/[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 1059 by chpe, Tue Oct 16 15:53:53 2012 UTC revision 1152 by chpe, Sun Oct 21 16:53:51 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 750  return (*p == CHAR_RIGHT_CURLY_BRACKET); Line 752  return (*p == CHAR_RIGHT_CURLY_BRACKET);
752    
753  /* This function is called when a \ has been encountered. It either returns a  /* This function is called when a \ has been encountered. It either returns a
754  positive value for a simple escape such as \n, or 0 for a data character  positive value for a simple escape such as \n, or 0 for a data character
755  which will be placed in chptr. A backreference to group  which will be placed in chptr. A backreference to group n is returned as
756  n is returned as ESC_REF + n; ESC_REF is the highest ESC_xxx macro. When  negative n. When UTF-8 is enabled, a positive value greater than 255 may
757  UTF-8 is enabled, a positive value greater than 255 may be returned in chptr.  be returned in chptr.
758  On entry,ptr is pointing at the \. On exit, it is on the final character of the  On entry,ptr is pointing at the \. On exit, it is on the final character of the
759  escape sequence.  escape sequence.
760    
# Line 766  Arguments: Line 768  Arguments:
768    
769  Returns:         zero => a data character  Returns:         zero => a data character
770                   positive => a special escape sequence                   positive => a special escape sequence
771                     negative => a back reference
772                   on error, errorcodeptr is set                   on error, errorcodeptr is set
773  */  */
774    
775  static int  static int
776  check_escape(const pcre_uchar **ptrptr, int *chptr, int *errorcodeptr,  check_escape(const pcre_uchar **ptrptr, pcre_uint32 *chptr, int *errorcodeptr,
777    int bracount, int options, BOOL isclass)    int bracount, int options, BOOL isclass)
778  {  {
779  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
780  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
781  const pcre_uchar *ptr = *ptrptr + 1;  const pcre_uchar *ptr = *ptrptr + 1;
782  pcre_int32 c;  pcre_uint32 c;
783  int escape = 0;  int escape = 0;
784  int i;  int i;
785    
# Line 794  Otherwise further processing may be requ Line 797  Otherwise further processing may be requ
797  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
798  /* Not alphanumeric */  /* Not alphanumeric */
799  else if (c < CHAR_0 || c > CHAR_z) {}  else if (c < CHAR_0 || c > CHAR_z) {}
800  else if ((i = escapes[c - CHAR_0]) != 0) { if (i > 0) c = i; else escape = -i; }  else if ((i = escapes[c - CHAR_0]) != 0) { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
801    
802  #else           /* EBCDIC coding */  #else           /* EBCDIC coding */
803  /* Not alphanumeric */  /* Not alphanumeric */
804  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}
805  else if ((i = escapes[c - 0x48]) != 0)  { if (i > 0) c = i; else escape = -i; }  else if ((i = escapes[c - 0x48]) != 0)  { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
806  #endif  #endif
807    
808  /* Escapes that need further processing, or are illegal. */  /* Escapes that need further processing, or are illegal. */
# Line 807  else if ((i = escapes[c - 0x48]) != 0) Line 810  else if ((i = escapes[c - 0x48]) != 0)
810  else  else
811    {    {
812    const pcre_uchar *oldptr;    const pcre_uchar *oldptr;
813    BOOL braced, negated;    BOOL braced, negated, overflow;
814      int s;
815    
816    switch (c)    switch (c)
817      {      {
# Line 832  else Line 836  else
836          c = 0;          c = 0;
837          for (i = 0; i < 4; ++i)          for (i = 0; i < 4; ++i)
838            {            {
839            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
840  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
841            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
842            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 914  else Line 918  else
918      else negated = FALSE;      else negated = FALSE;
919    
920      /* The integer range is limited by the machine's int representation. */      /* The integer range is limited by the machine's int representation. */
921      c = 0;      s = 0;
922        overflow = FALSE;
923      while (IS_DIGIT(ptr[1]))      while (IS_DIGIT(ptr[1]))
924        {        {
925        if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */        if (s > INT_MAX / 10 - 1) /* Integer overflow */
926          {          {
927          c = -1;          overflow = TRUE;
928          break;          break;
929          }          }
930        c = c * 10 + *(++ptr) - CHAR_0;        s = s * 10 + (int)(*(++ptr) - CHAR_0);
931        }        }
932      if (((unsigned int)c) > INT_MAX) /* Integer overflow */      if (overflow) /* Integer overflow */
933        {        {
934        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
935          ptr++;          ptr++;
# Line 938  else Line 943  else
943        break;        break;
944        }        }
945    
946      if (c == 0)      if (s == 0)
947        {        {
948        *errorcodeptr = ERR58;        *errorcodeptr = ERR58;
949        break;        break;
# Line 946  else Line 951  else
951    
952      if (negated)      if (negated)
953        {        {
954        if (c > bracount)        if (s > bracount)
955          {          {
956          *errorcodeptr = ERR15;          *errorcodeptr = ERR15;
957          break;          break;
958          }          }
959        c = bracount - (c - 1);        s = bracount - (s - 1);
960        }        }
961    
962      escape = ESC_REF + c;      escape = -s;
963      break;      break;
964    
965      /* The handling of escape sequences consisting of a string of digits      /* The handling of escape sequences consisting of a string of digits
# Line 976  else Line 981  else
981        {        {
982        oldptr = ptr;        oldptr = ptr;
983        /* The integer range is limited by the machine's int representation. */        /* The integer range is limited by the machine's int representation. */
984        c -= CHAR_0;        s = (int)(c -CHAR_0);
985          overflow = FALSE;
986        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
987          {          {
988          if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */          if (s > INT_MAX / 10 - 1) /* Integer overflow */
989            {            {
990            c = -1;            overflow = TRUE;
991            break;            break;
992            }            }
993          c = c * 10 + *(++ptr) - CHAR_0;          s = s * 10 + (int)(*(++ptr) - CHAR_0);
994          }          }
995        if (((unsigned int)c) > INT_MAX) /* Integer overflow */        if (overflow) /* Integer overflow */
996          {          {
997          while (IS_DIGIT(ptr[1]))          while (IS_DIGIT(ptr[1]))
998            ptr++;            ptr++;
999          *errorcodeptr = ERR61;          *errorcodeptr = ERR61;
1000          break;          break;
1001          }          }
1002        if (c < 10 || c <= bracount)        if (s < 10 || s <= bracount)
1003          {          {
1004          escape = ESC_REF + c;          escape = -s;
1005          break;          break;
1006          }          }
1007        ptr = oldptr;      /* Put the pointer back and fall through */        ptr = oldptr;      /* Put the pointer back and fall through */
# Line 1042  else Line 1048  else
1048          c = 0;          c = 0;
1049          for (i = 0; i < 2; ++i)          for (i = 0; i < 2; ++i)
1050            {            {
1051            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
1052  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1053            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1054            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 1060  else Line 1066  else
1066        const pcre_uchar *pt = ptr + 2;        const pcre_uchar *pt = ptr + 2;
1067    
1068        c = 0;        c = 0;
1069          overflow = FALSE;
1070        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)
1071          {          {
