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

Diff of /code/trunk/pcre_exec.c

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

revision 184 by ph10, Tue Jun 19 13:26:46 2007 UTC revision 231 by ph10, Tue Sep 11 11:15:33 2007 UTC
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
43  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
44    
45    #ifdef HAVE_CONFIG_H
46    #include <config.h>
47    #endif
48    
49  #define NLBLOCK md             /* Block containing newline information */  #define NLBLOCK md             /* Block containing newline information */
50  #define PSSTART start_subject  /* Field containing processed string start */  #define PSSTART start_subject  /* Field containing processed string start */
51  #define PSEND   end_subject    /* Field containing processed string end */  #define PSEND   end_subject    /* Field containing processed string end */
# Line 53  possible. There are also some static sup Line 57  possible. There are also some static sup
57  #undef min  #undef min
58  #undef max  #undef max
59    
 /* The chain of eptrblocks for tail recursions uses memory in stack workspace,  
 obtained at top level, the size of which is defined by EPTR_WORK_SIZE. */  
   
 #define EPTR_WORK_SIZE (1000)  
   
60  /* Flag bits for the match() function */  /* Flag bits for the match() function */
61    
62  #define match_condassert     0x01  /* Called to check a condition assertion */  #define match_condassert     0x01  /* Called to check a condition assertion */
63  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */
 #define match_tail_recursed  0x04  /* Tail recursive call */  
64    
65  /* Non-error returns from the match() function. Error returns are externally  /* Non-error returns from the match() function. Error returns are externally
66  defined PCRE_ERROR_xxx codes, which are all negative. */  defined PCRE_ERROR_xxx codes, which are all negative. */
# Line 70  defined PCRE_ERROR_xxx codes, which are Line 68  defined PCRE_ERROR_xxx codes, which are
68  #define MATCH_MATCH        1  #define MATCH_MATCH        1
69  #define MATCH_NOMATCH      0  #define MATCH_NOMATCH      0
70    
71    /* Special internal returns from the match() function. Make them sufficiently
72    negative to avoid the external error codes. */
73    
74    #define MATCH_COMMIT       (-999)
75    #define MATCH_PRUNE        (-998)
76    #define MATCH_SKIP         (-997)
77    #define MATCH_THEN         (-996)
78    
79  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
80  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
81  because the offset vector is always a multiple of 3 long. */  because the offset vector is always a multiple of 3 long. */
# Line 205  variable instead of being passed in the Line 211  variable instead of being passed in the
211  ****************************************************************************  ****************************************************************************
212  ***************************************************************************/  ***************************************************************************/
213    
214    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
215  /* Numbers for RMATCH calls */  below must be updated in sync.  */
216    
217  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
218         RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,         RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
219         RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,         RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
220         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
221         RM41,  RM42, RM43, RM44, RM45, RM46, RM47 };         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
222           RM51,  RM52, RM53, RM54 };
223    
224  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
225  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 384  Arguments: Line 390  Arguments:
390                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
391                   match_cbegroup - this is the start of an unlimited repeat                   match_cbegroup - this is the start of an unlimited repeat
392                     group that can match an empty string                     group that can match an empty string
                  match_tail_recursed - this is a tail_recursed group  
393     rdepth      the recursion depth     rdepth      the recursion depth
394    
395  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
# Line 586  original_ims = ims; /* Save for reset Line 591  original_ims = ims; /* Save for reset
591  string, the match_cbegroup flag is set. When this is the case, add the current  string, the match_cbegroup flag is set. When this is the case, add the current
592  subject pointer to the chain of such remembered pointers, to be checked when we  subject pointer to the chain of such remembered pointers, to be checked when we
593  hit the closing ket, in order to break infinite loops that match no characters.  hit the closing ket, in order to break infinite loops that match no characters.
