/[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 1074 by chpe, Tue Oct 16 15:54:47 2012 UTC revision 1219 by ph10, Sun Nov 11 18:04:37 2012 UTC
# Line 84  static int Line 84  static int
84      const pcre_uint32 *, unsigned int);      const pcre_uint32 *, unsigned int);
85    
86  static BOOL  static BOOL
87    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL,    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int,
88      int, int, int *, int *, branch_chain *, compile_data *, int *);      pcre_uint32 *, pcre_int32 *, pcre_uint32 *, pcre_int32 *, branch_chain *,
89        compile_data *, int *);
90    
91    
92    
# Line 121  overrun before it actually does run off Line 122  overrun before it actually does run off
122    
123  /* Private flags added to firstchar and reqchar. */  /* Private flags added to firstchar and reqchar. */
124    
125  #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */  #define REQ_CASELESS    (1 << 0)        /* Indicates caselessness */
126  #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */  #define REQ_VARY        (1 << 1)        /* Reqchar followed non-literal item */
127  #define REQ_MASK       (REQ_CASELESS | REQ_VARY)  /* Negative values for the firstchar and reqchar flags */
128    #define REQ_UNSET       (-2)
129    #define REQ_NONE        (-1)
130    
131  /* Repeated character flags. */  /* Repeated character flags. */
132    
# Line 645  static const pcre_uint8 ebcdic_chartab[] Line 648  static const pcre_uint8 ebcdic_chartab[]
648    
649    
650    
   
651  /*************************************************  /*************************************************
652  *            Find an error text                  *  *            Find an error text                  *
653  *************************************************/  *************************************************/
# Line 665  find_error_text(int n) Line 667  find_error_text(int n)
667  const char *s = error_texts;  const char *s = error_texts;
668  for (; n > 0; n--)  for (; n > 0; n--)
669    {    {
670    while (*s++ != 0) {};    while (*s++ != CHAR_NULL) {};
671    if (*s == 0) return "Error text not found (please report)";    if (*s == CHAR_NULL) return "Error text not found (please report)";
672    }    }
673  return s;  return s;
674  }  }
# Line 786  ptr--;                            /* Set Line 788  ptr--;                            /* Set
788    
789  /* If backslash is at the end of the pattern, it's an error. */  /* If backslash is at the end of the pattern, it's an error. */
790    
791  if (c == 0) *errorcodeptr = ERR1;  if (c == CHAR_NULL) *errorcodeptr = ERR1;
792    
793  /* Non-alphanumerics are literals. For digits or letters, do an initial lookup  /* Non-alphanumerics are literals. For digits or letters, do an initial lookup
794  in a table. A non-zero result is something that can be returned immediately.  in a table. A non-zero result is something that can be returned immediately.
# Line 896  else Line 898  else
898      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
899        {        {
900        const pcre_uchar *p;        const pcre_uchar *p;
901        for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++)        for (p = ptr+2; *p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET; p++)
902          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;
903        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)        if (*p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET)
904          {          {
905          escape = ESC_k;          escape = ESC_k;
906          break;          break;
# Line 1132  else Line 1134  else
1134    
1135      case CHAR_c:      case CHAR_c:
1136      c = *(++ptr);      c = *(++ptr);
1137      if (c == 0)      if (c == CHAR_NULL)
1138        {        {
1139        *errorcodeptr = ERR2;        *errorcodeptr = ERR2;
1140        break;        break;
# Line 1201  escape sequence. Line 1203  escape sequence.
1203  Argument:  Argument:
1204    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
1205    negptr         points to a boolean that is set TRUE for negation else FALSE    negptr         points to a boolean that is set TRUE for negation else FALSE
1206    dptr           points to an int that is set to the detailed property value    ptypeptr       points to an unsigned int that is set to the type value
1207      pdataptr       points to an unsigned int that is set to the detailed property value
1208    errorcodeptr   points to the error code variable    errorcodeptr   points to the error code variable
1209    
1210  Returns:         type value from ucp_type_table, or -1 for an invalid type  Returns:         TRUE if the type value was found, or FALSE for an invalid type
1211  */  */
1212    
1213  static int  static BOOL
1214  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr)  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, unsigned int *ptypeptr,
1215      unsigned int *pdataptr, int *errorcodeptr)
1216  {  {
1217  pcre_uchar c;  pcre_uchar c;
1218  int i, bot, top;  int i, bot, top;
# Line 1216  const pcre_uchar *ptr = *ptrptr; Line 1220  const pcre_uchar *ptr = *ptrptr;
1220  pcre_uchar name[32];  pcre_uchar name[32];
1221    
1222  c = *(++ptr);  c = *(++ptr);
1223  if (c == 0) goto ERROR_RETURN;  if (c == CHAR_NULL) goto ERROR_RETURN;
1224    
1225  *negptr = FALSE;  *negptr = FALSE;
1226    
# Line 1233  if (c == CHAR_LEFT_CURLY_BRACKET) Line 1237  if (c == CHAR_LEFT_CURLY_BRACKET)
1237    for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)    for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)
1238      {      {
1239      c = *(++ptr);      c = *(++ptr);
1240      if (c == 0) goto ERROR_RETURN;      if (c == CHAR_NULL) goto ERROR_RETURN;
1241      if (c == CHAR_RIGHT_CURLY_BRACKET) break;      if (c == CHAR_RIGHT_CURLY_BRACKET) break;
1242      name[i] = c;      name[i] = c;
1243      }      }
# Line 1263  while (bot < top) Line 1267  while (bot < top)
1267    r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);    r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);
1268    if (r == 0)    if (r == 0)
1269      {      {
1270      *dptr = PRIV(utt)[i].value;      *ptypeptr = PRIV(utt)[i].type;
1271      return PRIV(utt)[i].type;      *pdataptr = PRIV(utt)[i].value;
1272        return TRUE;
1273      }      }
1274    if (r > 0) bot = i + 1; else top = i;    if (r > 0) bot = i + 1; else top = i;
1275    }    }
1276    
1277  *errorcodeptr = ERR47;  *errorcodeptr = ERR47;
1278  *ptrptr = ptr;  *ptrptr = ptr;
1279  return -1;  return FALSE;
1280    
1281  ERROR_RETURN:  ERROR_RETURN:
1282  *errorcodeptr = ERR46;  *errorcodeptr = ERR46;
1283  *ptrptr = ptr;  *ptrptr = ptr;
1284  return -1;  return FALSE;
1285  }  }
1286  #endif  #endif
1287    
# Line 1427  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1432  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1432    
1433    else if (ptr[2] == CHAR_NUMBER_SIGN)    else if (ptr[2] == CHAR_NUMBER_SIGN)
1434      {      {
1435      for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break;      for (ptr += 3; *ptr != CHAR_NULL; ptr++)
1436          if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
1437      goto FAIL_EXIT;      goto FAIL_EXIT;
1438      }      }
1439    
# Line 1440  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1446  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1446      ptr += 2;      ptr += 2;
1447      if (ptr[1] != CHAR_QUESTION_MARK)      if (ptr[1] != CHAR_QUESTION_MARK)
1448        {        {
1449        while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;        while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
1450        if (*ptr != 0) ptr++;        if (*ptr != CHAR_NULL) ptr++;
1451        }        }
1452      }      }
1453    
# Line 1484  for (; ptr < cd->end_pattern; ptr++) Line 1490  for (; ptr < cd->end_pattern; ptr++)
1490    
1491    if (*ptr == CHAR_BACKSLASH)    if (*ptr == CHAR_BACKSLASH)
1492      {      {
1493      if (*(++ptr) == 0) goto FAIL_EXIT;      if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
1494      if (*ptr == CHAR_Q) for (;;)      if (*ptr == CHAR_Q) for (;;)
1495        {        {
1496        while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};        while (*(++ptr) != CHAR_NULL && *ptr != CHAR_BACKSLASH) {};
1497        if (*ptr == 0) goto FAIL_EXIT;        if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1498        if (*(++ptr) == CHAR_E) break;        if (*(++ptr) == CHAR_E) break;
1499        }        }
1500      continue;      continue;
