086    const char *subject, int length, int start_offset, int options, int *offsets,    const char *subject, int length, int start_offset, int options, int *offsets,
3087    int offsetcount, int *workspace, int wscount)    int offsetcount, int *workspace, int wscount)
3088    #elif defined COMPILE_PCRE16
3089    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
3090    pcre16_dfa_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
3091      PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
3092      int offsetcount, int *workspace, int wscount)
3093    #elif defined COMPILE_PCRE32
3094    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
3095    pcre32_dfa_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
3096      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
3097      int offsetcount, int *workspace, int wscount)
3098    #endif
3099  {  {
3100  real_pcre *re = (real_pcre *)argument_re;  REAL_PCRE *re = (REAL_PCRE *)argument_re;
3101  dfa_match_data match_block;  dfa_match_data match_block;
3102  dfa_match_data *md = &match_block;  dfa_match_data *md = &match_block;
3103  BOOL utf8, anchored, startline, firstline;  BOOL utf, anchored, startline, firstline;
3104  const uschar *current_subject, *end_subject, *lcc;  const pcre_uchar *current_subject, *end_subject;
   
 pcre_study_data internal_study;  
3105  const pcre_study_data *study = NULL;  const pcre_study_data *study = NULL;
 real_pcre internal_re;  
3106    
3107  const uschar *req_byte_ptr;  const pcre_uchar *req_char_ptr;
3108  const uschar *start_bits = NULL;  const pcre_uint8 *start_bits = NULL;
3109  BOOL first_byte_caseless = FALSE;  BOOL has_first_char = FALSE;
3110  BOOL req_byte_caseless = FALSE;  BOOL has_req_char = FALSE;
3111  int first_byte = -1;  pcre_uchar first_char = 0;
3112  int req_byte = -1;  pcre_uchar first_char2 = 0;
3113  int req_byte2 = -1;  pcre_uchar req_char = 0;
3114    pcre_uchar req_char2 = 0;
3115  int newline;  int newline;
3116    
3117  /* Plausibility checks */  /* Plausibility checks */
# Line 2104  if (re == NULL || subject == NULL || wor Line 3121  if (re == NULL || subject == NULL || wor
3121     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;
3122  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
3123  if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE;  if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE;
3124    if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
3125    
3126  /* We need to find the pointer to any study data before we test for byte  /* Check that the first field in the block is the magic number. If it is not,
3127  flipping, so we scan the extra_data block first. This may set two fields in the  return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to
3128  match block, so we must initialize them beforehand. However, the other fields  REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which
3129  in the match block must not be set until after the byte flipping. */  means that the pattern is likely compiled with different endianness. */
3130    
3131    if (re->magic_number != MAGIC_NUMBER)
3132      return re->magic_number == REVERSED_MAGIC_NUMBER?
3133        PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC;
3134    if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE;
3135    
3136    /* If restarting after a partial match, do some sanity checks on the contents
3137    of the workspace. */
3138    
3139    if ((options & PCRE_DFA_RESTART) != 0)
3140      {
3141      if ((workspace[0] & (-2)) != 0 || workspace[1] < 1 ||
3142        workspace[1] > (wscount - 2)/INTS_PER_STATEBLOCK)
3143          return PCRE_ERROR_DFA_BADRESTART;
3144      }
3145    
3146    /* Set up study, callout, and table data */
3147    
3148  md->tables = re->tables;  md->tables = re->tables;
3149  md->callout_data = NULL;  md->callout_data = NULL;
# Line 2127  if (extra_data != NULL) Line 3162  if (extra_data != NULL)
3162      md->tables = extra_data->tables;      md->tables = extra_data->tables;
3163    }    }
3164    
 /* 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;  
   }  
   
3165  /* Set some local values */  /* Set some local values */
3166    
3167  current_subject = (const unsigned char *)subject + start_offset;  current_subject = (const pcre_uchar *)subject + start_offset;