1072          register int cc = *pt++;          register pcre_uint32 cc = *pt++;
1073          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */
1074    
1075    #ifdef COMPILE_PCRE32
1076            if (c >= 0x10000000l) { overflow = TRUE; break; }
1077    #endif
1078    
1079  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1080          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1081          c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 1074  else Line 1085  else
1085  #endif  #endif
1086    
1087  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
1088          if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }          if (c > (utf ? 0x10ffff : 0xff)) { overflow = TRUE; break; }
1089  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
1090          if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }          if (c > (utf ? 0x10ffff : 0xffff)) { overflow = TRUE; break; }
1091  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
1092          if (utf && c > 0x10ffff) { c = -1; break; }          if (utf && c > 0x10ffff) { overflow = TRUE; break; }
1093  #endif  #endif
1094          }          }
1095    
1096        if (c < 0)        if (overflow)
1097          {          {
1098          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;
1099          *errorcodeptr = ERR34;          *errorcodeptr = ERR34;
# Line 1104  else Line 1115  else
1115      c = 0;      c = 0;
1116      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)
1117        {        {
1118        int cc;                                  /* Some compilers don't like */        pcre_uint32 cc;                          /* Some compilers don't like */
1119        cc = *(++ptr);                           /* ++ in initializers */        cc = *(++ptr);                           /* ++ in initializers */
1120  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1121        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */
# Line 1192  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  int c, i, bot, top;  pcre_uchar c;
1218    int i, bot, top;
1219  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
1220  pcre_uchar name[32];  pcre_uchar name[32];
1221    
# Line 1248  top = PRIV(utt_size); Line 1262  top = PRIV(utt_size);
1262    
1263  while (bot < top)  while (bot < top)
1264    {    {
1265      int r;
1266    i = (bot + top) >> 1;    i = (bot + top) >> 1;
1267    c = 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 (c == 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 (c > 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 1300  int max = -1; Line 1316  int max = -1;
1316  /* Read the minimum value and do a paranoid check: a negative value indicates  /* Read the minimum value and do a paranoid check: a negative value indicates
1317  an integer overflow. */  an integer overflow. */
1318    
1319  while (IS_DIGIT(*p)) min = min * 10 + *p++ - CHAR_0;  while (IS_DIGIT(*p)) min = min * 10 + (int)(*p++ - CHAR_0);
1320  if (min < 0 || min > 65535)  if (min < 0 || min > 65535)
1321    {    {
1322    *errorcodeptr = ERR5;    *errorcodeptr = ERR5;
# Line 1315  if (*p == CHAR_RIGHT_CURLY_BRACKET) max Line 1331  if (*p == CHAR_RIGHT_CURLY_BRACKET) max
1331    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
1332      {      {
1333      max = 0;      max = 0;
1334      while(IS_DIGIT(*p)) max = max * 10 + *p++ - CHAR_0;      while(IS_DIGIT(*p)) max = max * 10 + (int)(*p++ - CHAR_0);
1335      if (max < 0 || max > 65535)      if (max < 0 || max > 65535)
1336        {        {
1337        *errorcodeptr = ERR5;        *errorcodeptr = ERR5;
# Line 1446  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1462  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1462      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&
1463          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)
1464        {        {
1465        int term;        pcre_uchar term;
1466        const pcre_uchar *thisname;        const pcre_uchar *thisname;
1467        *count += 1;        *count += 1;
1468        if (name == NULL && *count == lorn) return *count;        if (name == NULL && *count == lorn) return *count;
# Line 1454  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1470  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1470        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;
1471        thisname = ptr;        thisname = ptr;
1472        while (*ptr != term) ptr++;        while (*ptr != term) ptr++;
1473        if (name != NULL && lorn == ptr - thisname &&        if (name != NULL && lorn == (int)(ptr - thisname) &&
1474            STRNCMP_UC_UC(name, thisname, lorn) == 0)            STRNCMP_UC_UC(name, thisname, (unsigned int)lorn) == 0)
1475          return *count;          return *count;
1476        term++;        term++;
1477        }        }
# Line 1733  for (;;) Line 1749  for (;;)
1749    {    {
1750    int d;    int d;
1751    pcre_uchar *ce, *cs;    pcre_uchar *ce, *cs;
1752    register int op = *cc;    register pcre_uchar op = *cc;
1753    
1754    switch (op)    switch (op)
1755      {      {
# Line 1841  for (;;) Line 1857  for (;;)
1857      case OP_NOTI:      case OP_NOTI:
1858      branchlength++;      branchlength++;
1859      cc += 2;      cc += 2;
1860  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1861      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1862  #endif  #endif
1863      break;      break;
# Line 1853  for (;;) Line 1869  for (;;)
1869      case OP_EXACTI:      case OP_EXACTI:
1870      case OP_NOTEXACT:      case OP_NOTEXACT:
1871      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1872      branchlength += GET2(cc,1);      branchlength += (int)GET2(cc,1);
1873      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1874  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1875      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1876  #endif  #endif
1877      break;      break;
# Line 1898  for (;;) Line 1914  for (;;)
1914    
1915      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1916    
 #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  
   
1917      case OP_CLASS:      case OP_CLASS:
1918      case OP_NCLASS:      case OP_NCLASS:
1919    #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32
1920        case OP_XCLASS:
1921        /* The original code caused an unsigned overflow in 64 bit systems,
1922        so now we use a conditional statement. */
1923        if (op == OP_XCLASS)
1924          cc += GET(cc, 1);
1925        else
1926          cc += PRIV(OP_lengths)[OP_CLASS];
1927    #else
1928      cc += PRIV(OP_lengths)[OP_CLASS];      cc += PRIV(OP_lengths)[OP_CLASS];
1929    #endif
1930    
1931      switch (*cc)      switch (*cc)
1932        {        {
# Line 1921  for (;;) Line 1941  for (;;)
1941        case OP_CRRANGE:        case OP_CRRANGE:
1942        case OP_CRMINRANGE:        case OP_CRMINRANGE:
1943        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;
1944        branchlength += GET2(cc,1);        branchlength += (int)GET2(cc,1);
1945        cc += 1 + 2 * IMM2_SIZE;        cc += 1 + 2 * IMM2_SIZE;
1946        break;        break;
1947    
# Line 2048  PRIV(find_bracket)(const pcre_uchar *cod Line 2068  PRIV(find_bracket)(const pcre_uchar *cod
2068  {  {
2069  for (;;)  for (;;)
2070    {    {
2071    register int c = *code;    register pcre_uchar c = *code;
2072    
2073    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2074    
# Line 2071  for (;;) Line 2091  for (;;)
2091    else if (c == OP_CBRA || c == OP_SCBRA ||    else if (c == OP_CBRA || c == OP_SCBRA ||
2092             c == OP_CBRAPOS || c == OP_SCBRAPOS)             c == OP_CBRAPOS || c == OP_SCBRAPOS)
2093      {      {
2094      int n = GET2(code, 1+LINK_SIZE);      int n = (int)GET2(code, 1+LINK_SIZE);
2095      if (n == number) return (pcre_uchar *)code;      if (n == number) return (pcre_uchar *)code;
2096      code += PRIV(OP_lengths)[c];      code += PRIV(OP_lengths)[c];
2097      }      }
# Line 2186  find_recurse(const pcre_uchar *code, BOO Line 2206  find_recurse(const pcre_uchar *code, BOO
2206  {  {
2207  for (;;)  for (;;)