# Line 1532  for (; ptr < cd->end_pattern; ptr++) Line 1538  for (; ptr < cd->end_pattern; ptr++)
1538    
1539      while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)      while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)
1540        {        {
1541        if (*ptr == 0) return -1;        if (*ptr == CHAR_NULL) return -1;
1542        if (*ptr == CHAR_BACKSLASH)        if (*ptr == CHAR_BACKSLASH)
1543          {          {
1544          if (*(++ptr) == 0) goto FAIL_EXIT;          if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
1545          if (*ptr == CHAR_Q) for (;;)          if (*ptr == CHAR_Q) for (;;)
1546            {            {
1547            while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};            while (*(++ptr) != CHAR_NULL && *ptr != CHAR_BACKSLASH) {};
1548            if (*ptr == 0) goto FAIL_EXIT;            if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1549            if (*(++ptr) == CHAR_E) break;            if (*(++ptr) == CHAR_E) break;
1550            }            }
1551          continue;          continue;
# Line 1553  for (; ptr < cd->end_pattern; ptr++) Line 1559  for (; ptr < cd->end_pattern; ptr++)
1559    if (xmode && *ptr == CHAR_NUMBER_SIGN)    if (xmode && *ptr == CHAR_NUMBER_SIGN)
1560      {      {
1561      ptr++;      ptr++;
1562      while (*ptr != 0)      while (*ptr != CHAR_NULL)
1563        {        {
1564        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
1565        ptr++;        ptr++;
# Line 1561  for (; ptr < cd->end_pattern; ptr++) Line 1567  for (; ptr < cd->end_pattern; ptr++)
1567        if (utf) FORWARDCHAR(ptr);        if (utf) FORWARDCHAR(ptr);
1568  #endif  #endif
1569        }        }
1570      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1571      continue;      continue;
1572      }      }
1573    
# Line 1571  for (; ptr < cd->end_pattern; ptr++) Line 1577  for (; ptr < cd->end_pattern; ptr++)
1577      {      {
1578      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);
1579      if (rc > 0) return rc;      if (rc > 0) return rc;
1580      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1581      }      }
1582    
1583    else if (*ptr == CHAR_RIGHT_PARENTHESIS)    else if (*ptr == CHAR_RIGHT_PARENTHESIS)
# Line 1636  matching closing parens. That is why we Line 1642  matching closing parens. That is why we
1642  for (;;)  for (;;)
1643    {    {
1644    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);
1645    if (rc > 0 || *ptr++ == 0) break;    if (rc > 0 || *ptr++ == CHAR_NULL) break;
1646    }    }
1647    
1648  return rc;  return rc;
# Line 1852  for (;;) Line 1858  for (;;)
1858      case OP_NOTI:      case OP_NOTI:
1859      branchlength++;      branchlength++;
1860      cc += 2;      cc += 2;
1861  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1862      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1863  #endif  #endif
1864      break;      break;
# Line 1866  for (;;) Line 1872  for (;;)
1872      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1873      branchlength += (int)GET2(cc,1);      branchlength += (int)GET2(cc,1);
1874      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1875  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1876      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1877  #endif  #endif
1878      break;      break;
# Line 1909  for (;;) Line 1915  for (;;)
1915    
1916      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1917    
 #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32  
     case OP_XCLASS:  
     cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];  
     /* Fall through */  
 #endif  
   
1918      case OP_CLASS:      case OP_CLASS:
1919      case OP_NCLASS:      case OP_NCLASS:
1920    #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32
1921        case OP_XCLASS:
1922        /* The original code caused an unsigned overflow in 64 bit systems,
1923        so now we use a conditional statement. */
1924        if (op == OP_XCLASS)
1925          cc += GET(cc, 1);
1926        else
1927          cc += PRIV(OP_lengths)[OP_CLASS];
1928    #else
1929      cc += PRIV(OP_lengths)[OP_CLASS];      cc += PRIV(OP_lengths)[OP_CLASS];
1930    #endif
1931    
1932      switch (*cc)      switch (*cc)
1933        {        {
# Line 2708  check_posix_syntax(const pcre_uchar *ptr Line 2718  check_posix_syntax(const pcre_uchar *ptr
2718  {  {
2719  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */
2720  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
2721  for (++ptr; *ptr != 0; ptr++)  for (++ptr; *ptr != CHAR_NULL; ptr++)
2722    {    {
2723    if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)    if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
2724      ptr++;      ptr++;
# Line 2913  get_othercase_range(pcre_uint32 *cptr, p Line 2923  get_othercase_range(pcre_uint32 *cptr, p
2923    pcre_uint32 *odptr)    pcre_uint32 *odptr)
2924  {  {
2925  pcre_uint32 c, othercase, next;  pcre_uint32 c, othercase, next;
2926  int co;  unsigned int co;
2927    
2928  /* Find the first character that has an other case. If it has multiple other  /* Find the first character that has an other case. If it has multiple other
2929  cases, return its case offset value. */  cases, return its case offset value. */
# Line 2924  for (c = *cptr; c <= d; c++) Line 2934  for (c = *cptr; c <= d; c++)
2934      {      {
2935      *ocptr = c++;   /* Character that has the set */      *ocptr = c++;   /* Character that has the set */
2936      *cptr = c;      /* Rest of input range */      *cptr = c;      /* Rest of input range */
2937      return co;      return (int)co;
2938      }      }
2939    if ((othercase = UCD_OTHERCASE(c)) != c) break;    if ((othercase = UCD_OTHERCASE(c)) != c) break;
2940    }    }
# Line 2964  Returns:       TRUE if auto-possessifyin Line 2974  Returns:       TRUE if auto-possessifyin
2974  */  */
2975    
2976  static BOOL  static BOOL
2977  check_char_prop(pcre_uint32 c, int ptype, int pdata, BOOL negated)  check_char_prop(pcre_uint32 c, unsigned int ptype, unsigned int pdata, BOOL negated)
2978  {  {
2979  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2980  const pcre_uint32 *p;  const pcre_uint32 *p;
# Line 3015  switch(ptype) Line 3025  switch(ptype)
3025    p = PRIV(ucd_caseless_sets) + prop->caseset;    p = PRIV(ucd_caseless_sets) + prop->caseset;
3026    for (;;)    for (;;)
3027      {      {
3028      if ((unsigned int)c < *p) return !negated;      if (c < *p) return !negated;
3029      if ((unsigned int)c == *p++) return negated;      if (c == *p++) return negated;
3030      }      }
3031    break;  /* Control never reaches here */    break;  /* Control never reaches here */
3032  #endif  #endif
# Line 3053  check_auto_possessive(const pcre_uchar * Line 3063  check_auto_possessive(const pcre_uchar *
3063  pcre_uint32 c = NOTACHAR;  pcre_uint32 c = NOTACHAR;
3064  pcre_uint32 next;  pcre_uint32 next;
3065  int escape;  int escape;
3066  int op_code = *previous++;  pcre_uchar op_code = *previous++;
3067    
3068  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
3069    
# Line 3065  if ((options & PCRE_EXTENDED) != 0) Line 3075  if ((options & PCRE_EXTENDED) != 0)
3075      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
3076        {        {
3077        ptr++;        ptr++;
3078        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3079          {          {
3080          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
3081          ptr++;          ptr++;
# Line 3108  if ((options & PCRE_EXTENDED) != 0) Line 3118  if ((options & PCRE_EXTENDED) != 0)
3118      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
3119        {        {
3120        ptr++;        ptr++;
3121        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3122          {          {
3123          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
3124          ptr++;          ptr++;
# Line 3131  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3141  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3141    
3142  if (op_code == OP_CHAR || op_code == OP_CHARI ||  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3143      op_code == OP_NOT || op_code == OP_NOTI)      op_code == OP_NOT || op_code == OP_NOTI)
   //if (escape == 0) switch(op_code)  
3144    {    {
3145  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3146    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
# Line 3151  if (escape == 0) Line 3160  if (escape == 0)
3160  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3161    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3162      {      {