3168  end_subject = (const unsigned char *)subject + length;  end_subject = (const pcre_uchar *)subject + length;
3169  req_byte_ptr = current_subject - 1;  req_char_ptr = current_subject - 1;
3170    
3171  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3172  utf8 = (re->options & PCRE_UTF8) != 0;  /* PCRE_UTF(16|32) have the same value as PCRE_UTF8. */
3173    utf = (re->options & PCRE_UTF8) != 0;
3174  #else  #else
3175  utf8 = FALSE;  utf = FALSE;
3176  #endif  #endif
3177    
3178  anchored = (options & (PCRE_ANCHORED|PCRE_DFA_RESTART)) != 0 ||  anchored = (options & (PCRE_ANCHORED|PCRE_DFA_RESTART)) != 0 ||
# Line 2156  anchored = (options & (PCRE_ANCHORED|PCR Line 3180  anchored = (options & (PCRE_ANCHORED|PCR
3180    
3181  /* The remaining fixed data for passing around. */  /* The remaining fixed data for passing around. */
3182    
3183  md->start_code = (const uschar *)argument_re +  md->start_code = (const pcre_uchar *)argument_re +
3184      re->name_table_offset + re->name_count * re->name_entry_size;      re->name_table_offset + re->name_count * re->name_entry_size;
3185  md->start_subject = (const unsigned char *)subject;  md->start_subject = (const pcre_uchar *)subject;
3186  md->end_subject = end_subject;  md->end_subject = end_subject;
3187    md->start_offset = start_offset;
3188  md->moptions = options;  md->moptions = options;
3189  md->poptions = re->options;  md->poptions = re->options;
3190    
3191    /* If the BSR option is not set at match time, copy what was set
3192    at compile time. */
3193    
3194    if ((md->moptions & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == 0)
3195      {
3196      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
3197        md->moptions |= re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE);
3198    #ifdef BSR_ANYCRLF
3199      else md->moptions |= PCRE_BSR_ANYCRLF;
3200    #endif
3201      }
3202    
3203  /* Handle different types of newline. The three bits give eight cases. If  /* Handle different types of newline. The three bits give eight cases. If
3204  nothing is set at run time, whatever was used at compile time applies. */  nothing is set at run time, whatever was used at compile time applies. */
3205    
# Line 2170  switch ((((options & PCRE_NEWLINE_BITS) Line 3207  switch ((((options & PCRE_NEWLINE_BITS)
3207           PCRE_NEWLINE_BITS)           PCRE_NEWLINE_BITS)
3208    {    {
3209    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
3210    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = CHAR_CR; break;
3211    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = CHAR_NL; break;
3212    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
3213         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break;
3214    case PCRE_NEWLINE_ANY: newline = -1; break;    case PCRE_NEWLINE_ANY: newline = -1; break;
3215    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
3216    default: return PCRE_ERROR_BADNEWLINE;    default: return PCRE_ERROR_BADNEWLINE;
# Line 2206  else Line 3243  else
3243  /* 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
3244  back the character offset. */  back the character offset. */
3245    
3246  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3247  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0)
3248    {    {
3249    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    int erroroffset;
3250      return PCRE_ERROR_BADUTF8;    int errorcode = PRIV(valid_utf)((pcre_uchar *)subject, length, &erroroffset);
3251    if (start_offset > 0 && start_offset < length)    if (errorcode != 0)
3252      {      {
3253      int tb = ((uschar *)subject)[start_offset];      if (offsetcount >= 2)
     if (tb > 127)  
3254        {        {
3255        tb &= 0xc0;        offsets[0] = erroroffset;
3256        if (tb != 0 && tb != 0xc0) return PCRE_ERROR_BADUTF8_OFFSET;        offsets[1] = errorcode;
3257        }        }
3258    #if defined COMPILE_PCRE8
3259        return (errorcode <= PCRE_UTF8_ERR5 && (options & PCRE_PARTIAL_HARD) != 0) ?
3260          PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
3261    #elif defined COMPILE_PCRE16
3262        return (errorcode <= PCRE_UTF16_ERR1 && (options & PCRE_PARTIAL_HARD) != 0) ?