594  When match() is called in other circumstances, don't add to the chain. If this  When match() is called in other circumstances, don't add to the chain. The
595  is a tail recursion, use a block from the workspace, as the one on the stack is  match_cbegroup flag must NOT be used with tail recursion, because the memory
596  already used. */  block that is used is on the stack, so a new one may be required for each
597    match(). */
598    
599  if ((flags & match_cbegroup) != 0)  if ((flags & match_cbegroup) != 0)
600    {    {
601    eptrblock *p;    newptrb.epb_saved_eptr = eptr;
602    if ((flags & match_tail_recursed) != 0)    newptrb.epb_prev = eptrb;
603      {    eptrb = &newptrb;
     if (md->eptrn >= EPTR_WORK_SIZE) RRETURN(PCRE_ERROR_NULLWSLIMIT);  
     p = md->eptrchain + md->eptrn++;  
     }  
   else p = &newptrb;  
   p->epb_saved_eptr = eptr;  
   p->epb_prev = eptrb;  
   eptrb = p;  
604    }    }
605    
606  /* Now start processing the opcodes. */  /* Now start processing the opcodes. */
# Line 621  for (;;) Line 620  for (;;)
620    
621    switch(op)    switch(op)
622      {      {
623        case OP_FAIL:
624        RRETURN(MATCH_NOMATCH);
625    
626        case OP_PRUNE:
627        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
628          ims, eptrb, flags, RM51);
629        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
630        RRETURN(MATCH_PRUNE);
631    
632        case OP_COMMIT:
633        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
634          ims, eptrb, flags, RM52);
635        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
636        RRETURN(MATCH_COMMIT);
637    
638        case OP_SKIP:
639        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
640          ims, eptrb, flags, RM53);
641        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
642        md->start_match_ptr = eptr;   /* Pass back current position */
643        RRETURN(MATCH_SKIP);
644    
645        case OP_THEN:
646        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
647          ims, eptrb, flags, RM54);
648        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
649        RRETURN(MATCH_THEN);
650    
651      /* Handle a capturing bracket. If there is space in the offset vector, save      /* Handle a capturing bracket. If there is space in the offset vector, save
652      the current subject position in the working slot at the top of the vector.      the current subject position in the working slot at the top of the vector.
653      We mustn't change the current values of the data slot, because they may be      We mustn't change the current values of the data slot, because they may be
# Line 662  for (;;) Line 689  for (;;)
689          {          {
690          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
691            ims, eptrb, flags, RM1);            ims, eptrb, flags, RM1);
692          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
693          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
694          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
695          }          }
# Line 677  for (;;) Line 704  for (;;)
704        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
705        }        }
706    
707      /* Insufficient room for saving captured contents. Treat as a non-capturing      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
708      bracket. */      as a non-capturing bracket. */
709    
710        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
711        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
712    
713      DPRINTF(("insufficient capture room: treat as non-capturing\n"));      DPRINTF(("insufficient capture room: treat as non-capturing\n"));
714    
715        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
716        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
717    
718      /* Non-capturing bracket. Loop for all the alternatives. When we get to the      /* Non-capturing bracket. Loop for all the alternatives. When we get to the
719      final alternative within the brackets, we would return the result of a      final alternative within the brackets, we would return the result of a
720      recursive call to match() whatever happened. We can reduce stack usage by      recursive call to match() whatever happened. We can reduce stack usage by
721      turning this into a tail recursion. */      turning this into a tail recursion, except in the case when match_cbegroup
722        is set.*/
723    
724      case OP_BRA:      case OP_BRA:
725      case OP_SBRA:      case OP_SBRA:
# Line 693  for (;;) Line 727  for (;;)
727      flags = (op >= OP_SBRA)? match_cbegroup : 0;      flags = (op >= OP_SBRA)? match_cbegroup : 0;
728      for (;;)      for (;;)
729        {        {
730        if (ecode[GET(ecode, 1)] != OP_ALT)        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
731          {          {
732          ecode += _pcre_OP_lengths[*ecode];          if (flags == 0)    /* Not a possibly empty group */
733          flags |= match_tail_recursed;            {
734          DPRINTF(("bracket 0 tail recursion\n"));            ecode += _pcre_OP_lengths[*ecode];
735          goto TAIL_RECURSE;            DPRINTF(("bracket 0 tail recursion\n"));
736              goto TAIL_RECURSE;
737              }
738    
739            /* Possibly empty group; can't use tail recursion. */
740    
741            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
742              eptrb, flags, RM48);
743            RRETURN(rrc);
744          }          }
745    
746        /* For non-final alternatives, continue the loop for a NOMATCH result;        /* For non-final alternatives, continue the loop for a NOMATCH result;
# Line 706  for (;;) Line 748  for (;;)
748    
749        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
750          eptrb, flags, RM2);          eptrb, flags, RM2);
751        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
752        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
753        }        }
754      /* Control never reaches here. */      /* Control never reaches here. */