3163      int ocs = UCD_CASESET(next);      unsigned int ocs = UCD_CASESET(next);
3164      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);
3165      }      }
3166  #endif  #endif
# Line 3337  switch(op_code) Line 3346  switch(op_code)
3346      case ESC_p:      case ESC_p:
3347      case ESC_P:      case ESC_P:
3348        {        {
3349        int ptype, pdata, errorcodeptr;        unsigned int ptype = 0, pdata = 0;
3350          int errorcodeptr;
3351        BOOL negated;        BOOL negated;
3352    
3353        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
3354        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
3355        if (ptype < 0) return FALSE;          return FALSE;
3356        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
3357    
3358        /* If the property item is optional, we have to give up. (When generated        /* If the property item is optional, we have to give up. (When generated
# Line 3666  Arguments: Line 3676  Arguments:
3676    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3677    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3678    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3679    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr    place to put the first required character
3680    reqcharptr     set to the last literal character required, else < 0    firstcharflagsptr place to put the first character flags, or a negative number
3681      reqcharptr     place to put the last required character
3682      reqcharflagsptr place to put the last required character flags, or a negative number
3683    bcptr          points to current branch chain    bcptr          points to current branch chain
3684    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3685    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3680  Returns:         TRUE on success Line 3692  Returns:         TRUE on success
3692    
3693  static BOOL  static BOOL
3694  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3695    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,    const pcre_uchar **ptrptr, int *errorcodeptr,
3696    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,    pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
3697      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
3698      branch_chain *bcptr, int cond_depth,
3699    compile_data *cd, int *lengthptr)    compile_data *cd, int *lengthptr)
3700  {  {
3701  int repeat_type, op_type;  int repeat_type, op_type;
3702  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3703  int bravalue = 0;  int bravalue = 0;
3704  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3705  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
3706  pcre_int32 zeroreqchar, zerofirstchar;  pcre_int32 firstcharflags, reqcharflags;
3707    pcre_uint32 zeroreqchar, zerofirstchar;
3708    pcre_int32 zeroreqcharflags, zerofirstcharflags;
3709  pcre_int32 req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3710  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3711  int after_manual_callout = 0;  int after_manual_callout = 0;
# Line 3717  dynamically as we process the pattern. * Line 3733  dynamically as we process the pattern. *
3733  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3734  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3735  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3736    #ifndef COMPILE_PCRE32
3737  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3738    #endif
3739  #else  #else
3740  BOOL utf = FALSE;  BOOL utf = FALSE;
3741  #endif  #endif
# Line 3752  to take the zero repeat into account. Th Line 3770  to take the zero repeat into account. Th
3770  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3771  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3772    
3773  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
3774    firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;
3775    
3776  /* The variable req_caseopt contains either the REQ_CASELESS value  /* The variable req_caseopt contains either the REQ_CASELESS value
3777  or zero, according to the current setting of the caseless flag. The  or zero, according to the current setting of the caseless flag. The
# Line 3778  for (;; ptr++) Line 3797  for (;; ptr++)
3797    int recno;    int recno;
3798    int refsign;    int refsign;
3799    int skipbytes;    int skipbytes;
3800    int subreqchar;    pcre_uint32 subreqchar, subfirstchar;
3801    int subfirstchar;    pcre_int32 subreqcharflags, subfirstcharflags;
3802    int terminator;    int terminator;
3803    int mclength;    unsigned int mclength;
3804    int tempbracount;    unsigned int tempbracount;
3805    int ec; // FIXMEchpe pcre_uint32    pcre_uint32 ec;
3806    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3807    
3808    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 3793  for (;; ptr++) Line 3812  for (;; ptr++)
3812    /* If we are at the end of a nested substitution, revert to the outer level    /* If we are at the end of a nested substitution, revert to the outer level
3813    string. Nesting only happens one level deep. */    string. Nesting only happens one level deep. */
3814    
3815    if (c == 0 && nestptr != NULL)    if (c == CHAR_NULL && nestptr != NULL)
3816      {      {
3817      ptr = nestptr;      ptr = nestptr;
3818      nestptr = NULL;      nestptr = NULL;
# Line 3868  for (;; ptr++) Line 3887  for (;; ptr++)
3887    
3888    /* If in \Q...\E, check for the end; if not, we have a literal */    /* If in \Q...\E, check for the end; if not, we have a literal */
3889    
3890    if (inescq && c != 0)    if (inescq && c != CHAR_NULL)
3891      {      {
3892      if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)      if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)
3893        {        {
# Line 3916  for (;; ptr++) Line 3935  for (;; ptr++)
3935      if (c == CHAR_NUMBER_SIGN)      if (c == CHAR_NUMBER_SIGN)
3936        {        {
3937        ptr++;        ptr++;
3938        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3939          {          {
3940          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
3941          ptr++;          ptr++;
# Line 3924  for (;; ptr++) Line 3943  for (;; ptr++)
3943          if (utf) FORWARDCHAR(ptr);          if (utf) FORWARDCHAR(ptr);
3944  #endif  #endif
3945          }          }
3946        if (*ptr != 0) continue;        if (*ptr != CHAR_NULL) continue;
3947    
3948        /* Else fall through to handle end of string */        /* Else fall through to handle end of string */
3949        c = 0;        c = 0;
# Line 3946  for (;; ptr++) Line 3965  for (;; ptr++)
3965      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3966      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3967      *firstcharptr = firstchar;      *firstcharptr = firstchar;
3968        *firstcharflagsptr = firstcharflags;
3969      *reqcharptr = reqchar;      *reqcharptr = reqchar;
3970        *reqcharflagsptr = reqcharflags;
3971      *codeptr = code;      *codeptr = code;
3972      *ptrptr = ptr;      *ptrptr = ptr;
3973      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3970  for (;; ptr++) Line 3991  for (;; ptr++)
3991      previous = NULL;      previous = NULL;
3992      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3993        {        {
3994        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
3995        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3996        }        }
3997      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3985  for (;; ptr++) Line 4006  for (;; ptr++)
4006      repeats. The value of reqchar doesn't change either. */      repeats. The value of reqchar doesn't change either. */
4007    
4008      case CHAR_DOT:      case CHAR_DOT:
4009      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4010      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4011        zerofirstcharflags = firstcharflags;
4012      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4013        zeroreqcharflags = reqcharflags;
4014      previous = code;      previous = code;
4015      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
4016      break;      break;
# Line 4061  for (;; ptr++) Line 4084  for (;; ptr++)
4084          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
4085        {        {
4086        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
4087        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4088        zerofirstchar = firstchar;        zerofirstchar = firstchar;
