/[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 349 by ph10, Wed Jul 2 18:42:11 2008 UTC revision 409 by ph10, Sat Mar 28 17:10:56 2009 UTC
# Line 6  Line 6 
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.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2008 University of Cambridge             Copyright (c) 1997-2009 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 158  printf("\n"); Line 158  printf("\n");
158    
159  if (length > md->end_subject - eptr) return FALSE;  if (length > md->end_subject - eptr) return FALSE;
160    
161  /* Separate the caselesss case for speed */  /* Separate the caseless case for speed. In UTF-8 mode we can only do this
162    properly if Unicode properties are supported. Otherwise, we can check only
163    ASCII characters. */
164    
165  if ((ims & PCRE_CASELESS) != 0)  if ((ims & PCRE_CASELESS) != 0)
166    {    {
167    #ifdef SUPPORT_UTF8
168    #ifdef SUPPORT_UCP
169      if (md->utf8)
170        {
171        USPTR endptr = eptr + length;
172        while (eptr < endptr)
173          {
174          int c, d;
175          GETCHARINC(c, eptr);
176          GETCHARINC(d, p);
177          if (c != d && c != UCD_OTHERCASE(d)) return FALSE;
178          }
179        }
180      else
181    #endif
182    #endif
183    
184      /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
185      is no UCP support. */
186    
187    while (length-- > 0)    while (length-- > 0)
188      if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE;      { if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; }
189    }    }
190    
191    /* In the caseful case, we can just compare the bytes, whether or not we
192    are in UTF-8 mode. */
193    
194  else  else
195    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }
196    
# Line 296  typedef struct heapframe { Line 322  typedef struct heapframe {
322    
323    /* Function arguments that may change */    /* Function arguments that may change */
324    
325    const uschar *Xeptr;    USPTR Xeptr;
326    const uschar *Xecode;    const uschar *Xecode;
327    const uschar *Xmstart;    USPTR Xmstart;
328    int Xoffset_top;    int Xoffset_top;
329    long int Xims;    long int Xims;
330    eptrblock *Xeptrb;    eptrblock *Xeptrb;
# Line 307  typedef struct heapframe { Line 333  typedef struct heapframe {
333    
334    /* Function local variables */    /* Function local variables */
335    
336    const uschar *Xcallpat;    USPTR Xcallpat;
337    const uschar *Xcharptr;  #ifdef SUPPORT_UTF8
338    const uschar *Xdata;    USPTR Xcharptr;
339    const uschar *Xnext;  #endif
340    const uschar *Xpp;    USPTR Xdata;
341    const uschar *Xprev;    USPTR Xnext;
342    const uschar *Xsaved_eptr;    USPTR Xpp;
343      USPTR Xprev;
344      USPTR Xsaved_eptr;
345    
346    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
347    
# Line 334  typedef struct heapframe { Line 362  typedef struct heapframe {
362    uschar Xocchars[8];    uschar Xocchars[8];
363  #endif  #endif
364    
365      int Xcodelink;
366    int Xctype;    int Xctype;
367    unsigned int Xfc;    unsigned int Xfc;
368    int Xfi;    int Xfi;
# Line 399  Returns:       MATCH_MATCH if matched Line 428  Returns:       MATCH_MATCH if matched
428  */  */
429    
430  static int  static int
431  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, const uschar *mstart,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,
432    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,
433    int flags, unsigned int rdepth)    int flags, unsigned int rdepth)
434  {  {
# Line 413  register unsigned int c;   /* Character Line 442  register unsigned int c;   /* Character
442  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
443    
444  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
445    int condcode;
446    
447  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
448  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame" which is obtained from
# Line 455  HEAP_RECURSE: Line 485  HEAP_RECURSE:
485  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
486  #endif  #endif
487  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
488    #define codelink           frame->Xcodelink
489  #define data               frame->Xdata  #define data               frame->Xdata
490  #define next               frame->Xnext  #define next               frame->Xnext
491  #define pp                 frame->Xpp  #define pp                 frame->Xpp
# Line 535  int oclength; Line 566  int oclength;
566  uschar occhars[8];  uschar occhars[8];
567  #endif  #endif
568    
569    int codelink;
570  int ctype;  int ctype;
571  int length;  int length;
572  int max;  int max;
# Line 761  for (;;) Line 793  for (;;)
793    
794      case OP_COND:      case OP_COND:
795      case OP_SCOND:      case OP_SCOND:
796      if (ecode[LINK_SIZE+1] == OP_RREF)         /* Recursion test */      codelink= GET(ecode, 1);
797    
798        /* Because of the way auto-callout works during compile, a callout item is
799        inserted between OP_COND and an assertion condition. */
800    
801        if (ecode[LINK_SIZE+1] == OP_CALLOUT)
802          {
803          if (pcre_callout != NULL)
804            {
805            pcre_callout_block cb;
806            cb.version          = 1;   /* Version 1 of the callout block */
807            cb.callout_number   = ecode[LINK_SIZE+2];
808            cb.offset_vector    = md->offset_vector;
809            cb.subject          = (PCRE_SPTR)md->start_subject;
810            cb.subject_length   = md->end_subject - md->start_subject;
811            cb.start_match      = mstart - md->start_subject;
812            cb.current_position = eptr - md->start_subject;
813            cb.pattern_position = GET(ecode, LINK_SIZE + 3);
814            cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
815            cb.capture_top      = offset_top/2;
816            cb.capture_last     = md->capture_last;
817            cb.callout_data     = md->callout_data;
818            if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);
819            if (rrc < 0) RRETURN(rrc);
820            }
821          ecode += _pcre_OP_lengths[OP_CALLOUT];
822          }
823    