# Line 754  for (;;) Line 796  for (;;)
796          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
797          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
798          }          }
799        else if (rrc != MATCH_NOMATCH)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
800          {          {
801          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
802          }          }
# Line 766  for (;;) Line 808  for (;;)
808        }        }
809    
810      /* We are now at the branch that is to be obeyed. As there is only one,      /* We are now at the branch that is to be obeyed. As there is only one,
811      we can use tail recursion to avoid using another stack frame. If the second      we can use tail recursion to avoid using another stack frame, except when
812      alternative doesn't exist, we can just plough on. */      match_cbegroup is required for an unlimited repeat of a possibly empty
813        group. If the second alternative doesn't exist, we can just plough on. */
814    
815      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
816        {        {
817        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
818        flags = match_tail_recursed | ((op == OP_SCOND)? match_cbegroup : 0);        if (op == OP_SCOND)        /* Possibly empty group */
819        goto TAIL_RECURSE;          {
820            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);
821            RRETURN(rrc);
822            }
823          else                       /* Group must match something */
824            {
825            flags = 0;
826            goto TAIL_RECURSE;
827            }
828        }        }
829      else      else                         /* Condition false & no 2nd alternative */
830        {        {
831        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
832        }        }
833      break;      break;
834    
835    
836      /* End of the pattern. If we are in a top-level recursion, we should      /* End of the pattern, either real or forced. If we are in a top-level
837      restore the offsets appropriately and continue from after the call. */      recursion, we should restore the offsets appropriately and continue from
838        after the call. */
839    
840        case OP_ACCEPT:
841      case OP_END:      case OP_END:
842      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL && md->recursive->group_num == 0)
843        {        {
# Line 805  for (;;) Line 858  for (;;)
858      if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH);      if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH);
859      md->end_match_ptr = eptr;           /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
860      md->end_offset_top = offset_top;    /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
861      md->start_match_ptr = mstart;  /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
862      RRETURN(MATCH_MATCH);      RRETURN(MATCH_MATCH);
863    
864      /* Change option settings */      /* Change option settings */
# Line 829  for (;;) Line 882  for (;;)
882        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
883          RM4);          RM4);
884        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
885        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
886        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
887        }        }
888      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 856  for (;;) Line 909  for (;;)
909        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
910          RM5);          RM5);
911        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);
912        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
913        ecode += GET(ecode,1);        ecode += GET(ecode,1);
914        }        }
915      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 880  for (;;) Line 933  for (;;)
933          {          {
934          eptr--;          eptr--;
935          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
936          BACKCHAR(eptr)          BACKCHAR(eptr);
937          }          }
938        }        }
939      else      else
# Line 993  for (;;) Line 1046  for (;;)
1046              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1047            RRETURN(MATCH_MATCH);            RRETURN(MATCH_MATCH);
1048            }            }
1049          else if (rrc != MATCH_NOMATCH)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1050            {            {
1051            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1052            RRETURN(rrc);            RRETURN(rrc);
# Line 1027  for (;;) Line 1080  for (;;)
1080    
1081      do      do
1082        {        {
1083        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
         eptrb, 0, RM7);  
1084        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
1085        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1086        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1087        }        }
1088      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 1073  for (;;) Line 1125  for (;;)
1125    
1126      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1127        {        {
1128        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);
         RM8);  
1129        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1130        ecode = prev;        ecode = prev;
1131        flags = match_tail_recursed;        flags = 0;
1132        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1133        }        }
1134      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
# Line 1085  for (;;) Line 1136  for (;;)
1136        RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);
1137        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1138        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1139        flags = match_tail_recursed;        flags = 0;
1140        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1141        }        }
1142      /* Control never gets here */      /* Control never gets here */
# Line 1216  for (;;) Line 1267  for (;;)
1267    