4089          zerofirstcharflags = firstcharflags;
4090        break;        break;
4091        }        }
4092    
# Line 4097  for (;; ptr++) Line 4121  for (;; ptr++)
4121      means that an initial ] is taken as a data character. At the start of the      means that an initial ] is taken as a data character. At the start of the
4122      loop, c contains the first byte of the character. */      loop, c contains the first byte of the character. */
4123    
4124      if (c != 0) do      if (c != CHAR_NULL) do
4125        {        {
4126        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
4127    
4128  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
4129        if (utf && HAS_EXTRALEN(c))        if (utf && HAS_EXTRALEN(c))
4130          {                           /* Braces are required because the */          {                           /* Braces are required because the */
4131          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
# Line 4360  for (;; ptr++) Line 4384  for (;; ptr++)
4384              case ESC_P:              case ESC_P:
4385                {                {
4386                BOOL negated;                BOOL negated;
4387                int pdata;                unsigned int ptype = 0, pdata = 0;
4388                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
4389                if (ptype < 0) goto FAILED;                  goto FAILED;
4390                *class_uchardata++ = ((escape == ESC_p) != negated)?                *class_uchardata++ = ((escape == ESC_p) != negated)?
4391                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4392                *class_uchardata++ = ptype;                *class_uchardata++ = ptype;
# Line 4416  for (;; ptr++) Line 4440  for (;; ptr++)
4440    
4441        if (!inescq && ptr[1] == CHAR_MINUS)        if (!inescq && ptr[1] == CHAR_MINUS)
4442          {          {
4443          int d;          pcre_uint32 d;
4444          ptr += 2;          ptr += 2;
4445          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;
4446    
# Line 4435  for (;; ptr++) Line 4459  for (;; ptr++)
4459          /* Minus (hyphen) at the end of a class is treated as a literal, so put          /* Minus (hyphen) at the end of a class is treated as a literal, so put
4460          back the pointer and jump to handle the character that preceded it. */          back the pointer and jump to handle the character that preceded it. */
4461    
4462          if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))          if (*ptr == CHAR_NULL || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))
4463            {            {
4464            ptr = oldptr;            ptr = oldptr;
4465            goto CLASS_SINGLE_CHARACTER;            goto CLASS_SINGLE_CHARACTER;
# Line 4522  for (;; ptr++) Line 4546  for (;; ptr++)
4546          {          {
4547          ptr++;          ptr++;
4548          zeroreqchar = reqchar;          zeroreqchar = reqchar;
4549            zeroreqcharflags = reqcharflags;
4550    
4551          if (negate_class)          if (negate_class)
4552            {            {
4553  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
             // FIXMEchpe pcreuint32?  
4554            int d;            int d;
4555  #endif  #endif
4556            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;            if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4557            zerofirstchar = firstchar;            zerofirstchar = firstchar;
4558              zerofirstcharflags = firstcharflags;
4559    
4560            /* For caseless UTF-8 mode when UCP support is available, check            /* For caseless UTF-8 mode when UCP support is available, check
4561            whether this character has more than one other case. If so, generate            whether this character has more than one other case. If so, generate
# Line 4589  for (;; ptr++) Line 4614  for (;; ptr++)
4614      If we are at the end of an internal nested string, revert to the outer      If we are at the end of an internal nested string, revert to the outer
4615      string. */      string. */
4616    
4617      while (((c = *(++ptr)) != 0 ||      while (((c = *(++ptr)) != CHAR_NULL ||
4618             (nestptr != NULL &&             (nestptr != NULL &&
4619               (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != 0)) &&               (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != CHAR_NULL)) &&
4620             (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));             (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));
4621    
4622      /* Check for missing terminating ']' */      /* Check for missing terminating ']' */
4623    
4624      if (c == 0)      if (c == CHAR_NULL)
4625        {        {
4626        *errorcodeptr = ERR6;        *errorcodeptr = ERR6;
4627        goto FAILED;        goto FAILED;
# Line 4618  for (;; ptr++) Line 4643  for (;; ptr++)
4643      setting, whatever the repeat count. Any reqchar setting must remain      setting, whatever the repeat count. Any reqchar setting must remain
4644      unchanged after any kind of repeat. */      unchanged after any kind of repeat. */
4645    
4646      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4647      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4648        zerofirstcharflags = firstcharflags;
4649      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4650        zeroreqcharflags = reqcharflags;
4651    
4652      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4653      extended class, with its own opcode, unless there was a negated special      extended class, with its own opcode, unless there was a negated special
# Line 4715  for (;; ptr++) Line 4742  for (;; ptr++)
4742      if (repeat_min == 0)      if (repeat_min == 0)
4743        {        {
4744        firstchar = zerofirstchar;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4745          firstcharflags = zerofirstcharflags;
4746        reqchar = zeroreqchar;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4747          reqcharflags = zeroreqcharflags;
4748        }        }
4749    
4750      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4818  for (;; ptr++) Line 4847  for (;; ptr++)
4847          {          {
4848          c = code[-1];          c = code[-1];
4849          if (*previous <= OP_CHARI && repeat_min > 1)          if (*previous <= OP_CHARI && repeat_min > 1)
4850            reqchar = c | req_caseopt | cd->req_varyopt;            {
4851              reqchar = c;
4852              reqcharflags = req_caseopt | cd->req_varyopt;
4853              }
4854          }          }
4855    
4856        /* If the repetition is unlimited, it pays to see if the next thing on        /* If the repetition is unlimited, it pays to see if the next thing on
# Line 5200  for (;; ptr++) Line 5232  for (;; ptr++)
5232    
5233            else            else
5234              {              {
5235              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;              if (groupsetfirstchar && reqcharflags < 0)
5236                  {
5237                  reqchar = firstchar;
5238                  reqcharflags = firstcharflags;
5239                  }
5240    
5241              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5242                {                {
# Line 5479  for (;; ptr++) Line 5515  for (;; ptr++)
5515        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)
5516          {          {
5517          tempcode += PRIV(OP_lengths)[*tempcode];          tempcode += PRIV(OP_lengths)[*tempcode];
5518  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
5519          if (utf && HAS_EXTRALEN(tempcode[-1]))          if (utf && HAS_EXTRALEN(tempcode[-1]))
5520            tempcode += GET_EXTRALEN(tempcode[-1]);            tempcode += GET_EXTRALEN(tempcode[-1]);
5521  #endif  #endif
# Line 5575  for (;; ptr++) Line 5611  for (;; ptr++)
5611        if (*ptr == CHAR_COLON)        if (*ptr == CHAR_COLON)
5612          {          {
5613          arg = ++ptr;          arg = ++ptr;
5614          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5615          arglen = (int)(ptr - arg);          arglen = (int)(ptr - arg);
5616          if ((unsigned int)arglen > MAX_MARK)          if ((unsigned int)arglen > MAX_MARK)
5617            {            {
# Line 5620  for (;; ptr++) Line 5656  for (;; ptr++)
5656                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5657    