824        condcode = ecode[LINK_SIZE+1];
825    
826        /* Now see what the actual condition is */
827    
828        if (condcode == OP_RREF)         /* Recursion test */
829        {        {
830        offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/        offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/
831        condition = md->recursive != NULL &&        condition = md->recursive != NULL &&
# Line 769  for (;;) Line 833  for (;;)
833        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 3 : GET(ecode, 1);
834        }        }
835    
836      else if (ecode[LINK_SIZE+1] == OP_CREF)    /* Group used test */      else if (condcode == OP_CREF)    /* Group used test */
837        {        {
838        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
839        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
840        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 3 : GET(ecode, 1);
841        }        }
842    
843      else if (ecode[LINK_SIZE+1] == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
844        {        {
845        condition = FALSE;        condition = FALSE;
846        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 803  for (;;) Line 867  for (;;)
867        else        else
868          {          {
869          condition = FALSE;          condition = FALSE;
870          ecode += GET(ecode, 1);          ecode += codelink;
871          }          }
872        }        }
873    
# Line 826  for (;;) Line 890  for (;;)
890          goto TAIL_RECURSE;          goto TAIL_RECURSE;
891          }          }
892        }        }
893      else                         /* Condition false & no 2nd alternative */      else                         /* Condition false & no alternative */
894        {        {
895        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
896        }        }
# Line 1049  for (;;) Line 1113  for (;;)
1113          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1114            {            {
1115            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1116              if (new_recursive.offset_save != stacksave)
1117                (pcre_free)(new_recursive.offset_save);
1118            RRETURN(rrc);            RRETURN(rrc);
1119            }            }
1120    
# Line 1395  for (;;) Line 1461  for (;;)
1461          {          {
1462          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
1463            {            {
1464            const uschar *lastptr = eptr - 1;            USPTR lastptr = eptr - 1;
1465            while((*lastptr & 0xc0) == 0x80) lastptr--;            while((*lastptr & 0xc0) == 0x80) lastptr--;
1466            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
1467            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;
# Line 1653  for (;;) Line 1719  for (;;)
1719      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1720      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1721        {        {
1722        const ucd_record * prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
1723    
1724        switch(ecode[1])        switch(ecode[1])
1725          {          {
# Line 1669  for (;;) Line 1735  for (;;)
1735           break;           break;
1736    
1737          case PT_GC:          case PT_GC:
1738          if ((ecode[2] != ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))
1739            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1740          break;          break;
1741    
# Line 2021  for (;;) Line 2087  for (;;)
2087    
2088    
2089      /* Match an extended character class. This opcode is encountered only      /* Match an extended character class. This opcode is encountered only
2090      in UTF-8 mode, because that's the only time it is compiled. */      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
2091        mode, because Unicode properties are supported in non-UTF-8 mode. */
2092    
2093  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2094      case OP_XCLASS:      case OP_XCLASS:
# Line 2063  for (;;) Line 2130  for (;;)
2130        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
2131          {          {
2132          if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);          if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2133          GETCHARINC(c, eptr);          GETCHARINCTEST(c, eptr);
2134          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);
2135          }          }
2136    
# Line 2082  for (;;) Line 2149  for (;;)
2149            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2150            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2151            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2152            GETCHARINC(c, eptr);            GETCHARINCTEST(c, eptr);
2153            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);
2154            }            }
2155          /* Control never gets here */          /* Control never gets here */
# Line 2097  for (;;) Line 2164  for (;;)
2164            {            {
2165            int len = 1;            int len = 1;
2166            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
2167            GETCHARLEN(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
2168            if (!_pcre_xclass(c, data)) break;            if (!_pcre_xclass(c, data)) break;
2169            eptr += len;            eptr += len;
2170            }            }
# Line 2583  for (;;) Line 2650  for (;;)
2650              {              {
2651              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2652              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2653                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2654              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2655              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
2656              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fc == d) RRETURN(MATCH_NOMATCH);
2657                RRETURN(MATCH_NOMATCH);  
2658              }              }
2659            }            }
2660          else          else
# Line 2692  for (;;) Line 2760  for (;;)
2760              {              {