1268      /* The repeating kets try the rest of the pattern or restart from the      /* The repeating kets try the rest of the pattern or restart from the
1269      preceding bracket, in the appropriate order. In the second case, we can use      preceding bracket, in the appropriate order. In the second case, we can use
1270      tail recursion to avoid using another stack frame. */      tail recursion to avoid using another stack frame, unless we have an
1271        unlimited repeat of a group that can match an empty string. */
1272    
1273      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1274    
1275      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1276        {        {
1277        RMATCH(eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);
         RM12);  
1278        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1279          if (flags != 0)    /* Could match an empty string */
1280            {
1281            RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1282            RRETURN(rrc);
1283            }
1284        ecode = prev;        ecode = prev;
       flags |= match_tail_recursed;  
1285        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1286        }        }
1287      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
# Line 1234  for (;;) Line 1289  for (;;)
1289        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1290        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1291        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1292        flags = match_tail_recursed;        flags = 0;
1293        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1294        }        }
1295      /* Control never gets here */      /* Control never gets here */
# Line 1471  for (;;) Line 1526  for (;;)
1526        case 0x000d:        case 0x000d:
1527        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1528        break;        break;
1529    
1530        case 0x000a:        case 0x000a:
1531          break;
1532    
1533        case 0x000b:        case 0x000b:
1534        case 0x000c:        case 0x000c:
1535        case 0x0085:        case 0x0085:
1536        case 0x2028:        case 0x2028:
1537        case 0x2029:        case 0x2029:
1538          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1539        break;        break;
1540        }        }
1541      ecode++;      ecode++;
# Line 2033  for (;;) Line 2092  for (;;)
2092            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2093            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2094            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2095            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2096            }            }
2097          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2098          }          }
# Line 2899  for (;;) Line 2958  for (;;)
2958              case 0x000d:              case 0x000d:
2959              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2960              break;              break;
2961    
2962              case 0x000a:              case 0x000a:
2963                break;
2964    
2965              case 0x000b:              case 0x000b:
2966              case 0x000c:              case 0x000c:
2967              case 0x0085:              case 0x0085:
2968              case 0x2028:              case 0x2028:
2969              case 0x2029:              case 0x2029:
2970                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2971              break;              break;
2972              }              }
2973            }            }
# Line 3038  for (;;) Line 3101  for (;;)
3101          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3102            {            {
3103            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3104               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3105              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3106            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3107            }            }
3108          break;          break;
3109    
# Line 3058  for (;;) Line 3121  for (;;)
3121          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3122            {            {
3123            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3124               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3125              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3126            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3127            }            }
3128          break;          break;
3129    
# Line 3117  for (;;) Line 3180  for (;;)
3180              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3181              break;              break;
3182              case 0x000a:              case 0x000a:
3183                break;
3184    
3185              case 0x000b:              case 0x000b:
3186              case 0x000c:              case 0x000c:
3187              case 0x0085:              case 0x0085:
3188                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3189              break;              break;
3190              }              }
3191            }            }
# Line 3371  for (;;) Line 3437  for (;;)
3437                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3438                break;                break;
3439                case 0x000a:                case 0x000a:
3440                  break;
3441    
3442                case 0x000b:                case 0x000b:
3443                case 0x000c:                case 0x000c:
3444                case 0x0085:                case 0x0085:
3445                case 0x2028:                case 0x2028:
3446                case 0x2029:                case 0x2029:
3447                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3448                break;                break;
3449                }                }
3450              break;              break;
# Line 3527  for (;;) Line 3596  for (;;)
3596                case 0x000d:                case 0x000d:
3597                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3598                break;                break;