5658              /* Do not set firstchar after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5659              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
5660              }              }
5661    
5662            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5689  for (;; ptr++) Line 5725  for (;; ptr++)
5725          {          {
5726          case CHAR_NUMBER_SIGN:                 /* Comment; skip to ket */          case CHAR_NUMBER_SIGN:                 /* Comment; skip to ket */
5727          ptr++;          ptr++;
5728          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5729          if (*ptr == 0)          if (*ptr == CHAR_NULL)
5730            {            {
5731            *errorcodeptr = ERR18;            *errorcodeptr = ERR18;
5732            goto FAILED;            goto FAILED;
# Line 5766  for (;; ptr++) Line 5802  for (;; ptr++)
5802            }            }
5803          else          else
5804            {            {
5805            terminator = 0;            terminator = CHAR_NULL;
5806            if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);            if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);
5807            }            }
5808    
# Line 5786  for (;; ptr++) Line 5822  for (;; ptr++)
5822          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)
5823            {            {
5824            if (recno >= 0)            if (recno >= 0)
5825              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;              recno = (IS_DIGIT(*ptr))? recno * 10 + (int)(*ptr - CHAR_0) : -1;
5826            ptr++;            ptr++;
5827            }            }
5828          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5829    
5830          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) ||
5831              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
5832            {            {
5833            ptr--;      /* Error offset */            ptr--;      /* Error offset */
# Line 5856  for (;; ptr++) Line 5892  for (;; ptr++)
5892            code[1+LINK_SIZE]++;            code[1+LINK_SIZE]++;
5893            }            }
5894    
5895          /* If terminator == 0 it means that the name followed directly after          /* If terminator == CHAR_NULL it means that the name followed directly
5896          the opening parenthesis [e.g. (?(abc)...] and in this case there are          after the opening parenthesis [e.g. (?(abc)...] and in this case there
5897          some further alternatives to try. For the cases where terminator != 0          are some further alternatives to try. For the cases where terminator !=
5898          [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have          0 [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have
5899          now checked all the possibilities, so give an error. */          now checked all the possibilities, so give an error. */
5900    
5901          else if (terminator != 0)          else if (terminator != CHAR_NULL)
5902            {            {
5903            *errorcodeptr = ERR15;            *errorcodeptr = ERR15;
5904            goto FAILED;            goto FAILED;
# Line 6031  for (;; ptr++) Line 6067  for (;; ptr++)
6067    
6068            if (lengthptr != NULL)            if (lengthptr != NULL)
6069              {              {
6070              if (*ptr != terminator)              if (*ptr != (pcre_uchar)terminator)
6071                {                {
6072                *errorcodeptr = ERR42;                *errorcodeptr = ERR42;
6073                goto FAILED;                goto FAILED;
# Line 6173  for (;; ptr++) Line 6209  for (;; ptr++)
6209              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
6210              goto FAILED;              goto FAILED;
6211              }              }
6212            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6213              {              {
6214              *errorcodeptr = ERR42;              *errorcodeptr = ERR42;
6215              goto FAILED;              goto FAILED;
# Line 6279  for (;; ptr++) Line 6315  for (;; ptr++)
6315            while(IS_DIGIT(*ptr))            while(IS_DIGIT(*ptr))
6316              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6317    
6318            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6319              {              {
6320              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
6321              goto FAILED;              goto FAILED;
# Line 6383  for (;; ptr++) Line 6419  for (;; ptr++)
6419    
6420          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6421    
6422          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6423          continue;          continue;
6424    
6425    
# Line 6517  for (;; ptr++) Line 6553  for (;; ptr++)
6553           cond_depth +           cond_depth +
6554             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6555           &subfirstchar,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6556             &subfirstcharflags,
6557           &subreqchar,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6558             &subreqcharflags,
6559           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6560           cd,                              /* Tables block */           cd,                              /* Tables block */
6561           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6578  for (;; ptr++) Line 6616  for (;; ptr++)
6616            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6617            goto FAILED;            goto FAILED;
6618            }            }
6619          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;          if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
6620          }          }
6621        }        }
6622    
# Line 6627  for (;; ptr++) Line 6665  for (;; ptr++)
6665      back off. */      back off. */
6666    
6667      zeroreqchar = reqchar;      zeroreqchar = reqchar;
6668        zeroreqcharflags = reqcharflags;
6669      zerofirstchar = firstchar;      zerofirstchar = firstchar;
6670        zerofirstcharflags = firstcharflags;
6671      groupsetfirstchar = FALSE;      groupsetfirstchar = FALSE;
6672    
6673      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
# Line 6638  for (;; ptr++) Line 6678  for (;; ptr++)
6678        no firstchar, set "none" for the whole branch. In both cases, a zero        no firstchar, set "none" for the whole branch. In both cases, a zero
6679        repeat forces firstchar to "none". */        repeat forces firstchar to "none". */
6680    
6681        if (firstchar == REQ_UNSET)        if (firstcharflags == REQ_UNSET)
6682          {          {
6683          if (subfirstchar >= 0)          if (subfirstcharflags >= 0)
6684            {            {
6685            firstchar = subfirstchar;            firstchar = subfirstchar;
6686              firstcharflags = subfirstcharflags;
6687            groupsetfirstchar = TRUE;            groupsetfirstchar = TRUE;
6688            }            }
6689          else firstchar = REQ_NONE;          else firstcharflags = REQ_NONE;
6690          zerofirstchar = REQ_NONE;          zerofirstcharflags = REQ_NONE;
6691          }          }
6692    
6693        /* If firstchar was previously set, convert the subpattern's firstchar        /* If firstchar was previously set, convert the subpattern's firstchar
6694        into reqchar if there wasn't one, using the vary flag that was in        into reqchar if there wasn't one, using the vary flag that was in
6695        existence beforehand. */        existence beforehand. */
6696    
6697        else if (subfirstchar >= 0 && subreqchar < 0)        else if (subfirstcharflags >= 0 && subreqcharflags < 0)
6698          subreqchar = subfirstchar | tempreqvary;          {
6699            subreqchar = subfirstchar;
6700            subreqcharflags = subfirstcharflags | tempreqvary;
6701            }
6702    
6703        /* If the subpattern set a required byte (or set a first byte that isn't        /* If the subpattern set a required byte (or set a first byte that isn't
6704        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6705    
6706        if (subreqchar >= 0) reqchar = subreqchar;        if (subreqcharflags >= 0)
6707            {
6708            reqchar = subreqchar;
6709            reqcharflags = subreqcharflags;
6710            }
6711        }        }
6712    
6713      /* For a forward assertion, we take the reqchar, if set. This can be      /* For a forward assertion, we take the reqchar, if set. This can be