2761              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2762              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2763                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2764              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2765              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fc == d) RRETURN(MATCH_NOMATCH);
               RRETURN(MATCH_NOMATCH);  
2766              }              }
2767            }            }
2768          else          else
# Line 4358  Returns:          > 0 => success; value Line 4426  Returns:          > 0 => success; value
4426                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4427  */  */
4428    
4429  PCRE_EXP_DEFN int  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
4430  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4431    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4432    int offsetcount)    int offsetcount)
# Line 4506  switch ((((options & PCRE_NEWLINE_BITS) Line 4574  switch ((((options & PCRE_NEWLINE_BITS)
4574          (pcre_uint32)options) & PCRE_NEWLINE_BITS)          (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4575    {    {
4576    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
4577    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = CHAR_CR; break;
4578    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = CHAR_NL; break;
4579    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
4580         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break;
4581    case PCRE_NEWLINE_ANY: newline = -1; break;    case PCRE_NEWLINE_ANY: newline = -1; break;
4582    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4583    default: return PCRE_ERROR_BADNEWLINE;    default: return PCRE_ERROR_BADNEWLINE;
# Line 4551  back the character offset. */ Line 4619  back the character offset. */
4619  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4620  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
4621    {    {
4622    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    if (_pcre_valid_utf8((USPTR)subject, length) >= 0)
4623      return PCRE_ERROR_BADUTF8;      return PCRE_ERROR_BADUTF8;
4624    if (start_offset > 0 && start_offset < length)    if (start_offset > 0 && start_offset < length)
4625      {      {
4626      int tb = ((uschar *)subject)[start_offset];      int tb = ((USPTR)subject)[start_offset];
4627      if (tb > 127)      if (tb > 127)
4628        {        {
4629        tb &= 0xc0;        tb &= 0xc0;
# Line 4661  for(;;) Line 4729  for(;;)
4729      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
4730      }      }
4731    
4732    /* 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
4733    start of the match is constrained to the first line of a multiline string.    line of a multiline string. That is, the match must be before or at the first
4734    That is, the match must be before or at the first newline. Implement this by    newline. Implement this by temporarily adjusting end_subject so that we stop
4735    temporarily adjusting end_subject so that we stop scanning at a newline. If    scanning at a newline. If the match fails at the newline, later code breaks
4736    the match fails at the newline, later code breaks this loop. */    this loop. */
4737    
4738    if (firstline)    if (firstline)
4739      {      {
4740      USPTR t = start_match;      USPTR t = start_match;
4741    #ifdef SUPPORT_UTF8
4742        if (utf8)
4743          {
4744          while (t < md->end_subject && !IS_NEWLINE(t))
4745            {
4746            t++;
4747            while (t < end_subject && (*t & 0xc0) == 0x80) t++;
4748            }
4749          }
4750        else
4751    #endif
4752      while (t < md->end_subject && !IS_NEWLINE(t)) t++;      while (t < md->end_subject && !IS_NEWLINE(t)) t++;
4753      end_subject = t;      end_subject = t;
4754      }      }
4755    
4756    /* Now test for a unique first byte */    /* There are some optimizations that avoid running the match if a known
4757      starting point is not found, or if a known later character is not present.
4758      However, there is an option that disables these, for testing and for ensuring
4759      that all callouts do actually occur. */
4760    
4761    if (first_byte >= 0)    if ((options & PCRE_NO_START_OPTIMIZE) == 0)
4762      {      {
4763      if (first_byte_caseless)      /* Advance to a unique first byte if there is one. */
4764        while (start_match < end_subject &&  
4765               md->lcc[*start_match] != first_byte)      if (first_byte >= 0)
4766          { NEXTCHAR(start_match); }        {
4767      else        if (first_byte_caseless)
4768        while (start_match < end_subject && *start_match != first_byte)          while (start_match < end_subject && md->lcc[*start_match] != first_byte)
4769          { NEXTCHAR(start_match); }            start_match++;
4770      }        else
4771            while (start_match < end_subject && *start_match != first_byte)
4772              start_match++;
4773          }
4774    
4775    /* Or to just after a linebreak for a multiline match if possible */      /* Or to just after a linebreak for a multiline match */
4776    
4777    else if (startline)      else if (startline)
     {  
     if (start_match > md->start_subject + start_offset)  
4778        {        {
4779        while (start_match <= end_subject && !WAS_NEWLINE(start_match))        if (start_match > md->start_subject + start_offset)
4780          { NEXTCHAR(start_match); }          {
4781    #ifdef SUPPORT_UTF8
4782            if (utf8)
4783              {
4784              while (start_match < end_subject && !WAS_NEWLINE(start_match))
4785                {
4786                start_match++;
4787                while(start_match < end_subject && (*start_match & 0xc0) == 0x80)
4788                  start_match++;
4789                }
4790              }
4791            else
4792    #endif
4793            while (start_match < end_subject && !WAS_NEWLINE(start_match))
4794              start_match++;
4795    
4796            /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4797            and we are now at a LF, advance the match position by one more character.