3599    
3600                case 0x000a:                case 0x000a:
3601                  break;
3602    
3603                case 0x000b:                case 0x000b:
3604                case 0x000c:                case 0x000c:
3605                case 0x0085:                case 0x0085:
3606                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3607                break;                break;
3608                }                }
3609              break;              break;
# Line 3702  for (;;) Line 3775  for (;;)
3775            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3776            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3777            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3778            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3779            }            }
3780          }          }
3781    
# Line 3741  for (;;) Line 3814  for (;;)
3814            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3815              {              {
3816              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3817              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3818                {                {
3819                  BACKCHAR(eptr);
3820                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3821                }                }
3822              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
# Line 3764  for (;;) Line 3837  for (;;)
3837          switch(ctype)          switch(ctype)
3838            {            {
3839            case OP_ANY:            case OP_ANY:
   
           /* Special code is required for UTF8, but when the maximum is  
           unlimited we don't need it, so we repeat the non-UTF8 code. This is  
           probably worth it, because .* is quite a common idiom. */  
   
3840            if (max < INT_MAX)            if (max < INT_MAX)
3841              {              {
3842              if ((ims & PCRE_DOTALL) == 0)              if ((ims & PCRE_DOTALL) == 0)
# Line 3801  for (;;) Line 3869  for (;;)
3869                  {                  {
3870                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3871                  eptr++;                  eptr++;
3872                    while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3873                  }                  }
               break;  
3874                }                }
3875              else              else
3876                {                {
3877                c = max - min;                eptr = md->end_subject;
               if (c > (unsigned int)(md->end_subject - eptr))  
                 c = md->end_subject - eptr;  
               eptr += c;  
3878                }                }
3879              }              }
3880            break;            break;
# Line 3836  for (;;) Line 3901  for (;;)
3901                }                }
3902              else              else
3903                {                {
3904                if (c != 0x000a && c != 0x000b && c != 0x000c &&                if (c != 0x000a &&
3905                    c != 0x0085 && c != 0x2028 && c != 0x2029)                    (md->bsr_anycrlf ||
3906                       (c != 0x000b && c != 0x000c &&
3907                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3908                  break;                  break;
3909                eptr += len;                eptr += len;
3910                }                }
# Line 3990  for (;;) Line 4057  for (;;)
4057            }            }
4058          }          }
4059        else        else
4060  #endif  #endif  /* SUPPORT_UTF8 */
4061    
4062        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4063          {          {
# Line 4027  for (;;) Line 4094  for (;;)
4094                }                }
4095              else              else
4096                {                {
4097                if (c != 0x000a && c != 0x000b && c != 0x000c && c != 0x0085)                if (c != 0x000a &&
4098                      (md->bsr_anycrlf ||
4099                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4100                  break;                  break;
4101                eptr++;                eptr++;
4102                }                }
# Line 4181  switch (frame->Xwhere) Line 4250  switch (frame->Xwhere)
4250    LBL(17) LBL(18) LBL(19) LBL(20) LBL(21) LBL(22) LBL(23) LBL(24)    LBL(17) LBL(18) LBL(19) LBL(20) LBL(21) LBL(22) LBL(23) LBL(24)
4251    LBL(25) LBL(26) LBL(27) LBL(28) LBL(29) LBL(30) LBL(31) LBL(32)    LBL(25) LBL(26) LBL(27) LBL(28) LBL(29) LBL(30) LBL(31) LBL(32)
4252    LBL(33) LBL(34) LBL(35) LBL(36) LBL(37) LBL(38) LBL(39) LBL(40)    LBL(33) LBL(34) LBL(35) LBL(36) LBL(37) LBL(38) LBL(39) LBL(40)
4253    LBL(41) LBL(42) LBL(43) LBL(44) LBL(45) LBL(46) LBL(47)    LBL(41) LBL(42) LBL(43) LBL(44) LBL(45) LBL(46) LBL(47) LBL(48)
4254      LBL(49) LBL(50) LBL(51) LBL(52) LBL(53) LBL(54)
4255    default:    default:
4256    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4257    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
# Line 4298  const uschar *start_bits = NULL; Line 4368  const uschar *start_bits = NULL;
4368  USPTR start_match = (USPTR)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
4369  USPTR end_subject;  USPTR end_subject;
4370  USPTR req_byte_ptr = start_match - 1;  USPTR req_byte_ptr = start_match - 1;
 eptrblock eptrchain[EPTR_WORK_SIZE];  
4371    
4372  pcre_study_data internal_study;  pcre_study_data internal_study;
4373  const pcre_study_data *study;  const pcre_study_data *study;
# Line 4361  if (re->magic_number != MAGIC_NUMBER) Line 4430  if (re->magic_number != MAGIC_NUMBER)
4430  /* Set up other data */  /* Set up other data */
4431    
4432  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4433  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4434  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4435    
4436  /* The code starts after the real_pcre block and the capture name table. */  /* The code starts after the real_pcre block and the capture name table. */