2208    {    {
2209    register int c = *code;    register pcre_uchar c = *code;
2210    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2211    if (c == OP_RECURSE) return code;    if (c == OP_RECURSE) return code;
2212    
# Line 2340  static BOOL Line 2360  static BOOL
2360  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,
2361    BOOL utf, compile_data *cd)    BOOL utf, compile_data *cd)
2362  {  {
2363  register int c;  register pcre_uchar c;
2364  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
2365       code < endcode;       code < endcode;
2366       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))
# Line 2374  for (code = first_significant_code(code Line 2394  for (code = first_significant_code(code
2394      /* Test for forward reference */      /* Test for forward reference */
2395    
2396      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)
2397        if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE;        if ((int)GET(scode, 0) == (int)(code + 1 - cd->start_code)) return TRUE;
2398    
2399      /* Not a forward reference, test for completed backward reference */      /* Not a forward reference, test for completed backward reference */
2400    
# Line 2695  Returns:   TRUE or FALSE Line 2715  Returns:   TRUE or FALSE
2715  static BOOL  static BOOL
2716  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)
2717  {  {
2718  int terminator;          /* Don't combine these lines; the Solaris cc */  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */
2719  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
2720  for (++ptr; *ptr != 0; ptr++)  for (++ptr; *ptr != 0; ptr++)
2721    {    {
# Line 2744  register int yield = 0; Line 2764  register int yield = 0;
2764  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
2765    {    {
2766    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
2767      STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;      STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield;
2768    pn += posix_name_lengths[yield] + 1;    pn += posix_name_lengths[yield] + 1;
2769    yield++;    yield++;
2770    }    }
# Line 2799  while ((ptr = (pcre_uchar *)find_recurse Line 2819  while ((ptr = (pcre_uchar *)find_recurse
2819    
2820    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
2821      {      {
2822      offset = GET(hc, 0);      offset = (int)GET(hc, 0);
2823      if (cd->start_code + offset == ptr + 1)      if (cd->start_code + offset == ptr + 1)
2824        {        {
2825        PUT(hc, 0, offset + adjust);        PUT(hc, 0, offset + adjust);
# Line 2812  while ((ptr = (pcre_uchar *)find_recurse Line 2832  while ((ptr = (pcre_uchar *)find_recurse
2832    
2833    if (hc >= cd->hwm)    if (hc >= cd->hwm)
2834      {      {
2835      offset = GET(ptr, 1);      offset = (int)GET(ptr, 1);
2836      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
2837      }      }
2838    
# Line 2898  Yield:        -1 when no more Line 2918  Yield:        -1 when no more
2918  */  */
2919    
2920  static int  static int
2921  get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr,  get_othercase_range(pcre_uint32 *cptr, pcre_uint32 d, pcre_uint32 *ocptr,
2922    unsigned int *odptr)    pcre_uint32 *odptr)
2923  {  {
2924  unsigned int c, othercase, next;  pcre_uint32 c, othercase, next;
2925  int co;  unsigned int co;
2926    
2927  /* 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
2928  cases, return its case offset value. */  cases, return its case offset value. */
# Line 2913  for (c = *cptr; c <= d; c++) Line 2933  for (c = *cptr; c <= d; c++)
2933      {      {
2934      *ocptr = c++;   /* Character that has the set */      *ocptr = c++;   /* Character that has the set */
2935      *cptr = c;      /* Rest of input range */      *cptr = c;      /* Rest of input range */
2936      return co;      return (int)co;
2937      }      }
2938    if ((othercase = UCD_OTHERCASE(c)) != c) break;    if ((othercase = UCD_OTHERCASE(c)) != c) break;
2939    }    }
# Line 2953  Returns:       TRUE if auto-possessifyin Line 2973  Returns:       TRUE if auto-possessifyin
2973  */  */
2974    
2975  static BOOL  static BOOL
2976  check_char_prop(int c, int ptype, int pdata, BOOL negated)  check_char_prop(pcre_uint32 c, unsigned int ptype, unsigned int pdata, BOOL negated)
2977  {  {
2978  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2979  const pcre_uint32 *p;  const pcre_uint32 *p;
# Line 3039  static BOOL Line 3059  static BOOL
3059  check_auto_possessive(const pcre_uchar *previous, BOOL utf,  check_auto_possessive(const pcre_uchar *previous, BOOL utf,
3060    const pcre_uchar *ptr, int options, compile_data *cd)    const pcre_uchar *ptr, int options, compile_data *cd)
3061  {  {
3062  pcre_int32 c = NOTACHAR; // FIXMEchpe pcre_uint32  pcre_uint32 c = NOTACHAR;
3063  pcre_int32 next;  pcre_uint32 next;
3064  int escape;  int escape;
3065  int op_code = *previous++;  pcre_uchar op_code = *previous++;
3066    
3067  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
3068    
# Line 3120  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3140  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3140    
3141  if (op_code == OP_CHAR || op_code == OP_CHARI ||  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3142      op_code == OP_NOT || op_code == OP_NOTI)      op_code == OP_NOT || op_code == OP_NOTI)
   //if (escape == 0) switch(op_code)  
3143    {    {
3144  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3145    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
# Line 3138  if (escape == 0) Line 3157  if (escape == 0)
3157    case, which maps to the special PT_CLIST property. Check this first. */    case, which maps to the special PT_CLIST property. Check this first. */
3158    
3159  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3160    if (utf && (unsigned int)c != NOTACHAR && (options & PCRE_CASELESS) != 0)    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3161      {      {
3162      int ocs = UCD_CASESET(next);      unsigned int ocs = UCD_CASESET(next);
3163      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);
3164      }      }
3165  #endif  #endif
# Line 3160  if (escape == 0) Line 3179  if (escape == 0)
3179  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3180      if (utf)      if (utf)
3181        {        {
3182        unsigned int othercase;        pcre_uint32 othercase;
3183        if (next < 128) othercase = cd->fcc[next]; else        if (next < 128) othercase = cd->fcc[next]; else
3184  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3185        othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE(next);
3186  #else  #else
3187        othercase = NOTACHAR;        othercase = NOTACHAR;
3188  #endif  #endif
3189        return (unsigned int)c != othercase;        return c != othercase;
3190        }        }
3191      else      else
3192  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3193      return (c != TABLE_GET((unsigned int)next, cd->fcc, next));  /* Not UTF */      return (c != TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3194    
3195      case OP_NOT:      case OP_NOT:
3196      return c == next;      return c == next;
# Line 3181  if (escape == 0) Line 3200  if (escape == 0)
3200  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3201      if (utf)      if (utf)
3202        {        {
3203        unsigned int othercase;        pcre_uint32 othercase;
3204        if (next < 128) othercase = cd->fcc[next]; else        if (next < 128) othercase = cd->fcc[next]; else
3205  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3206        othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE(next);
3207  #else  #else
3208        othercase = NOTACHAR;        othercase = NOTACHAR;
3209  #endif  #endif
3210        return (unsigned int)c == othercase;        return c == othercase;
3211        }        }
3212      else      else
3213  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3214      return (c == TABLE_GET((unsigned int)next, cd->fcc, next));  /* Not UTF */      return (c == TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3215    