4798            */
4799    
4800        /* If we have just passed a CR and the newline option is ANY or ANYCRLF,          if (start_match[-1] == CHAR_CR &&
4801        and we are now at a LF, advance the match position by one more character.               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
4802        */               start_match < end_subject &&
4803                 *start_match == CHAR_NL)
4804        if (start_match[-1] == '\r' &&            start_match++;
4805             (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&          }
            start_match < end_subject &&  
            *start_match == '\n')  
         start_match++;  
4806        }        }
     }  
4807    
4808    /* Or to a non-unique first char after study */      /* Or to a non-unique first byte after study */
4809    
4810    else if (start_bits != NULL)      else if (start_bits != NULL)
     {  
     while (start_match < end_subject)  
4811        {        {
4812        register unsigned int c = *start_match;        while (start_match < end_subject)
4813        if ((start_bits[c/8] & (1 << (c&7))) == 0)          {
4814          { NEXTCHAR(start_match); }          register unsigned int c = *start_match;
4815        else break;          if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++;
4816              else break;
4817            }
4818        }        }
4819      }      }   /* Starting optimizations */
4820    
4821    /* Restore fudged end_subject */    /* Restore fudged end_subject */
4822    
# Line 4731  for(;;) Line 4828  for(;;)
4828    printf("\n");    printf("\n");
4829  #endif  #endif
4830    
4831    /* If req_byte is set, we know that that character must appear in the subject    /* If req_byte is set, we know that that character must appear in the
4832    for the match to succeed. If the first character is set, req_byte must be    subject for the match to succeed. If the first character is set, req_byte
4833    later in the subject; otherwise the test starts at the match point. This    must be later in the subject; otherwise the test starts at the match point.
4834    optimization can save a huge amount of backtracking in patterns with nested    This optimization can save a huge amount of backtracking in patterns with
4835    unlimited repeats that aren't going to match. Writing separate code for    nested unlimited repeats that aren't going to match. Writing separate code
4836    cased/caseless versions makes it go faster, as does using an autoincrement    for cased/caseless versions makes it go faster, as does using an
4837    and backing off on a match.    autoincrement and backing off on a match.
4838    
4839    HOWEVER: when the subject string is very, very long, searching to its end can    HOWEVER: when the subject string is very, very long, searching to its end
4840    take a long time, and give bad performance on quite ordinary patterns. This    can take a long time, and give bad performance on quite ordinary patterns.
4841    showed up when somebody was matching something like /^\d+C/ on a 32-megabyte    This showed up when somebody was matching something like /^\d+C/ on a
4842    string... so we don't do this when the string is sufficiently long.    32-megabyte string... so we don't do this when the string is sufficiently
4843      long.
4844    
4845    ALSO: this processing is disabled when partial matching is requested.    ALSO: this processing is disabled when partial matching is requested, or if
4846    */    disabling is explicitly requested. */
4847    
4848    if (req_byte >= 0 &&    if ((options & PCRE_NO_START_OPTIMIZE) == 0 &&
4849          req_byte >= 0 &&
4850        end_subject - start_match < REQ_BYTE_MAX &&        end_subject - start_match < REQ_BYTE_MAX &&
4851        !md->partial)        !md->partial)
4852      {      {
# Line 4855  for(;;) Line 4954  for(;;)
4954    not contain any explicit matches for \r or \n, and the newline option is CRLF    not contain any explicit matches for \r or \n, and the newline option is CRLF
4955    or ANY or ANYCRLF, advance the match position by one more character. */    or ANY or ANYCRLF, advance the match position by one more character. */
4956    
4957    if (start_match[-1] == '\r' &&    if (start_match[-1] == CHAR_CR &&
4958        start_match < end_subject &&        start_match < end_subject &&
4959        *start_match == '\n' &&        *start_match == CHAR_NL &&
4960        (re->flags & PCRE_HASCRORLF) == 0 &&        (re->flags & PCRE_HASCRORLF) == 0 &&
4961          (md->nltype == NLTYPE_ANY ||          (md->nltype == NLTYPE_ANY ||
4962           md->nltype == NLTYPE_ANYCRLF ||           md->nltype == NLTYPE_ANYCRLF ||
Legend:
Removed from v.349  
changed lines
  Added in v.409
  ViewVC Help
Powered by ViewVC 1.1.5