# Line 6670  for (;; ptr++) Line 6718  for (;; ptr++)
6718      of a firstchar. This is overcome by a scan at the end if there's no      of a firstchar. This is overcome by a scan at the end if there's no
6719      firstchar, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6720    
6721      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;      else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
6722          {
6723          reqchar = subreqchar;
6724          reqcharflags = subreqcharflags;
6725          }
6726      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6727    
6728    
# Line 6706  for (;; ptr++) Line 6758  for (;; ptr++)
6758        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6759        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6760    
6761        if (firstchar == REQ_UNSET && escape > ESC_b && escape < ESC_Z)        if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6762          firstchar = REQ_NONE;          firstcharflags = REQ_NONE;
6763    
6764        /* Set values to reset to if this is followed by a zero repeat. */        /* Set values to reset to if this is followed by a zero repeat. */
6765    
6766        zerofirstchar = firstchar;        zerofirstchar = firstchar;
6767          zerofirstcharflags = firstcharflags;
6768        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6769          zeroreqcharflags = reqcharflags;
6770    
6771        /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'        /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'
6772        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
# Line 6741  for (;; ptr++) Line 6795  for (;; ptr++)
6795          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
6796            {            {
6797            BOOL is_a_number = TRUE;            BOOL is_a_number = TRUE;
6798            for (p = ptr + 1; *p != 0 && *p != terminator; p++)            for (p = ptr + 1; *p != CHAR_NULL && *p != (pcre_uchar)terminator; p++)
6799              {              {
6800              if (!MAX_255(*p)) { is_a_number = FALSE; break; }              if (!MAX_255(*p)) { is_a_number = FALSE; break; }
6801              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
6802              if ((cd->ctypes[*p] & ctype_word) == 0) break;              if ((cd->ctypes[*p] & ctype_word) == 0) break;
6803              }              }
6804            if (*p != terminator)            if (*p != (pcre_uchar)terminator)
6805              {              {
6806              *errorcodeptr = ERR57;              *errorcodeptr = ERR57;
6807              break;              break;
# Line 6765  for (;; ptr++) Line 6819  for (;; ptr++)
6819    
6820          p = ptr + 2;          p = ptr + 2;
6821          while (IS_DIGIT(*p)) p++;          while (IS_DIGIT(*p)) p++;
6822          if (*p != terminator)          if (*p != (pcre_uchar)terminator)
6823            {            {
6824            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
6825            break;            break;
# Line 6802  for (;; ptr++) Line 6856  for (;; ptr++)
6856          recno = -escape;          recno = -escape;
6857    
6858          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6859          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6860          previous = code;          previous = code;
6861          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6862          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6829  for (;; ptr++) Line 6883  for (;; ptr++)
6883        else if (escape == ESC_P || escape == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6884          {          {
6885          BOOL negated;          BOOL negated;
6886          int pdata;          unsigned int ptype = 0, pdata = 0;
6887          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
6888          if (ptype < 0) goto FAILED;            goto FAILED;
6889          previous = code;          previous = code;
6890          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6891          *code++ = ptype;          *code++ = ptype;
# Line 6905  for (;; ptr++) Line 6959  for (;; ptr++)
6959      mclength = 1;      mclength = 1;
6960      mcbuffer[0] = c;      mcbuffer[0] = c;
6961    
6962  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
6963      if (utf && HAS_EXTRALEN(c))      if (utf && HAS_EXTRALEN(c))
6964        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));
6965  #endif  #endif
# Line 6929  for (;; ptr++) Line 6983  for (;; ptr++)
6983          *code++ = OP_PROP;          *code++ = OP_PROP;
6984          *code++ = PT_CLIST;          *code++ = PT_CLIST;
6985          *code++ = c;          *code++ = c;
6986          if (firstchar == REQ_UNSET) firstchar = zerofirstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = zerofirstcharflags = REQ_NONE;
6987          break;          break;
6988          }          }
6989        }        }
# Line 6950  for (;; ptr++) Line 7004  for (;; ptr++)
7004      Otherwise, leave the firstchar value alone, and don't change it on a zero      Otherwise, leave the firstchar value alone, and don't change it on a zero
7005      repeat. */      repeat. */
7006    
7007      if (firstchar == REQ_UNSET)      if (firstcharflags == REQ_UNSET)
7008        {        {
7009        zerofirstchar = REQ_NONE;        zerofirstcharflags = REQ_NONE;
7010        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7011          zeroreqcharflags = reqcharflags;
7012    
7013        /* If the character is more than one byte long, we can set firstchar        /* If the character is more than one byte long, we can set firstchar
7014        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
# Line 6961  for (;; ptr++) Line 7016  for (;; ptr++)
7016        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7017          {          {
7018          firstchar = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
7019          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;          firstchar = mcbuffer[0];
7020            firstcharflags = req_caseopt;
7021    
7022            if (mclength != 1)
7023              {
7024              reqchar = code[-1];
7025              reqcharflags = cd->req_varyopt;
7026              }
7027          }          }
7028        else firstchar = reqchar = REQ_NONE;        else firstcharflags = reqcharflags = REQ_NONE;
7029        }        }
7030    
7031      /* firstchar was previously set; we can set reqchar only if the length is      /* firstchar was previously set; we can set reqchar only if the length is
# Line 6972  for (;; ptr++) Line 7034  for (;; ptr++)
7034      else      else
7035        {        {
7036        zerofirstchar = firstchar;        zerofirstchar = firstchar;
7037          zerofirstcharflags = firstcharflags;
7038        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7039          zeroreqcharflags = reqcharflags;
7040        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7041          reqchar = code[-1] | req_caseopt | cd->req_varyopt;          {
7042            reqchar = code[-1];
7043            reqcharflags = req_caseopt | cd->req_varyopt;
7044            }
7045        }        }
7046    
7047      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6993  return FALSE; Line 7060  return FALSE;
7060    
7061    
7062    
   
7063  /*************************************************  /*************************************************
7064  *     Compile sequence of alternatives           *  *     Compile sequence of alternatives           *
7065  *************************************************/  *************************************************/
# Line 7014  Arguments: Line 7080  Arguments:
7080    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
7081    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
7082    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
7083    firstcharptr   place to put the first required character, or a negative number    firstcharptr    place to put the first required character
7084    reqcharptr     place to put the last required character, or a negative number    firstcharflagsptr place to put the first character flags, or a negative number
7085      reqcharptr     place to put the last required character
7086      reqcharflagsptr place to put the last required character flags, or a negative number
7087    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
7088    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
7089    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 7027  Returns:         TRUE on success Line 7095  Returns:         TRUE on success