3216      /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.      /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
3217      When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */      When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
# Line 3326  switch(op_code) Line 3345  switch(op_code)
3345      case ESC_p:      case ESC_p:
3346      case ESC_P:      case ESC_P:
3347        {        {
3348        int ptype, pdata, errorcodeptr;        unsigned int ptype = 0, pdata = 0;
3349          int errorcodeptr;
3350        BOOL negated;        BOOL negated;
3351    
3352        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
3353        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
3354        if (ptype < 0) return FALSE;          return FALSE;
3355        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
3356    
3357        /* 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 3655  Arguments: Line 3675  Arguments:
3675    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3676    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3677    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3678    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr    place to put the first required character
3679    reqcharptr     set to the last literal character required, else < 0    firstcharflagsptr place to put the first character flags, or a negative number
3680      reqcharptr     place to put the last required character
3681      reqcharflagsptr place to put the last required character flags, or a negative number
3682    bcptr          points to current branch chain    bcptr          points to current branch chain
3683    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3684    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3669  Returns:         TRUE on success Line 3691  Returns:         TRUE on success
3691    
3692  static BOOL  static BOOL
3693  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3694    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,    const pcre_uchar **ptrptr, int *errorcodeptr,
3695    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,    pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
3696      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
3697      branch_chain *bcptr, int cond_depth,
3698    compile_data *cd, int *lengthptr)    compile_data *cd, int *lengthptr)
3699  {  {
3700  int repeat_type, op_type;  int repeat_type, op_type;
3701  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3702  int bravalue = 0;  int bravalue = 0;
3703  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3704  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
3705  pcre_int32 zeroreqchar, zerofirstchar;  pcre_int32 firstcharflags, reqcharflags;
3706    pcre_uint32 zeroreqchar, zerofirstchar;
3707    pcre_int32 zeroreqcharflags, zerofirstcharflags;
3708  pcre_int32 req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3709  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3710  int after_manual_callout = 0;  int after_manual_callout = 0;
3711  int length_prevgroup = 0;  int length_prevgroup = 0;
3712  register int c;  register pcre_uint32 c;
3713  int escape;  int escape;
3714  register pcre_uchar *code = *codeptr;  register pcre_uchar *code = *codeptr;
3715  pcre_uchar *last_code = code;  pcre_uchar *last_code = code;
# Line 3706  dynamically as we process the pattern. * Line 3732  dynamically as we process the pattern. *
3732  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3733  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3734  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3735    #ifndef COMPILE_PCRE32
3736  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3737    #endif
3738  #else  #else
3739  BOOL utf = FALSE;  BOOL utf = FALSE;
3740  #endif  #endif
# Line 3741  to take the zero repeat into account. Th Line 3769  to take the zero repeat into account. Th
3769  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3770  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3771    
3772  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
3773    firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;
3774    
3775  /* The variable req_caseopt contains either the REQ_CASELESS value  /* The variable req_caseopt contains either the REQ_CASELESS value
3776  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 3767  for (;; ptr++) Line 3796  for (;; ptr++)
3796    int recno;    int recno;
3797    int refsign;    int refsign;
3798    int skipbytes;    int skipbytes;
3799    int subreqchar;    pcre_uint32 subreqchar, subfirstchar;
3800    int subfirstchar;    pcre_int32 subreqcharflags, subfirstcharflags;
3801    int terminator;    int terminator;
3802    int mclength;    unsigned int mclength;
3803    int tempbracount;    unsigned int tempbracount;
3804    int ec; // FIXMEchpe pcre_uint32    pcre_uint32 ec;
3805    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3806    
3807    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 3935  for (;; ptr++) Line 3964  for (;; ptr++)
3964      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3965      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3966      *firstcharptr = firstchar;      *firstcharptr = firstchar;
3967        *firstcharflagsptr = firstcharflags;
3968      *reqcharptr = reqchar;      *reqcharptr = reqchar;
3969        *reqcharflagsptr = reqcharflags;
3970      *codeptr = code;      *codeptr = code;
3971      *ptrptr = ptr;      *ptrptr = ptr;
3972      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3959  for (;; ptr++) Line 3990  for (;; ptr++)
3990      previous = NULL;      previous = NULL;
3991      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3992        {        {
3993        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
3994        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3995        }        }
3996      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3974  for (;; ptr++) Line 4005  for (;; ptr++)
4005      repeats. The value of reqchar doesn't change either. */      repeats. The value of reqchar doesn't change either. */
4006    
4007      case CHAR_DOT:      case CHAR_DOT:
4008      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4009      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4010        zerofirstcharflags = firstcharflags;
4011      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4012        zeroreqcharflags = reqcharflags;
4013      previous = code;      previous = code;
4014      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
4015      break;      break;
# Line 4050  for (;; ptr++) Line 4083  for (;; ptr++)
4083          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
4084        {        {
4085        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
4086        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4087        zerofirstchar = firstchar;        zerofirstchar = firstchar;
4088          zerofirstcharflags = firstcharflags;
4089        break;        break;
4090        }        }
4091    
# Line 4090  for (;; ptr++) Line 4124  for (;; ptr++)
4124        {        {
4125        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
4126    
4127  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
4128        if (utf && HAS_EXTRALEN(c))        if (utf && HAS_EXTRALEN(c))
4129          {                           /* Braces are required because the */          {                           /* Braces are required because the */
4130          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
# Line 4349  for (;; ptr++) Line 4383  for (;; ptr++)
4383              case ESC_P:              case ESC_P:
4384                {                {
4385                BOOL negated;                BOOL negated;
4386                int pdata;                unsigned int ptype = 0, pdata = 0;
4387                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
4388                if (ptype < 0) goto FAILED;                  goto FAILED;
4389                *class_uchardata++ = ((escape == ESC_p) != negated)?                *class_uchardata++ = ((escape == ESC_p) != negated)?
