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

Diff of /code/trunk/pcre_compile.c

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

revision 305 by ph10, Sun Jan 20 20:07:32 2008 UTC revision 335 by ph10, Sat Apr 12 14:36:14 2008 UTC
# Line 158  static const char verbnames[] = Line 158  static const char verbnames[] =
158    "SKIP\0"    "SKIP\0"
159    "THEN";    "THEN";
160    
161  static verbitem verbs[] = {  static const verbitem verbs[] = {
162    { 6, OP_ACCEPT },    { 6, OP_ACCEPT },
163    { 6, OP_COMMIT },    { 6, OP_COMMIT },
164    { 1, OP_FAIL },    { 1, OP_FAIL },
# Line 168  static verbitem verbs[] = { Line 168  static verbitem verbs[] = {
168    { 4, OP_THEN  }    { 4, OP_THEN  }
169  };  };
170    
171  static int verbcount = sizeof(verbs)/sizeof(verbitem);  static const int verbcount = sizeof(verbs)/sizeof(verbitem);
172    
173    
174  /* Tables of names of POSIX character classes and their lengths. The names are  /* Tables of names of POSIX character classes and their lengths. The names are
# Line 295  static const char error_texts[] = Line 295  static const char error_texts[] =
295    /* 55 */    /* 55 */
296    "repeating a DEFINE group is not allowed\0"    "repeating a DEFINE group is not allowed\0"
297    "inconsistent NEWLINE options\0"    "inconsistent NEWLINE options\0"
298    "\\g is not followed by a braced name or an optionally braced non-zero number\0"    "\\g is not followed by a braced, angle-bracketed, or quoted name/number or by a plain number\0"
299    "(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number\0"    "a numbered reference must not be zero\0"
300    "(*VERB) with an argument is not supported\0"    "(*VERB) with an argument is not supported\0"
301    /* 60 */    /* 60 */
302    "(*VERB) not recognized\0"    "(*VERB) not recognized\0"
# Line 531  else Line 531  else
531      *errorcodeptr = ERR37;      *errorcodeptr = ERR37;
532      break;      break;
533    
534      /* \g must be followed by a number, either plain or braced. If positive, it      /* \g must be followed by one of a number of specific things:
535      is an absolute backreference. If negative, it is a relative backreference.  
536      This is a Perl 5.10 feature. Perl 5.10 also supports \g{name} as a      (1) A number, either plain or braced. If positive, it is an absolute
537      reference to a named group. This is part of Perl's movement towards a      backreference. If negative, it is a relative backreference. This is a Perl
538      unified syntax for back references. As this is synonymous with \k{name}, we      5.10 feature.
539      fudge it up by pretending it really was \k. */  
540        (2) Perl 5.10 also supports \g{name} as a reference to a named group. This
541        is part of Perl's movement towards a unified syntax for back references. As
542        this is synonymous with \k{name}, we fudge it up by pretending it really
543        was \k.
544    
545        (3) For Oniguruma compatibility we also support \g followed by a name or a
546        number either in angle brackets or in single quotes. However, these are
547        (possibly recursive) subroutine calls, _not_ backreferences. Just return
548        the -ESC_g code (cf \k). */
549    
550      case 'g':      case 'g':
551        if (ptr[1] == '<' || ptr[1] == '\'')
552          {
553          c = -ESC_g;
554          break;
555          }
556    
557        /* Handle the Perl-compatible cases */
558    
559      if (ptr[1] == '{')      if (ptr[1] == '{')
560        {        {
561        const uschar *p;        const uschar *p;
# Line 565  else Line 582  else
582      while ((digitab[ptr[1]] & ctype_digit) != 0)      while ((digitab[ptr[1]] & ctype_digit) != 0)
583        c = c * 10 + *(++ptr) - '0';        c = c * 10 + *(++ptr) - '0';
584    
585      if (c < 0)      if (c < 0)   /* Integer overflow */
586        {        {
587        *errorcodeptr = ERR61;        *errorcodeptr = ERR61;
588        break;        break;
589        }        }
590    
591      if (c == 0 || (braced && *(++ptr) != '}'))      if (braced && *(++ptr) != '}')
592        {        {
593        *errorcodeptr = ERR57;        *errorcodeptr = ERR57;
594        break;        break;
595        }        }
596    
597        if (c == 0)
598          {
599          *errorcodeptr = ERR58;
600          break;
601          }
602    
603      if (negated)      if (negated)
604        {        {
# Line 611  else Line 634  else
634        c -= '0';        c -= '0';
635        while ((digitab[ptr[1]] & ctype_digit) != 0)        while ((digitab[ptr[1]] & ctype_digit) != 0)
636          c = c * 10 + *(++ptr) - '0';          c = c * 10 + *(++ptr) - '0';
637        if (c < 0)        if (c < 0)    /* Integer overflow */
638          {          {
639          *errorcodeptr = ERR61;          *errorcodeptr = ERR61;
640          break;          break;
# Line 1544  for (code = first_significant_code(code Line 1567  for (code = first_significant_code(code
1567    
1568    /* Groups with zero repeats can of course be empty; skip them. */    /* Groups with zero repeats can of course be empty; skip them. */
1569    
1570    if (c == OP_BRAZERO || c == OP_BRAMINZERO)    if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO)
1571      {      {
1572      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1573      do code += GET(code, 1); while (*code == OP_ALT);      do code += GET(code, 1); while (*code == OP_ALT);