7095  static BOOL  static BOOL
7096  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
7097    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
7098    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,    int cond_depth,
7099      pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
7100      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
7101    branch_chain *bcptr, compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
7102  {  {
7103  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
# Line 7037  pcre_uchar *start_bracket = code; Line 7107  pcre_uchar *start_bracket = code;
7107  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
7108  open_capitem capitem;  open_capitem capitem;
7109  int capnumber = 0;  int capnumber = 0;
7110  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7111  pcre_int32 branchfirstchar, branchreqchar;  pcre_int32 firstcharflags, reqcharflags;
7112    pcre_uint32 branchfirstchar, branchreqchar;
7113    pcre_int32 branchfirstcharflags, branchreqcharflags;
7114  int length;  int length;
7115  int orig_bracount;  unsigned int orig_bracount;
7116  int max_bracount;  unsigned int max_bracount;
7117  branch_chain bc;  branch_chain bc;
7118    
7119  bc.outer = bcptr;  bc.outer = bcptr;
7120  bc.current_branch = code;  bc.current_branch = code;
7121    
7122  firstchar = reqchar = REQ_UNSET;  firstchar = reqchar = 0;
7123    firstcharflags = reqcharflags = REQ_UNSET;
7124    
7125  /* Accumulate the length for use in the pre-compile phase. Start with the  /* Accumulate the length for use in the pre-compile phase. Start with the
7126  length of the BRA and KET and any extra bytes that are required at the  length of the BRA and KET and any extra bytes that are required at the
# Line 7107  for (;;) Line 7180  for (;;)
7180    into the length. */    into the length. */
7181    
7182    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
7183          &branchreqchar, &bc, cond_depth, cd,          &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
7184          (lengthptr == NULL)? NULL : &length))          cond_depth, cd, (lengthptr == NULL)? NULL : &length))
7185      {      {
7186      *ptrptr = ptr;      *ptrptr = ptr;
7187      return FALSE;      return FALSE;
# Line 7129  for (;;) Line 7202  for (;;)
7202      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
7203        {        {
7204        firstchar = branchfirstchar;        firstchar = branchfirstchar;
7205          firstcharflags = branchfirstcharflags;
7206        reqchar = branchreqchar;        reqchar = branchreqchar;
7207          reqcharflags = branchreqcharflags;
7208        }        }
7209    
7210      /* If this is not the first branch, the first char and reqchar have to      /* If this is not the first branch, the first char and reqchar have to
# Line 7143  for (;;) Line 7218  for (;;)
7218        we have to abandon the firstchar for the regex, but if there was        we have to abandon the firstchar for the regex, but if there was
7219        previously no reqchar, it takes on the value of the old firstchar. */        previously no reqchar, it takes on the value of the old firstchar. */
7220    
7221        if (firstchar >= 0 && firstchar != branchfirstchar)        if (firstcharflags >= 0 &&
7222              (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar))
7223          {          {
7224          if (reqchar < 0) reqchar = firstchar;          if (reqcharflags < 0)
7225          firstchar = REQ_NONE;            {
7226              reqchar = firstchar;
7227              reqcharflags = firstcharflags;
7228              }
7229            firstcharflags = REQ_NONE;
7230          }          }
7231    
7232        /* If we (now or from before) have no firstchar, a firstchar from the        /* If we (now or from before) have no firstchar, a firstchar from the
7233        branch becomes a reqchar if there isn't a branch reqchar. */        branch becomes a reqchar if there isn't a branch reqchar. */
7234    
7235        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)        if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
7236            branchreqchar = branchfirstchar;          {
7237            branchreqchar = branchfirstchar;
7238            branchreqcharflags = branchfirstcharflags;
7239            }
7240    
7241        /* Now ensure that the reqchars match */        /* Now ensure that the reqchars match */
7242    
7243        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))        if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) ||
7244          reqchar = REQ_NONE;            reqchar != branchreqchar)
7245        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */          reqcharflags = REQ_NONE;
7246          else
7247            {
7248            reqchar = branchreqchar;
7249            reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */
7250            }
7251        }        }
7252    
7253      /* If lookbehind, check that this branch matches a fixed-length string, and      /* If lookbehind, check that this branch matches a fixed-length string, and
# Line 7255  for (;;) Line 7343  for (;;)
7343      *codeptr = code;      *codeptr = code;
7344      *ptrptr = ptr;      *ptrptr = ptr;
7345      *firstcharptr = firstchar;      *firstcharptr = firstchar;
7346        *firstcharflagsptr = firstcharflags;
7347      *reqcharptr = reqchar;      *reqcharptr = reqchar;
7348        *reqcharflagsptr = reqcharflags;
7349      if (lengthptr != NULL)      if (lengthptr != NULL)
7350        {        {
7351        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7542  we return that char, otherwise -1. Line 7632  we return that char, otherwise -1.
7632    
7633  Arguments:  Arguments:
7634    code       points to start of expression (the bracket)    code       points to start of expression (the bracket)
7635      flags       points to the first char flags, or to REQ_NONE
7636    inassert   TRUE if in an assertion    inassert   TRUE if in an assertion
7637    
7638  Returns:     -1 or the fixed first char  Returns:     the fixed first char, or 0 with REQ_NONE in flags
7639  */  */
7640    
7641  static int  static pcre_uint32
7642  find_firstassertedchar(const pcre_uchar *code, BOOL inassert)  find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags,
7643      BOOL inassert)
7644  {  {
7645  register int c = -1;  register pcre_uint32 c = 0;
7646    int cflags = REQ_NONE;
7647    
7648    *flags = REQ_NONE;
7649  do {  do {
7650     int d;     pcre_uint32 d;
7651       int dflags;
7652     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7653               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7654     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7655       TRUE);       TRUE);
7656     register int op = *scode;     register pcre_uchar op = *scode;
7657    
7658     switch(op)     switch(op)
7659       {       {
7660       default:       default:
7661       return -1;       return 0;
7662    
7663       case OP_BRA:       case OP_BRA:
7664       case OP_BRAPOS:       case OP_BRAPOS:
# Line 7574  do { Line 7670  do {
7670       case OP_ONCE:       case OP_ONCE:
7671       case OP_ONCE_NC:       case OP_ONCE_NC:
7672       case OP_COND:       case OP_COND:
7673       if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)       d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
7674         return -1;       if (dflags < 0)
7675       if (c < 0) c = d; else if (c != d) return -1;         return 0;
7676         if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0;
7677       break;       break;
7678    
7679       case OP_EXACT:       case OP_EXACT:
# Line 7587  do { Line 7684  do {
7684       case OP_PLUS:       case OP_PLUS:
7685       case OP_MINPLUS:       case OP_MINPLUS:
7686       case OP_POSPLUS:       case OP_POSPLUS:
7687       if (!inassert) return -1;       if (!inassert) return 0;
7688       if (c < 0) c = scode[1];       if (cflags < 0) { c = scode[1]; cflags = 0; }
7689         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7690       break;       break;
7691    
7692       case OP_EXACTI:       case OP_EXACTI:
# Line 7600  do { Line 7697  do {
7697       case OP_PLUSI:       case OP_PLUSI:
7698       case OP_MINPLUSI:       case OP_MINPLUSI:
7699       case OP_POSPLUSI:       case OP_POSPLUSI:
7700       if (!inassert) return -1;       if (!inassert) return 0;
7701       if (c < 0) c = scode[1] | REQ_CASELESS;       if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
7702         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7703       break;       break;