4390                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4391                *class_uchardata++ = ptype;                *class_uchardata++ = ptype;
# Line 4405  for (;; ptr++) Line 4439  for (;; ptr++)
4439    
4440        if (!inescq && ptr[1] == CHAR_MINUS)        if (!inescq && ptr[1] == CHAR_MINUS)
4441          {          {
4442          int d;          pcre_uint32 d;
4443          ptr += 2;          ptr += 2;
4444          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;
4445    
# Line 4453  for (;; ptr++) Line 4487  for (;; ptr++)
4487    
4488            /* \b is backspace; any other special means the '-' was literal. */            /* \b is backspace; any other special means the '-' was literal. */
4489    
4490            if (descape > 0)            if (descape != 0)
4491              {              {
4492              if (descape == ESC_b) d = CHAR_BS; else              if (descape == ESC_b) d = CHAR_BS; else
4493                {                {
# Line 4511  for (;; ptr++) Line 4545  for (;; ptr++)
4545          {          {
4546          ptr++;          ptr++;
4547          zeroreqchar = reqchar;          zeroreqchar = reqchar;
4548            zeroreqcharflags = reqcharflags;
4549    
4550          if (negate_class)          if (negate_class)
4551            {            {
4552  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
             // FIXMEchpe pcreuint32?  
4553            int d;            int d;
4554  #endif  #endif
4555            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;            if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4556            zerofirstchar = firstchar;            zerofirstchar = firstchar;
4557              zerofirstcharflags = firstcharflags;
4558    
4559            /* For caseless UTF-8 mode when UCP support is available, check            /* For caseless UTF-8 mode when UCP support is available, check
4560            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 4607  for (;; ptr++) Line 4642  for (;; ptr++)
4642      setting, whatever the repeat count. Any reqchar setting must remain      setting, whatever the repeat count. Any reqchar setting must remain
4643      unchanged after any kind of repeat. */      unchanged after any kind of repeat. */
4644    
4645      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4646      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4647        zerofirstcharflags = firstcharflags;
4648      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4649        zeroreqcharflags = reqcharflags;
4650    
4651      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4652      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 4704  for (;; ptr++) Line 4741  for (;; ptr++)
4741      if (repeat_min == 0)      if (repeat_min == 0)
4742        {        {
4743        firstchar = zerofirstchar;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4744          firstcharflags = zerofirstcharflags;
4745        reqchar = zeroreqchar;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4746          reqcharflags = zeroreqcharflags;
4747        }        }
4748    
4749      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4807  for (;; ptr++) Line 4846  for (;; ptr++)
4846          {          {
4847          c = code[-1];          c = code[-1];
4848          if (*previous <= OP_CHARI && repeat_min > 1)          if (*previous <= OP_CHARI && repeat_min > 1)
4849            reqchar = c | req_caseopt | cd->req_varyopt;            {
4850              reqchar = c;
4851              reqcharflags = req_caseopt | cd->req_varyopt;
4852              }
4853          }          }
4854    
4855        /* 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 5189  for (;; ptr++) Line 5231  for (;; ptr++)
5231    
5232            else            else
5233              {              {
5234              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;              if (groupsetfirstchar && reqcharflags < 0)
5235                  {
5236                  reqchar = firstchar;
5237                  reqcharflags = firstcharflags;
5238                  }
5239    
5240              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5241                {                {
# Line 5468  for (;; ptr++) Line 5514  for (;; ptr++)
5514        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)
5515          {          {
5516          tempcode += PRIV(OP_lengths)[*tempcode];          tempcode += PRIV(OP_lengths)[*tempcode];
5517  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
5518          if (utf && HAS_EXTRALEN(tempcode[-1]))          if (utf && HAS_EXTRALEN(tempcode[-1]))
5519            tempcode += GET_EXTRALEN(tempcode[-1]);            tempcode += GET_EXTRALEN(tempcode[-1]);
5520  #endif  #endif
# Line 5609  for (;; ptr++) Line 5655  for (;; ptr++)
5655                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5656    
5657              /* Do not set firstchar after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5658              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
5659              }              }
5660    
5661            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5775  for (;; ptr++) Line 5821  for (;; ptr++)
5821          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)
5822            {            {
5823            if (recno >= 0)            if (recno >= 0)
5824              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;              recno = (IS_DIGIT(*ptr))? recno * 10 + (int)(*ptr - CHAR_0) : -1;
5825            ptr++;            ptr++;
5826            }            }
5827          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5828    
5829          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) ||
5830              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
5831            {            {
5832            ptr--;      /* Error offset */            ptr--;      /* Error offset */
# Line 6020  for (;; ptr++) Line 6066  for (;; ptr++)
6066    
6067            if (lengthptr != NULL)            if (lengthptr != NULL)
6068              {              {
6069              if (*ptr != terminator)              if (*ptr != (pcre_uchar)terminator)
6070                {                {
6071                *errorcodeptr = ERR42;                *errorcodeptr = ERR42;
6072                goto FAILED;                goto FAILED;
# Line 6162  for (;; ptr++) Line 6208  for (;; ptr++)
6208              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
6209              goto FAILED;              goto FAILED;
6210              }              }
6211            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6212              {              {
6213              *errorcodeptr = ERR42;              *errorcodeptr = ERR42;
6214              goto FAILED;              goto FAILED;
# Line 6268  for (;; ptr++) Line 6314  for (;; ptr++)
6314            while(IS_DIGIT(*ptr))            while(IS_DIGIT(*ptr))
6315              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6316    
6317            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6318              {              {
6319              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
6320              goto FAILED;              goto FAILED;
# Line 6372  for (;; ptr++) Line 6418  for (;; ptr++)
6418    
6419          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6420    
6421          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6422          continue;          continue;
6423    
6424    
# Line 6506  for (;; ptr++) Line 6552  for (;; ptr++)
6552           cond_depth +           cond_depth +
6553             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6554           &subfirstchar,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6555             &subfirstcharflags,
6556           &subreqchar,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6557             &subreqcharflags,
6558           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6559           cd,                              /* Tables block */           cd,                              /* Tables block */
6560           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6567  for (;; ptr++) Line 6615  for (;; ptr++)
6615            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6616            goto FAILED;            goto FAILED;
6617            }            }
6618          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;          if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
6619          }          }
6620        }        }
6621    
# Line 6616  for (;; ptr++) Line 6664  for (;; ptr++)
6664      back off. */      back off. */
6665    