# Line 1824  return -1; Line 1847  return -1;
1847  that is referenced. This means that groups can be replicated for fixed  that is referenced. This means that groups can be replicated for fixed
1848  repetition simply by copying (because the recursion is allowed to refer to  repetition simply by copying (because the recursion is allowed to refer to
1849  earlier groups that are outside the current group). However, when a group is  earlier groups that are outside the current group). However, when a group is
1850  optional (i.e. the minimum quantifier is zero), OP_BRAZERO is inserted before  optional (i.e. the minimum quantifier is zero), OP_BRAZERO or OP_SKIPZERO is
1851  it, after it has been compiled. This means that any OP_RECURSE items within it  inserted before it, after it has been compiled. This means that any OP_RECURSE
1852  that refer to the group itself or any contained groups have to have their  items within it that refer to the group itself or any contained groups have to
1853  offsets adjusted. That one of the jobs of this function. Before it is called,  have their offsets adjusted. That one of the jobs of this function. Before it
1854  the partially compiled regex must be temporarily terminated with OP_END.  is called, the partially compiled regex must be temporarily terminated with
1855    OP_END.
1856    
1857  This function has been extended with the possibility of forward references for  This function has been extended with the possibility of forward references for
1858  recursions and subroutine calls. It must also check the list of such references  recursions and subroutine calls. It must also check the list of such references
# Line 1859  while ((ptr = (uschar *)find_recurse(ptr Line 1883  while ((ptr = (uschar *)find_recurse(ptr
1883    
1884    /* See if this recursion is on the forward reference list. If so, adjust the    /* See if this recursion is on the forward reference list. If so, adjust the
1885    reference. */    reference. */
1886    
1887    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
1888      {      {
1889      offset = GET(hc, 0);      offset = GET(hc, 0);
# Line 2113  if (next >= 0) switch(op_code) Line 2137  if (next >= 0) switch(op_code)
2137    /* For OP_NOT, "item" must be a single-byte character. */    /* For OP_NOT, "item" must be a single-byte character. */
2138    
2139    case OP_NOT:    case OP_NOT:
   if (next < 0) return FALSE;  /* Not a character */  
2140    if (item == next) return TRUE;    if (item == next) return TRUE;
2141    if ((options & PCRE_CASELESS) == 0) return FALSE;    if ((options & PCRE_CASELESS) == 0) return FALSE;
2142  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 2436  for (;; ptr++) Line 2459  for (;; ptr++)
2459    /* Get next byte in the pattern */    /* Get next byte in the pattern */
2460    
2461    c = *ptr;    c = *ptr;
2462    
2463    /* If we are in the pre-compile phase, accumulate the length used for the    /* If we are in the pre-compile phase, accumulate the length used for the
2464    previous cycle of this loop. */    previous cycle of this loop. */
2465    
# Line 2688  for (;; ptr++) Line 2711  for (;; ptr++)
2711  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2712      class_utf8 = FALSE;                       /* No chars >= 256 */      class_utf8 = FALSE;                       /* No chars >= 256 */
2713      class_utf8data = code + LINK_SIZE + 2;    /* For UTF-8 items */      class_utf8data = code + LINK_SIZE + 2;    /* For UTF-8 items */
2714      class_utf8data_base = class_utf8data;     /* For resetting in pass 1 */      class_utf8data_base = class_utf8data;     /* For resetting in pass 1 */
2715  #endif  #endif
2716    
2717      /* Process characters until ] is reached. By writing this as a "do" it      /* Process characters until ] is reached. By writing this as a "do" it
# Line 2704  for (;; ptr++) Line 2727  for (;; ptr++)
2727          {                           /* Braces are required because the */          {                           /* Braces are required because the */
2728          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
2729          }          }
2730    
2731        /* In the pre-compile phase, accumulate the length of any UTF-8 extra        /* In the pre-compile phase, accumulate the length of any UTF-8 extra
2732        data and reset the pointer. This is so that very large classes that        data and reset the pointer. This is so that very large classes that
2733        contain a zillion UTF-8 characters no longer overwrite the work space        contain a zillion UTF-8 characters no longer overwrite the work space
2734        (which is on the stack). */        (which is on the stack). */
2735    
2736        if (lengthptr != NULL)        if (lengthptr != NULL)
2737          {          {
2738          *lengthptr += class_utf8data - class_utf8data_base;          *lengthptr += class_utf8data - class_utf8data_base;
2739          class_utf8data = class_utf8data_base;          class_utf8data = class_utf8data_base;
2740          }          }
2741    
2742  #endif  #endif
2743    
2744        /* Inside \Q...\E everything is literal except \E */        /* Inside \Q...\E everything is literal except \E */
# Line 3820  we set the flag only if there is a liter Line 3843  we set the flag only if there is a liter
3843    
3844        if (repeat_min == 0)        if (repeat_min == 0)
3845          {          {
3846          /* If the maximum is also zero, we just omit the group from the output          /* If the maximum is also zero, we used to just omit the group from the