7704       }       }
7705    
7706     code += GET(code, 1);     code += GET(code, 1);
7707     }     }
7708  while (*code == OP_ALT);  while (*code == OP_ALT);
7709    
7710    *flags = cflags;
7711  return c;  return c;
7712  }  }
7713    
# Line 7676  pcre32_compile2(PCRE_SPTR32 pattern, int Line 7775  pcre32_compile2(PCRE_SPTR32 pattern, int
7775  {  {
7776  REAL_PCRE *re;  REAL_PCRE *re;
7777  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7778  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7779    pcre_int32 firstcharflags, reqcharflags;
7780  int newline;  int newline;
7781  int errorcode = 0;  int errorcode = 0;
7782  int skipatstart = 0;  int skipatstart = 0;
# Line 7748  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7848  while (ptr[skipatstart] == CHAR_LEFT_PAR
7848    {    {
7849    int newnl = 0;    int newnl = 0;
7850    int newbsr = 0;    int newbsr = 0;
7851    
7852    /* For completeness and backward compatibility, (*UTFn) is supported in the
7853    relevant libraries, but (*UTF) is generic and always supported. Note that
7854    PCRE_UTF8 == PCRE_UTF16 == PCRE_UTF32. */
7855    
7856  #ifdef COMPILE_PCRE8  #ifdef COMPILE_PCRE8
7857    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 5) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF8_RIGHTPAR, 5) == 0)
7858      { skipatstart += 7; options |= PCRE_UTF8; continue; }      { skipatstart += 7; options |= PCRE_UTF8; continue; }
7859  #endif  #endif
7860  #ifdef COMPILE_PCRE16  #ifdef COMPILE_PCRE16
7861    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF16_RIGHTPAR, 6) == 0)
7862      { skipatstart += 8; options |= PCRE_UTF16; continue; }      { skipatstart += 8; options |= PCRE_UTF16; continue; }
7863  #endif  #endif
7864  #ifdef COMPILE_PCRE32  #ifdef COMPILE_PCRE32
7865    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF32_RIGHTPAR, 6) == 0)
7866      { skipatstart += 8; options |= PCRE_UTF32; continue; }      { skipatstart += 8; options |= PCRE_UTF32; continue; }
7867  #endif  #endif
7868    
7869      else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 4) == 0)
7870        { skipatstart += 6; options |= PCRE_UTF8; continue; }
7871    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)
7872      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
7873    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)
# Line 7926  ptr += skipatstart; Line 8033  ptr += skipatstart;
8033  code = cworkspace;  code = cworkspace;
8034  *code = OP_BRA;  *code = OP_BRA;
8035  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8036    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8037      cd, &length);
8038  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
8039    
8040  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 8004  ptr = (const pcre_uchar *)pattern + skip Line 8112  ptr = (const pcre_uchar *)pattern + skip
8112  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
8113  *code = OP_BRA;  *code = OP_BRA;
8114  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
8115    &firstchar, &reqchar, NULL, cd, NULL);    &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
8116  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
8117  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
8118  re->max_lookbehind = cd->max_lookbehind;  re->max_lookbehind = cd->max_lookbehind;
8119  re->flags = cd->external_flags | PCRE_MODE;  re->flags = cd->external_flags | PCRE_MODE;
8120    
8121  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept)
8122      {
8123      reqchar = 0;              /* Must disable after (*ACCEPT) */
8124      reqcharflags = REQ_NONE;
8125      }
8126    
8127  /* If not reached end of pattern on success, there's an excess bracket. */  /* If not reached end of pattern on success, there's an excess bracket. */
8128    
8129  if (errorcode == 0 && *ptr != 0) errorcode = ERR22;  if (errorcode == 0 && *ptr != CHAR_NULL) errorcode = ERR22;
8130    
8131  /* Fill in the terminating state and check for disastrous overflow, but  /* Fill in the terminating state and check for disastrous overflow, but
8132  if debugging, leave the test till after things are printed out. */  if debugging, leave the test till after things are printed out. */
# Line 8025  if debugging, leave the test till after Line 8137  if debugging, leave the test till after
8137  if (code - codestart > length) errorcode = ERR23;  if (code - codestart > length) errorcode = ERR23;
8138  #endif  #endif
8139    
8140    #ifdef SUPPORT_VALGRIND
8141    /* If the estimated length exceeds the really used length, mark the extra
8142    allocated memory as unadressable, so that any out-of-bound reads can be
8143    detected. */
8144    VALGRIND_MAKE_MEM_NOACCESS(code, (length - (code - codestart)) * sizeof(pcre_uchar));
8145    #endif
8146    
8147  /* Fill in any forward references that are required. There may be repeated  /* Fill in any forward references that are required. There may be repeated
8148  references; optimize for them, as searching a large regex takes time. */  references; optimize for them, as searching a large regex takes time. */
8149    
# Line 8060  if (errorcode == 0 && re->top_backref > Line 8179  if (errorcode == 0 && re->top_backref >
8179    
8180  /* If there were any lookbehind assertions that contained OP_RECURSE  /* If there were any lookbehind assertions that contained OP_RECURSE
8181  (recursions or subroutine calls), a flag is set for them to be checked here,  (recursions or subroutine calls), a flag is set for them to be checked here,
8182  because they may contain forward references. Actual recursions can't be fixed  because they may contain forward references. Actual recursions cannot be fixed
8183  length, but subroutine calls can. It is done like this so that those without  length, but subroutine calls can. It is done like this so that those without
8184  OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The  OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The
8185  exceptional ones forgo this. We scan the pattern to check that they are fixed  exceptional ones forgo this. We scan the pattern to check that they are fixed
# Line 8131  if ((re->options & PCRE_ANCHORED) == 0) Line 8250  if ((re->options & PCRE_ANCHORED) == 0)
8250    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
8251    else    else
8252      {      {
8253      if (firstchar < 0)      if (firstcharflags < 0)
8254        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
8255      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstcharflags >= 0)   /* Remove caseless flag for non-caseable chars */
8256        {        {
8257  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8258        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8259  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8260        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8261  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8262        re->first_char = firstchar & ~REQ_MASK;        re->first_char = firstchar;
8263  #endif  #endif
8264        if ((firstchar & REQ_CASELESS) != 0)        if ((firstcharflags & REQ_CASELESS) != 0)
8265          {          {
8266  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8267          /* We ignore non-ASCII first chars in 8 bit mode. */          /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8174  if ((re->options & PCRE_ANCHORED) == 0) Line 8293  if ((re->options & PCRE_ANCHORED) == 0)
8293  variable length item in the regex. Remove the caseless flag for non-caseable  variable length item in the regex. Remove the caseless flag for non-caseable
8294  bytes. */  bytes. */
8295    
8296  if (reqchar >= 0 &&  if (reqcharflags >= 0 &&
8297       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0))
8298    {    {
8299  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8300    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8301  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8302    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8303  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8304    re->req_char = reqchar & ~REQ_MASK;    re->req_char = reqchar;
8305  #endif  #endif
8306    if ((reqchar & REQ_CASELESS) != 0)    if ((reqcharflags & REQ_CASELESS) != 0)
8307      {      {
8308  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8309      /* We ignore non-ASCII first chars in 8 bit mode. */      /* We ignore non-ASCII first chars in 8 bit mode. */

Legend:
Removed from v.1074  
changed lines
  Added in v.1219

  ViewVC Help
Powered by ViewVC 1.1.5