6666      zeroreqchar = reqchar;      zeroreqchar = reqchar;
6667        zeroreqcharflags = reqcharflags;
6668      zerofirstchar = firstchar;      zerofirstchar = firstchar;
6669        zerofirstcharflags = firstcharflags;
6670      groupsetfirstchar = FALSE;      groupsetfirstchar = FALSE;
6671    
6672      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
# Line 6627  for (;; ptr++) Line 6677  for (;; ptr++)
6677        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
6678        repeat forces firstchar to "none". */        repeat forces firstchar to "none". */
6679    
6680        if (firstchar == REQ_UNSET)        if (firstcharflags == REQ_UNSET)
6681          {          {
6682          if (subfirstchar >= 0)          if (subfirstcharflags >= 0)
6683            {            {
6684            firstchar = subfirstchar;            firstchar = subfirstchar;
6685              firstcharflags = subfirstcharflags;
6686            groupsetfirstchar = TRUE;            groupsetfirstchar = TRUE;
6687            }            }
6688          else firstchar = REQ_NONE;          else firstcharflags = REQ_NONE;
6689          zerofirstchar = REQ_NONE;          zerofirstcharflags = REQ_NONE;
6690          }          }
6691    
6692        /* If firstchar was previously set, convert the subpattern's firstchar        /* If firstchar was previously set, convert the subpattern's firstchar
6693        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
6694        existence beforehand. */        existence beforehand. */
6695    
6696        else if (subfirstchar >= 0 && subreqchar < 0)        else if (subfirstcharflags >= 0 && subreqcharflags < 0)
6697          subreqchar = subfirstchar | tempreqvary;          {
6698            subreqchar = subfirstchar;
6699            subreqcharflags = subfirstcharflags | tempreqvary;
6700            }
6701    
6702        /* 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
6703        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6704    
6705        if (subreqchar >= 0) reqchar = subreqchar;        if (subreqcharflags >= 0)
6706            {
6707            reqchar = subreqchar;
6708            reqcharflags = subreqcharflags;
6709            }
6710        }        }
6711    
6712      /* 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 6659  for (;; ptr++) Line 6717  for (;; ptr++)
6717      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
6718      firstchar, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6719    
6720      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;      else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
6721          {
6722          reqchar = subreqchar;
6723          reqcharflags = subreqcharflags;
6724          }
6725      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6726    
6727    
# Line 6667  for (;; ptr++) Line 6729  for (;; ptr++)
6729      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
6730      are arranged to be the negation of the corresponding OP_values in the      are arranged to be the negation of the corresponding OP_values in the
6731      default case when PCRE_UCP is not set. For the back references, the values      default case when PCRE_UCP is not set. For the back references, the values
6732      are ESC_REF plus the reference number. Only back references and those types      are negative the reference number. Only back references and those types
6733      that consume a character may be repeated. We can test for values between      that consume a character may be repeated. We can test for values between
6734      ESC_b and ESC_Z for the latter; this may have to change if any new ones are      ESC_b and ESC_Z for the latter; this may have to change if any new ones are
6735      ever created. */      ever created. */
# Line 6695  for (;; ptr++) Line 6757  for (;; ptr++)
6757        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6758        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6759    
6760        if (firstchar == REQ_UNSET && escape > ESC_b && escape < ESC_Z)        if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6761          firstchar = REQ_NONE;          firstcharflags = REQ_NONE;
6762    
6763        /* 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. */
6764    
6765        zerofirstchar = firstchar;        zerofirstchar = firstchar;
6766          zerofirstcharflags = firstcharflags;
6767        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6768          zeroreqcharflags = reqcharflags;
6769    
6770        /* \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'
6771        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
6772        ESC_g is returned only for these cases. So we don't need to check for <        ESC_g is returned only for these cases. So we don't need to check for <
6773        or ' if the value is ESC_g. For the Perl syntax \g{n} the value is        or ' if the value is ESC_g. For the Perl syntax \g{n} the value is
6774        ESC_REF+n, and for the Perl syntax \g{name} the result is ESC_k (as        -n, and for the Perl syntax \g{name} the result is ESC_k (as
6775        that is a synonym for a named back reference). */        that is a synonym for a named back reference). */
6776    
6777        if (escape == ESC_g)        if (escape == ESC_g)
# Line 6730  for (;; ptr++) Line 6794  for (;; ptr++)
6794          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
6795            {            {
6796            BOOL is_a_number = TRUE;            BOOL is_a_number = TRUE;
6797            for (p = ptr + 1; *p != 0 && *p != terminator; p++)            for (p = ptr + 1; *p != 0 && *p != (pcre_uchar)terminator; p++)
6798              {              {
6799              if (!MAX_255(*p)) { is_a_number = FALSE; break; }              if (!MAX_255(*p)) { is_a_number = FALSE; break; }
6800              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
6801              if ((cd->ctypes[*p] & ctype_word) == 0) break;              if ((cd->ctypes[*p] & ctype_word) == 0) break;
6802              }              }
6803            if (*p != terminator)            if (*p != (pcre_uchar)terminator)
6804              {              {
6805              *errorcodeptr = ERR57;              *errorcodeptr = ERR57;
6806              break;              break;
# Line 6754  for (;; ptr++) Line 6818  for (;; ptr++)
6818    
6819          p = ptr + 2;          p = ptr + 2;
6820          while (IS_DIGIT(*p)) p++;          while (IS_DIGIT(*p)) p++;
6821          if (*p != terminator)          if (*p != (pcre_uchar)terminator)
6822            {            {
6823            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
6824            break;            break;
# Line 6785  for (;; ptr++) Line 6849  for (;; ptr++)
6849        not set to cope with cases like (?=(\w+))\1: which would otherwise set        not set to cope with cases like (?=(\w+))\1: which would otherwise set
6850        ':' later. */        ':' later. */
6851    
6852        if (escape >= ESC_REF)        if (escape < 0)
6853          {          {
6854          open_capitem *oc;          open_capitem *oc;
6855          recno = escape - ESC_REF;          recno = -escape;
6856    
6857          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6858          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6859          previous = code;          previous = code;
6860          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6861          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6818  for (;; ptr++) Line 6882  for (;; ptr++)
6882        else if (escape == ESC_P || escape == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6883          {          {
6884          BOOL negated;          BOOL negated;
6885          int pdata;          unsigned int ptype = 0, pdata = 0;
6886          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
6887          if (ptype < 0) goto FAILED;            goto FAILED;
6888          previous = code;          previous = code;
6889          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6890          *code++ = ptype;          *code++ = ptype;
# Line 6894  for (;; ptr++) Line 6958  for (;; ptr++)
6958      mclength = 1;      mclength = 1;
6959      mcbuffer[0] = c;      mcbuffer[0] = c;
6960    
6961  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
6962      if (utf && HAS_EXTRALEN(c))      if (utf && HAS_EXTRALEN(c))
6963        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));
6964  #endif  #endif
# Line 6918  for (;; ptr++) Line 6982  for (;; ptr++)
6982          *code++ = OP_PROP;          *code++ = OP_PROP;
6983          *code++ = PT_CLIST;          *code++ = PT_CLIST;
6984          *code++ = c;          *code++ = c;