# Line 4384  md->partial = (options & PCRE_PARTIAL) ! Line 4453  md->partial = (options & PCRE_PARTIAL) !
4453  md->hitend = FALSE;  md->hitend = FALSE;
4454    
4455  md->recursive = NULL;                   /* No recursion at top level */  md->recursive = NULL;                   /* No recursion at top level */
 md->eptrchain = eptrchain;              /* Make workspace generally available */  
4456    
4457  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
4458  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
4459    
4460    /* Handle different \R options. */
4461    
4462    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
4463      {
4464      case 0:
4465      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4466        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4467      else
4468    #ifdef BSR_ANYCRLF
4469      md->bsr_anycrlf = TRUE;
4470    #else
4471      md->bsr_anycrlf = FALSE;
4472    #endif
4473      break;
4474    
4475      case PCRE_BSR_ANYCRLF:
4476      md->bsr_anycrlf = TRUE;
4477      break;
4478    
4479      case PCRE_BSR_UNICODE:
4480      md->bsr_anycrlf = FALSE;
4481      break;
4482    
4483      default: return PCRE_ERROR_BADNEWLINE;
4484      }
4485    
4486  /* Handle different types of newline. The three bits give eight cases. If  /* Handle different types of newline. The three bits give eight cases. If
4487  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. */
4488    
4489  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : (pcre_uint32)options) &  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4490         PCRE_NEWLINE_BITS)          (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4491    {    {
4492    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
4493    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
# Line 4432  else Line 4526  else
4526  /* Partial matching is supported only for a restricted set of regexes at the  /* Partial matching is supported only for a restricted set of regexes at the
4527  moment. */  moment. */
4528    
4529  if (md->partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4530    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4531    
4532  /* 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
# Line 4509  studied, there may be a bitmap of possib Line 4603  studied, there may be a bitmap of possib
4603    
4604  if (!anchored)  if (!anchored)
4605    {    {
4606    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4607      {      {
4608      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4609      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
# Line 4524  if (!anchored) Line 4618  if (!anchored)
4618  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4619  character" set. */  character" set. */
4620    
4621  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4622    {    {
4623    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4624    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
# Line 4540  the loop runs just once. */ Line 4634  the loop runs just once. */
4634  for(;;)  for(;;)
4635    {    {
4636    USPTR save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4637      USPTR new_start_match;
4638    
4639    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4640    
# Line 4680  for(;;) Line 4775  for(;;)
4775    
4776    /* OK, we can now run the match. */    /* OK, we can now run the match. */
4777    
4778    md->start_match_ptr = start_match;      /* Insurance */    md->start_match_ptr = start_match;
4779    md->match_call_count = 0;    md->match_call_count = 0;
4780    md->eptrn = 0;                          /* Next free eptrchain slot */    rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
4781    rc = match(start_match, md->start_code, start_match, 2, md,  
4782      ims, NULL, 0, 0);    switch(rc)
4783        {
4784        /* NOMATCH and PRUNE advance by one character. THEN at this level acts
4785        exactly like PRUNE. */
4786    
4787        case MATCH_NOMATCH:
4788        case MATCH_PRUNE:
4789        case MATCH_THEN:
4790        new_start_match = start_match + 1;
4791    #ifdef SUPPORT_UTF8
4792        if (utf8)