3847          altogether. */          output altogether, like this:
3848    
3849          if (repeat_max == 0)          ** if (repeat_max == 0)
3850            {          **   {
3851            code = previous;          **   code = previous;
3852            goto END_REPEAT;          **   goto END_REPEAT;
3853            }          **   }
3854    
3855          /* If the maximum is 1 or unlimited, we just have to stick in the          However, that fails when a group is referenced as a subroutine from
3856          BRAZERO and do no more at this point. However, we do need to adjust          elsewhere in the pattern, so now we stick in OP_SKIPZERO in front of it
3857          any OP_RECURSE calls inside the group that refer to the group itself or          so that it is skipped on execution. As we don't have a list of which
3858          any internal or forward referenced group, because the offset is from          groups are referenced, we cannot do this selectively.
3859          the start of the whole regex. Temporarily terminate the pattern while  
3860          doing this. */          If the maximum is 1 or unlimited, we just have to stick in the BRAZERO
3861            and do no more at this point. However, we do need to adjust any
3862            OP_RECURSE calls inside the group that refer to the group itself or any
3863            internal or forward referenced group, because the offset is from the
3864            start of the whole regex. Temporarily terminate the pattern while doing
3865            this. */
3866    
3867          if (repeat_max <= 1)          if (repeat_max <= 1)    /* Covers 0, 1, and unlimited */
3868            {            {
3869            *code = OP_END;            *code = OP_END;
3870            adjust_recurse(previous, 1, utf8, cd, save_hwm);            adjust_recurse(previous, 1, utf8, cd, save_hwm);
3871            memmove(previous+1, previous, len);            memmove(previous+1, previous, len);
3872            code++;            code++;
3873              if (repeat_max == 0)
3874                {
3875                *previous++ = OP_SKIPZERO;
3876                goto END_REPEAT;
3877                }
3878            *previous++ = OP_BRAZERO + repeat_type;            *previous++ = OP_BRAZERO + repeat_type;
3879            }            }
3880    
# Line 4116  we set the flag only if there is a liter Line 4149  we set the flag only if there is a liter
4149      bravalue = OP_CBRA;      bravalue = OP_CBRA;
4150      save_hwm = cd->hwm;      save_hwm = cd->hwm;
4151      reset_bracount = FALSE;      reset_bracount = FALSE;
4152    
4153      /* First deal with various "verbs" that can be introduced by '*'. */      /* First deal with various "verbs" that can be introduced by '*'. */
4154    
4155      if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0)      if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0)
# Line 4568  we set the flag only if there is a liter Line 4601  we set the flag only if there is a liter
4601          references (?P=name) and recursion (?P>name), as well as falling          references (?P=name) and recursion (?P>name), as well as falling
4602          through from the Perl recursion syntax (?&name). We also come here from          through from the Perl recursion syntax (?&name). We also come here from
4603          the Perl \k<name> or \k'name' back reference syntax and the \k{name}          the Perl \k<name> or \k'name' back reference syntax and the \k{name}
4604          .NET syntax. */          .NET syntax, and the Oniguruma \g<...> and \g'...' subroutine syntax. */
4605    
4606          NAMED_REF_OR_RECURSE:          NAMED_REF_OR_RECURSE:
4607          name = ++ptr;          name = ++ptr;
# Line 4646  we set the flag only if there is a liter Line 4679  we set the flag only if there is a liter
4679          case '5': case '6': case '7': case '8': case '9':   /* subroutine */          case '5': case '6': case '7': case '8': case '9':   /* subroutine */
4680            {            {
4681            const uschar *called;            const uschar *called;
4682              terminator = ')';
4683    
4684              /* Come here from the \g<...> and \g'...' code (Oniguruma
4685              compatibility). However, the syntax has been checked to ensure that
4686              the ... are a (signed) number, so that neither ERR63 nor ERR29 will
4687              be called on this path, nor with the jump to OTHER_CHAR_AFTER_QUERY
4688              ever be taken. */
4689    
4690              HANDLE_NUMERICAL_RECURSION:
4691    
4692            if ((refsign = *ptr) == '+')            if ((refsign = *ptr) == '+')
4693              {              {
# Line 4667  we set the flag only if there is a liter Line 4709  we set the flag only if there is a liter
4709            while((digitab[*ptr] & ctype_digit) != 0)            while((digitab[*ptr] & ctype_digit) != 0)
4710              recno = recno * 10 + *ptr++ - '0';              recno = recno * 10 + *ptr++ - '0';
4711    
4712            if (*ptr != ')')            if (*ptr != terminator)
4713              {              {
4714              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
4715              goto FAILED;              goto FAILED;
# Line 5063  we set the flag only if there is a liter Line 5105  we set the flag only if there is a liter
5105      back references and those types that consume a character may be repeated.      back references and those types that consume a character may be repeated.