6985          if (firstchar == REQ_UNSET) firstchar = zerofirstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = zerofirstcharflags = REQ_NONE;
6986          break;          break;
6987          }          }
6988        }        }
# Line 6939  for (;; ptr++) Line 7003  for (;; ptr++)
7003      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
7004      repeat. */      repeat. */
7005    
7006      if (firstchar == REQ_UNSET)      if (firstcharflags == REQ_UNSET)
7007        {        {
7008        zerofirstchar = REQ_NONE;        zerofirstcharflags = REQ_NONE;
7009        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7010          zeroreqcharflags = reqcharflags;
7011    
7012        /* 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
7013        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
# Line 6950  for (;; ptr++) Line 7015  for (;; ptr++)
7015        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7016          {          {
7017          firstchar = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
7018          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;          firstchar = mcbuffer[0];
7019            firstcharflags = req_caseopt;
7020    
7021            if (mclength != 1)
7022              {
7023              reqchar = code[-1];
7024              reqcharflags = cd->req_varyopt;
7025              }
7026          }          }
7027        else firstchar = reqchar = REQ_NONE;        else firstcharflags = reqcharflags = REQ_NONE;
7028        }        }
7029    
7030      /* 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 6961  for (;; ptr++) Line 7033  for (;; ptr++)
7033      else      else
7034        {        {
7035        zerofirstchar = firstchar;        zerofirstchar = firstchar;
7036          zerofirstcharflags = firstcharflags;
7037        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7038          zeroreqcharflags = reqcharflags;
7039        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7040          reqchar = code[-1] | req_caseopt | cd->req_varyopt;          {
7041            reqchar = code[-1];
7042            reqcharflags = req_caseopt | cd->req_varyopt;
7043            }
7044        }        }
7045    
7046      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6982  return FALSE; Line 7059  return FALSE;
7059    
7060    
7061    
   
7062  /*************************************************  /*************************************************
7063  *     Compile sequence of alternatives           *  *     Compile sequence of alternatives           *
7064  *************************************************/  *************************************************/
# Line 7003  Arguments: Line 7079  Arguments:
7079    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
7080    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
7081    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
7082    firstcharptr   place to put the first required character, or a negative number    firstcharptr    place to put the first required character
7083    reqcharptr     place to put the last required character, or a negative number    firstcharflagsptr place to put the first character flags, or a negative number
7084      reqcharptr     place to put the last required character
7085      reqcharflagsptr place to put the last required character flags, or a negative number
7086    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
7087    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
7088    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 7016  Returns:         TRUE on success Line 7094  Returns:         TRUE on success
7094  static BOOL  static BOOL
7095  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
7096    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
7097    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,    int cond_depth,
7098      pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
7099      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
7100    branch_chain *bcptr, compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
7101  {  {
7102  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
# Line 7026  pcre_uchar *start_bracket = code; Line 7106  pcre_uchar *start_bracket = code;
7106  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
7107  open_capitem capitem;  open_capitem capitem;
7108  int capnumber = 0;  int capnumber = 0;
7109  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7110  pcre_int32 branchfirstchar, branchreqchar;  pcre_int32 firstcharflags, reqcharflags;
7111    pcre_uint32 branchfirstchar, branchreqchar;
7112    pcre_int32 branchfirstcharflags, branchreqcharflags;
7113  int length;  int length;
7114  int orig_bracount;  unsigned int orig_bracount;
7115  int max_bracount;  unsigned int max_bracount;
7116  branch_chain bc;  branch_chain bc;
7117    
7118  bc.outer = bcptr;  bc.outer = bcptr;
7119  bc.current_branch = code;  bc.current_branch = code;
7120    
7121  firstchar = reqchar = REQ_UNSET;  firstchar = reqchar = 0;
7122    firstcharflags = reqcharflags = REQ_UNSET;
7123    
7124  /* 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
7125  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 7096  for (;;) Line 7179  for (;;)
7179    into the length. */    into the length. */
7180    
7181    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
7182          &branchreqchar, &bc, cond_depth, cd,          &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
7183          (lengthptr == NULL)? NULL : &length))          cond_depth, cd, (lengthptr == NULL)? NULL : &length))
7184      {      {
7185      *ptrptr = ptr;      *ptrptr = ptr;
7186      return FALSE;      return FALSE;
# Line 7118  for (;;) Line 7201  for (;;)
7201      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
7202        {        {
7203        firstchar = branchfirstchar;        firstchar = branchfirstchar;
7204          firstcharflags = branchfirstcharflags;
7205        reqchar = branchreqchar;        reqchar = branchreqchar;
7206          reqcharflags = branchreqcharflags;
7207        }        }
7208    
7209      /* 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 7132  for (;;) Line 7217  for (;;)
7217        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
7218        previously no reqchar, it takes on the value of the old firstchar. */        previously no reqchar, it takes on the value of the old firstchar. */
7219    
7220        if (firstchar >= 0 && firstchar != branchfirstchar)        if (firstcharflags >= 0 &&
7221              (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar))
7222          {          {
7223          if (reqchar < 0) reqchar = firstchar;          if (reqcharflags < 0)
7224          firstchar = REQ_NONE;            {
7225              reqchar = firstchar;
7226              reqcharflags = firstcharflags;
7227              }
7228            firstcharflags = REQ_NONE;
7229          }          }
7230    
7231        /* 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
7232        branch becomes a reqchar if there isn't a branch reqchar. */        branch becomes a reqchar if there isn't a branch reqchar. */
7233    
7234        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)        if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
7235            branchreqchar = branchfirstchar;          {
7236            branchreqchar = branchfirstchar;
7237            branchreqcharflags = branchfirstcharflags;
7238            }
7239    
7240        /* Now ensure that the reqchars match */        /* Now ensure that the reqchars match */
7241    
7242        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))        if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) ||
7243          reqchar = REQ_NONE;            reqchar != branchreqchar)
7244        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */          reqcharflags = REQ_NONE;
7245          else
7246            {
7247            reqchar = branchreqchar;
7248            reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */
7249            }
7250        }        }
7251    
7252      /* 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 7244  for (;;) Line 7342  for (;;)
7342      *codeptr = code;      *codeptr = code;
7343      *ptrptr = ptr;      *ptrptr = ptr;
7344      *firstcharptr = firstchar;      *firstcharptr = firstchar;
7345        *firstcharflagsptr = firstcharflags;
7346      *reqcharptr = reqchar;      *reqcharptr = reqchar;
7347        *reqcharflagsptr = reqcharflags;
7348      if (lengthptr != NULL)      if (lengthptr != NULL)
7349        {        {
7350        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7531  we return that char, otherwise -1. Line 7631  we return that char, otherwise -1.