3263          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
3264    #elif defined COMPILE_PCRE32
3265        return PCRE_ERROR_BADUTF32;
3266    #endif
3267      }      }
3268    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
3269      if (start_offset > 0 && start_offset < length &&
3270            NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
3271        return PCRE_ERROR_BADUTF8_OFFSET;
3272    #endif
3273    }    }
3274  #endif  #endif
3275    
# Line 2227  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 3277  if (utf8 && (options & PCRE_NO_UTF8_CHEC
3277  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
3278  in other programs later. */  in other programs later. */
3279    
3280  if (md->tables == NULL) md->tables = _pcre_default_tables;  if (md->tables == NULL) md->tables = PRIV(default_tables);
3281    
3282  /* 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
3283  used in a loop when finding where to start. */  where to start. */
3284    
3285  lcc = md->tables + lcc_offset;  startline = (re->flags & PCRE_STARTLINE) != 0;
 startline = (re->options & PCRE_STARTLINE) != 0;  
3286  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
3287    
3288  /* Set up the first character to match, if available. The first_byte value is  /* Set up the first character to match, if available. The first_byte value is
# Line 2244  studied, there may be a bitmap of possib Line 3293  studied, there may be a bitmap of possib
3293    
3294  if (!anchored)  if (!anchored)
3295    {    {
3296    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
3297      {      {
3298      first_byte = re->first_byte & 255;      has_first_char = TRUE;
3299      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      first_char = first_char2 = (pcre_uchar)(re->first_char);
3300        first_byte = lcc[first_byte];      if ((re->flags & PCRE_FCH_CASELESS) != 0)
3301          {
3302          first_char2 = TABLE_GET(first_char, md->tables + fcc_offset, first_char);
3303    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
3304          if (utf && first_char > 127)
3305            first_char2 = UCD_OTHERCASE(first_char);
3306    #endif
3307          }
3308      }      }
3309    else    else
3310      {      {
3311      if (startline && study != NULL &&      if (!startline && study != NULL &&
3312           (study->options & PCRE_STUDY_MAPPED) != 0)           (study->flags & PCRE_STUDY_MAPPED) != 0)
3313        start_bits = study->start_bits;        start_bits = study->start_bits;
3314      }      }
3315    }    }
# Line 2261  if (!anchored) Line 3317  if (!anchored)
3317  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
3318  character" set. */  character" set. */
3319    
3320  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
3321    {    {
3322    req_byte = re->req_byte & 255;    has_req_char = TRUE;
3323    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_char = req_char2 = (pcre_uchar)(re->req_char);
3324    req_byte2 = (md->tables + fcc_offset)[req_byte];  /* case flipped */    if ((re->flags & PCRE_RCH_CASELESS) != 0)
3325        {
3326        req_char2 = TABLE_GET(req_char, md->tables + fcc_offset, req_char);
3327    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
3328        if (utf && req_char > 127)
3329          req_char2 = UCD_OTHERCASE(req_char);
3330    #endif
3331        }
3332    }    }
3333    
3334  /* 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
3335  failed match. Unless restarting, optimize by moving to the first match  failed match. If not restarting, perform certain optimizations at the start of
3336  character if possible, when not anchored. Then unless wanting a partial match,  a match. */
 check for a required later character. */  
3337    
3338  for (;;)  for (;;)
3339    {    {
# Line 2279  for (;;) Line 3341  for (;;)
3341    
3342    if ((options & PCRE_DFA_RESTART) == 0)    if ((options & PCRE_DFA_RESTART) == 0)
3343      {      {
3344      const uschar *save_end_subject = end_subject;      const pcre_uchar *save_end_subject = end_subject;
3345    
3346      /* Advance to a unique first char if possible. If firstline is TRUE, the      /* If firstline is TRUE, the start of the match is constrained to the first
3347      start of the match is constrained to the first line of a multiline string.      line of a multiline string. Implement this by temporarily adjusting
3348      Implement this by temporarily adjusting end_subject so that we stop      end_subject so that we stop scanning at a newline. If the match fails at
3349      scanning at a newline. If the match fails at the newline, later code breaks      the newline, later code breaks this loop. */
     this loop. */  
3350    
3351      if (firstline)      if (firstline)
3352        {        {
3353        const uschar *t = current_subject;        PCRE_PUCHAR t = current_subject;
3354    #ifdef SUPPORT_UTF
3355          if (utf)
3356            {
3357            while (t < md->end_subject && !IS_NEWLINE(t))
3358              {
3359              t++;
3360              ACROSSCHAR(t < end_subject, *t, t++);
3361              }
3362            }
3363          else
3364    #endif
3365        while (t < md->end_subject && !IS_NEWLINE(t)) t++;        while (t < md->end_subject && !IS_NEWLINE(t)) t++;
3366        end_subject = t;        end_subject = t;
3367        }        }
3368    
3369      if (first_byte >= 0)      /* There are some optimizations that avoid running the match if a known
3370        starting point is not found. However, there is an option that disables
3371        these, for testing and for ensuring that all callouts do actually occur.