5106      We can test for values between ESC_b and ESC_Z for the latter; this may      We can test for values between ESC_b and ESC_Z for the latter; this may
5107      have to change if any new ones are ever created. */      have to change if any new ones are ever created. */
5108    
5109      case '\\':      case '\\':
5110      tempptr = ptr;      tempptr = ptr;
5111      c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE);      c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE);
# Line 5090  we set the flag only if there is a liter Line 5132  we set the flag only if there is a liter
5132    
5133        zerofirstbyte = firstbyte;        zerofirstbyte = firstbyte;
5134        zeroreqbyte = reqbyte;        zeroreqbyte = reqbyte;
5135    
5136          /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'
5137          is a subroutine call by number (Oniguruma syntax). In fact, the value
5138          -ESC_g is returned only for these cases. So we don't need to check for <
5139          or ' if the value is -ESC_g. For the Perl syntax \g{n} the value is
5140          -ESC_REF+n, and for the Perl syntax \g{name} the result is -ESC_k (as
5141          that is a synonym for a named back reference). */
5142    
5143          if (-c == ESC_g)
5144            {
5145            const uschar *p;
5146            save_hwm = cd->hwm;   /* Normally this is set when '(' is read */
5147            terminator = (*(++ptr) == '<')? '>' : '\'';
5148    
5149            /* These two statements stop the compiler for warning about possibly
5150            unset variables caused by the jump to HANDLE_NUMERICAL_RECURSION. In
5151            fact, because we actually check for a number below, the paths that
5152            would actually be in error are never taken. */
5153    
5154            skipbytes = 0;
5155            reset_bracount = FALSE;
5156    
5157            /* Test for a name */
5158    
5159            if (ptr[1] != '+' && ptr[1] != '-')
5160              {
5161              BOOL isnumber = TRUE;
5162              for (p = ptr + 1; *p != 0 && *p != terminator; p++)
5163                {
5164                if ((cd->ctypes[*p] & ctype_digit) == 0) isnumber = FALSE;
5165                if ((cd->ctypes[*p] & ctype_word) == 0) break;
5166                }
5167              if (*p != terminator)
5168                {
5169                *errorcodeptr = ERR57;
5170                break;
5171                }
5172              if (isnumber)
5173                {
5174                ptr++;
5175                goto HANDLE_NUMERICAL_RECURSION;
5176                }
5177              is_recurse = TRUE;
5178              goto NAMED_REF_OR_RECURSE;
5179              }
5180    
5181            /* Test a signed number in angle brackets or quotes. */
5182    
5183            p = ptr + 2;
5184            while ((digitab[*p] & ctype_digit) != 0) p++;
5185            if (*p != terminator)
5186              {
5187              *errorcodeptr = ERR57;
5188              break;
5189              }
5190            ptr++;
5191            goto HANDLE_NUMERICAL_RECURSION;
5192            }
5193    
5194        /* \k<name> or \k'name' is a back reference by name (Perl syntax).        /* \k<name> or \k'name' is a back reference by name (Perl syntax).
5195        We also support \k{name} (.NET syntax) */        We also support \k{name} (.NET syntax) */
# Line 6109  while (errorcode == 0 && cd->hwm > cwork Line 6209  while (errorcode == 0 && cd->hwm > cwork
6209    if (groupptr == NULL) errorcode = ERR53;    if (groupptr == NULL) errorcode = ERR53;
6210      else PUT(((uschar *)codestart), offset, groupptr - codestart);      else PUT(((uschar *)codestart), offset, groupptr - codestart);
6211    }    }
6212    
6213  /* Give an error if there's back reference to a non-existent capturing  /* Give an error if there's back reference to a non-existent capturing
6214  subpattern. */  subpattern. */
6215    

Legend:
Removed from v.305  
changed lines
  Added in v.335

  ViewVC Help
Powered by ViewVC 1.1.5