7631    
7632  Arguments:  Arguments:
7633    code       points to start of expression (the bracket)    code       points to start of expression (the bracket)
7634      flags       points to the first char flags, or to REQ_NONE
7635    inassert   TRUE if in an assertion    inassert   TRUE if in an assertion
7636    
7637  Returns:     -1 or the fixed first char  Returns:     the fixed first char, or 0 with REQ_NONE in flags
7638  */  */
7639    
7640  static int  static pcre_uint32
7641  find_firstassertedchar(const pcre_uchar *code, BOOL inassert)  find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags,
7642      BOOL inassert)
7643  {  {
7644  register int c = -1;  register pcre_uint32 c = 0;
7645    int cflags = REQ_NONE;
7646    
7647    *flags = REQ_NONE;
7648  do {  do {
7649     int d;     pcre_uint32 d;
7650       int dflags;
7651     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7652               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7653     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7654       TRUE);       TRUE);
7655     register int op = *scode;     register pcre_uchar op = *scode;
7656    
7657     switch(op)     switch(op)
7658       {       {
7659       default:       default:
7660       return -1;       return 0;
7661    
7662       case OP_BRA:       case OP_BRA:
7663       case OP_BRAPOS:       case OP_BRAPOS:
# Line 7563  do { Line 7669  do {
7669       case OP_ONCE:       case OP_ONCE:
7670       case OP_ONCE_NC:       case OP_ONCE_NC:
7671       case OP_COND:       case OP_COND:
7672       if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)       d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
7673         return -1;       if (dflags < 0)
7674       if (c < 0) c = d; else if (c != d) return -1;         return 0;
7675         if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0;
7676       break;       break;
7677    
7678       case OP_EXACT:       case OP_EXACT:
# Line 7576  do { Line 7683  do {
7683       case OP_PLUS:       case OP_PLUS:
7684       case OP_MINPLUS:       case OP_MINPLUS:
7685       case OP_POSPLUS:       case OP_POSPLUS:
7686       if (!inassert) return -1;       if (!inassert) return 0;
7687       if (c < 0) c = scode[1];       if (cflags < 0) { c = scode[1]; cflags = 0; }
7688         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7689       break;       break;
7690    
7691       case OP_EXACTI:       case OP_EXACTI:
# Line 7589  do { Line 7696  do {
7696       case OP_PLUSI:       case OP_PLUSI:
7697       case OP_MINPLUSI:       case OP_MINPLUSI:
7698       case OP_POSPLUSI:       case OP_POSPLUSI:
7699       if (!inassert) return -1;       if (!inassert) return 0;
7700       if (c < 0) c = scode[1] | REQ_CASELESS;       if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
7701         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7702       break;       break;
7703       }       }
7704    
7705     code += GET(code, 1);     code += GET(code, 1);
7706     }     }
7707  while (*code == OP_ALT);  while (*code == OP_ALT);
7708    
7709    *flags = cflags;
7710  return c;  return c;
7711  }  }
7712    
# Line 7665  pcre32_compile2(PCRE_SPTR32 pattern, int Line 7774  pcre32_compile2(PCRE_SPTR32 pattern, int
7774  {  {
7775  REAL_PCRE *re;  REAL_PCRE *re;
7776  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7777  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7778    pcre_int32 firstcharflags, reqcharflags;
7779  int newline;  int newline;
7780  int errorcode = 0;  int errorcode = 0;
7781  int skipatstart = 0;  int skipatstart = 0;
# Line 7915  ptr += skipatstart; Line 8025  ptr += skipatstart;
8025  code = cworkspace;  code = cworkspace;
8026  *code = OP_BRA;  *code = OP_BRA;
8027  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8028    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8029      cd, &length);
8030  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
8031    
8032  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 7993  ptr = (const pcre_uchar *)pattern + skip Line 8104  ptr = (const pcre_uchar *)pattern + skip
8104  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
8105  *code = OP_BRA;  *code = OP_BRA;
8106  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
8107    &firstchar, &reqchar, NULL, cd, NULL);    &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
8108  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
8109  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
8110  re->max_lookbehind = cd->max_lookbehind;  re->max_lookbehind = cd->max_lookbehind;
8111  re->flags = cd->external_flags | PCRE_MODE;  re->flags = cd->external_flags | PCRE_MODE;
8112    
8113  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept)
8114      {
8115      reqchar = 0;              /* Must disable after (*ACCEPT) */
8116      reqcharflags = REQ_NONE;
8117      }
8118    
8119  /* 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. */
8120    
# Line 8014  if debugging, leave the test till after Line 8129  if debugging, leave the test till after
8129  if (code - codestart > length) errorcode = ERR23;  if (code - codestart > length) errorcode = ERR23;
8130  #endif  #endif
8131    
8132    #ifdef SUPPORT_VALGRIND
8133    /* If the estimated length exceeds the really used length, mark the extra
8134    allocated memory as unadressable, so that any out-of-bound reads can be
8135    detected. */
8136    VALGRIND_MAKE_MEM_NOACCESS(code, (length - (code - codestart)) * sizeof(pcre_uchar));
8137    #endif
8138    
8139  /* Fill in any forward references that are required. There may be repeated  /* Fill in any forward references that are required. There may be repeated
8140  references; optimize for them, as searching a large regex takes time. */  references; optimize for them, as searching a large regex takes time. */
8141    
# Line 8049  if (errorcode == 0 && re->top_backref > Line 8171  if (errorcode == 0 && re->top_backref >
8171    
8172  /* If there were any lookbehind assertions that contained OP_RECURSE  /* If there were any lookbehind assertions that contained OP_RECURSE
8173  (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,
8174  because they may contain forward references. Actual recursions can't be fixed  because they may contain forward references. Actual recursions cannot be fixed
8175  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
8176  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
8177  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 8120  if ((re->options & PCRE_ANCHORED) == 0) Line 8242  if ((re->options & PCRE_ANCHORED) == 0)
8242    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
8243    else    else
8244      {      {
8245      if (firstchar < 0)      if (firstcharflags < 0)
8246        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
8247      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstcharflags >= 0)   /* Remove caseless flag for non-caseable chars */
8248        {        {
8249  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8250        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8251  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8252        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8253  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8254        re->first_char = firstchar & ~REQ_MASK;        re->first_char = firstchar;
8255  #endif  #endif
8256        if ((firstchar & REQ_CASELESS) != 0)        if ((firstcharflags & REQ_CASELESS) != 0)
8257          {          {
8258  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8259          /* We ignore non-ASCII first chars in 8 bit mode. */          /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8163  if ((re->options & PCRE_ANCHORED) == 0) Line 8285  if ((re->options & PCRE_ANCHORED) == 0)
8285  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
8286  bytes. */  bytes. */
8287    
8288  if (reqchar >= 0 &&  if (reqcharflags >= 0 &&
8289       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0))
8290    {    {
8291  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8292    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8293  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8294    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8295  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8296    re->req_char = reqchar & ~REQ_MASK;    re->req_char = reqchar;
8297  #endif  #endif
8298    if ((reqchar & REQ_CASELESS) != 0)    if ((reqcharflags & REQ_CASELESS) != 0)
8299      {      {
8300  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8301      /* We ignore non-ASCII first chars in 8 bit mode. */      /* We ignore non-ASCII first chars in 8 bit mode. */

Legend:
Removed from v.1059  
changed lines
  Added in v.1152

  ViewVC Help
Powered by ViewVC 1.1.5