3372        The option can be set in the regex by (*NO_START_OPT) or passed in
3373        match-time options. */
3374    
3375        if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)
3376        {        {
3377        if (first_byte_caseless)        /* Advance to a known first char. */
         while (current_subject < end_subject &&  
                lcc[*current_subject] != first_byte)  
           current_subject++;  
       else  
         while (current_subject < end_subject && *current_subject != first_byte)  
           current_subject++;  
       }  
3378    
3379      /* Or to just after a linebreak for a multiline match if possible */        if (has_first_char)
3380            {
3381            if (first_char != first_char2)
3382              {
3383              pcre_uchar csc;
3384              while (current_subject < end_subject &&
3385                     (csc = RAWUCHARTEST(current_subject)) != first_char && csc != first_char2)
3386                current_subject++;
3387              }
3388            else
3389              while (current_subject < end_subject &&
3390                     RAWUCHARTEST(current_subject) != first_char)
3391                current_subject++;
3392            }
3393    
3394      else if (startline)        /* Or to just after a linebreak for a multiline match if possible */
3395        {  
3396        if (current_subject > md->start_subject + start_offset)        else if (startline)
3397          {          {
3398          while (current_subject <= end_subject && !WAS_NEWLINE(current_subject))          if (current_subject > md->start_subject + start_offset)
3399            current_subject++;            {
3400    #ifdef SUPPORT_UTF
3401              if (utf)
3402                {
3403                while (current_subject < end_subject &&
3404                       !WAS_NEWLINE(current_subject))
3405                  {
3406                  current_subject++;
3407                  ACROSSCHAR(current_subject < end_subject, *current_subject,
3408                    current_subject++);
3409                  }
3410                }
3411              else
3412    #endif
3413              while (current_subject < end_subject && !WAS_NEWLINE(current_subject))
3414                current_subject++;
3415    
3416          /* If we have just passed a CR and the newline option is ANY or            /* If we have just passed a CR and the newline option is ANY or
3417          ANYCRLF, and we are now at a LF, advance the match position by one more            ANYCRLF, and we are now at a LF, advance the match position by one
3418          character. */            more character. */
3419    
3420          if (current_subject[-1] == '\r' &&            if (RAWUCHARTEST(current_subject - 1) == CHAR_CR &&
3421               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&                 (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
3422               current_subject < end_subject &&                 current_subject < end_subject &&
3423               *current_subject == '\n')                 RAWUCHARTEST(current_subject) == CHAR_NL)
3424            current_subject++;              current_subject++;
3425              }
3426          }          }
       }  
3427    
3428      /* Or to a non-unique first char after study */        /* Or to a non-unique first char after study */
3429    
3430      else if (start_bits != NULL)        else if (start_bits != NULL)
       {  
       while (current_subject < end_subject)  
3431          {          {
3432          register unsigned int c = *current_subject;          while (current_subject < end_subject)
3433          if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++;            {
3434              register pcre_uint32 c = RAWUCHARTEST(current_subject);
3435    #ifndef COMPILE_PCRE8
3436              if (c > 255) c = 255;
3437    #endif
3438              if ((start_bits[c/8] & (1 << (c&7))) == 0)
3439                {
3440                current_subject++;
3441    #if defined SUPPORT_UTF && defined COMPILE_PCRE8
3442                /* In non 8-bit mode, the iteration will stop for
3443                characters > 255 at the beginning or not stop at all. */
3444                if (utf)
3445                  ACROSSCHAR(current_subject < end_subject, *current_subject,
3446                    current_subject++);
3447    #endif
3448                }
3449            else break;            else break;
3450              }
3451          }          }
3452        }        }
3453    
3454      /* Restore fudged end_subject */      /* Restore fudged end_subject */
3455    
3456      end_subject = save_end_subject;      end_subject = save_end_subject;
     }  
3457    
3458    /* If req_byte is set, we know that that character must appear in the subject      /* The following two optimizations are disabled for partial matching or if
3459    for the match to succeed. If the first character is set, req_byte must be      disabling is explicitly requested (and of course, by the test above, this
3460    later in the subject; otherwise the test starts at the match point. This      code is not obeyed when restarting after a partial match). */
   optimization can save a huge amount of work in patterns with nested unlimited  
   repeats that aren't going to match. Writing separate code for cased/caseless  
   versions makes it go faster, as does using an autoincrement and backing off  
   on a match.  
   
   HOWEVER: when the subject string is very, very long, searching to its end can  
   take a long time, and give bad performance on quite ordinary patterns. This  
   showed up when somebody was matching /^C/ on a 32-megabyte string... so we  
   don't do this when the string is sufficiently long.  
   
   ALSO: this processing is disabled when partial matching is requested.  