4793          while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)
4794            new_start_match++;
4795    #endif
4796        break;
4797    
4798        /* SKIP passes back the next starting point explicitly. */
4799    
4800        case MATCH_SKIP:
4801        new_start_match = md->start_match_ptr;
4802        break;
4803    
4804        /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
4805    
4806        case MATCH_COMMIT:
4807        rc = MATCH_NOMATCH;
4808        goto ENDLOOP;
4809    
4810        /* Any other return is some kind of error. */
4811    
4812        default:
4813        goto ENDLOOP;
4814        }
4815    
4816    /* Any return other than MATCH_NOMATCH breaks the loop. */    /* Control reaches here for the various types of "no match at this point"
4817      result. Reset the code to MATCH_NOMATCH for subsequent checking. */
4818    
4819    if (rc != MATCH_NOMATCH) break;    rc = MATCH_NOMATCH;
4820    
4821    /* If PCRE_FIRSTLINE is set, the match must happen before or at the first    /* If PCRE_FIRSTLINE is set, the match must happen before or at the first
4822    newline in the subject (though it may continue over the newline). Therefore,    newline in the subject (though it may continue over the newline). Therefore,
# Line 4696  for(;;) Line 4824  for(;;)
4824    
4825    if (firstline && IS_NEWLINE(start_match)) break;    if (firstline && IS_NEWLINE(start_match)) break;
4826    
4827    /* Advance the match position by one character. */    /* Advance to new matching position */
4828    
4829    start_match++;    start_match = new_start_match;
 #ifdef SUPPORT_UTF8  
   if (utf8)  
     while(start_match < end_subject && (*start_match & 0xc0) == 0x80)  
       start_match++;  
 #endif  
4830    
4831    /* Break the loop if the pattern is anchored or if we have passed the end of    /* Break the loop if the pattern is anchored or if we have passed the end of
4832    the subject. */    the subject. */
4833    
4834    if (anchored || start_match > end_subject) break;    if (anchored || start_match > end_subject) break;
4835    
4836    /* If we have just passed a CR and the newline option is CRLF or ANY or    /* If we have just passed a CR and we are now at a LF, and the pattern does
4837    ANYCRLF, and we are now at a LF, advance the match position by one more    not contain any explicit matches for \r or \n, and the newline option is CRLF
4838    character. */    or ANY or ANYCRLF, advance the match position by one more character. */
4839    
4840    if (start_match[-1] == '\r' &&    if (start_match[-1] == '\r' &&
4841         (md->nltype == NLTYPE_ANY ||        start_match < end_subject &&
4842          md->nltype == NLTYPE_ANYCRLF ||        *start_match == '\n' &&
4843          md->nllen == 2) &&        (re->flags & PCRE_HASCRORLF) == 0 &&
4844         start_match < end_subject &&          (md->nltype == NLTYPE_ANY ||
4845         *start_match == '\n')           md->nltype == NLTYPE_ANYCRLF ||
4846             md->nllen == 2))
4847      start_match++;      start_match++;
4848    
4849    }   /* End of for(;;) "bumpalong" loop */    }   /* End of for(;;) "bumpalong" loop */
# Line 4729  for(;;) Line 4853  for(;;)
4853  /* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping  /* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping
4854  conditions is true:  conditions is true:
4855    
4856  (1) The pattern is anchored;  (1) The pattern is anchored or the match was failed by (*COMMIT);
4857    
4858  (2) We are past the end of the subject;  (2) We are past the end of the subject;
4859    
# Line 4744  processing, copy those that we can. In t Line 4868  processing, copy those that we can. In t
4868  certain parts of the pattern were not used, even though there are more  certain parts of the pattern were not used, even though there are more
4869  capturing parentheses than vector slots. */  capturing parentheses than vector slots. */
4870    
4871    ENDLOOP:
4872    
4873  if (rc == MATCH_MATCH)  if (rc == MATCH_MATCH)
4874    {    {
4875    if (using_temporary_offsets)    if (using_temporary_offsets)
Legend:
Removed from v.184  
changed lines
  Added in v.231
webmaster@exim.org
 ViewVC Help
Powered by ViewVC 1.1.12