   */  
   
   if (req_byte >= 0 &&  
       end_subject - current_subject < REQ_BYTE_MAX &&  
       (options & PCRE_PARTIAL) == 0)  
     {  
     register const uschar *p = current_subject + ((first_byte >= 0)? 1 : 0);  
3461    
3462      /* We don't need to repeat the search if we haven't yet reached the      if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0 &&
3463      place we found it at last time. */          (options & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) == 0)
   
     if (p > req_byte_ptr)  
3464        {        {
3465        if (req_byte_caseless)        /* If the pattern was studied, a minimum subject length may be set. This
3466          {        is a lower bound; no actual string of that length may actually match the
3467          while (p < end_subject)        pattern. Although the value is, strictly, in characters, we treat it as
3468            {        bytes to avoid spending too much time in this optimization. */
3469            register int pp = *p++;  
3470            if (pp == req_byte || pp == req_byte2) { p--; break; }        if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 &&
3471            }            (pcre_uint32)(end_subject - current_subject) < study->minlength)
3472          }          return PCRE_ERROR_NOMATCH;
3473        else  
3474          /* If req_char is set, we know that that character must appear in the
3475          subject for the match to succeed. If the first character is set, req_char
3476          must be later in the subject; otherwise the test starts at the match
3477          point. This optimization can save a huge amount of work in patterns with
3478          nested unlimited repeats that aren't going to match. Writing separate
3479          code for cased/caseless versions makes it go faster, as does using an
3480          autoincrement and backing off on a match.
3481    
3482          HOWEVER: when the subject string is very, very long, searching to its end
3483          can take a long time, and give bad performance on quite ordinary
3484          patterns. This showed up when somebody was matching /^C/ on a 32-megabyte
3485          string... so we don't do this when the string is sufficiently long. */
3486    
3487          if (has_req_char && end_subject - current_subject < REQ_BYTE_MAX)
3488          {          {
3489          while (p < end_subject)          register PCRE_PUCHAR p = current_subject + (has_first_char? 1:0);
3490    
3491            /* We don't need to repeat the search if we haven't yet reached the
3492            place we found it at last time. */
3493    
3494            if (p > req_char_ptr)
3495            {            {
3496            if (*p++ == req_byte) { p--; break; }            if (req_char != req_char2)
3497            }              {
3498          }              while (p < end_subject)
3499                  {
3500                  register pcre_uint32 pp = RAWUCHARINCTEST(p);
3501                  if (pp == req_char || pp == req_char2) { p--; break; }
3502                  }
3503                }
3504              else
3505                {
3506                while (p < end_subject)
3507                  {
3508                  if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
3509                  }
3510                }
3511    
3512        /* If we can't find the required character, break the matching loop,            /* If we can't find the required character, break the matching loop,
3513        which will cause a return or PCRE_ERROR_NOMATCH. */            which will cause a return or PCRE_ERROR_NOMATCH. */
3514    
3515        if (p >= end_subject) break;            if (p >= end_subject) break;
3516    
3517        /* If we have found the required character, save the point where we            /* If we have found the required character, save the point where we
3518        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
3519        the start hasn't passed this character yet. */            the start hasn't passed this character yet. */
3520    
3521        req_byte_ptr = p;            req_char_ptr = p;
3522              }
3523            }
3524        }        }
3525      }      }   /* End of optimizations that are done when not restarting */
3526    
3527    /* OK, now we can do the business */    /* OK, now we can do the business */
3528    
3529      md->start_used_ptr = current_subject;
3530      md->recursive = NULL;
3531    
3532    rc = internal_dfa_exec(    rc = internal_dfa_exec(
3533      md,                                /* fixed match data */      md,                                /* fixed match data */
3534      md->start_code,                    /* this subexpression's code */      md->start_code,                    /* this subexpression's code */
# Line 2410  for (;;) Line 3538  for (;;)
3538      offsetcount,                       /* size of same */      offsetcount,                       /* size of same */
3539      workspace,                         /* workspace vector */      workspace,                         /* workspace vector */
3540      wscount,                           /* size of same */      wscount,                           /* size of same */
3541      re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL), /* ims flags */      0);                                /* function recurse level */
     0,                                 /* function recurse level */  
     0);                                /* regex recurse level */  
3542    
3543    /* Anything other than "no match" means we are done, always; otherwise, carry    /* Anything other than "no match" means we are done, always; otherwise, carry
3544    on only if not anchored. */    on only if not anchored. */
# Line 2424  for (;;) Line 3550  for (;;)
3550    
3551    if (firstline && IS_NEWLINE(current_subject)) break;    if (firstline && IS_NEWLINE(current_subject)) break;
3552    current_subject++;    current_subject++;
3553    if (utf8)  #ifdef SUPPORT_UTF
3554      if (utf)
3555      {      {
3556      while (current_subject < end_subject && (*current_subject & 0xc0) == 0x80)      ACROSSCHAR(current_subject < end_subject, *current_subject,