/[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 964 by ph10, Fri May 4 13:03:39 2012 UTC revision 1133 by chpe, Thu Oct 18 18:35:30 2012 UTC
# Line 53  supporting internal functions that are n Line 53  supporting internal functions that are n
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    
56  /* When PCRE_DEBUG is defined, we need the pcre(16)_printint() function, which  /* When PCRE_DEBUG is defined, we need the pcre(16|32)_printint() function, which
57  is also used by pcretest. PCRE_DEBUG is not defined when building a production  is also used by pcretest. PCRE_DEBUG is not defined when building a production
58  library. We do not need to select pcre16_printint.c specially, because the  library. We do not need to select pcre16_printint.c specially, because the
59  COMPILE_PCREx macro will already be appropriately set. */  COMPILE_PCREx macro will already be appropriately set. */
# Line 68  COMPILE_PCREx macro will already be appr Line 68  COMPILE_PCREx macro will already be appr
68    
69  /* Macro for setting individual bits in class bitmaps. */  /* Macro for setting individual bits in class bitmaps. */
70    
71  #define SETBIT(a,b) a[b/8] |= (1 << (b%8))  #define SETBIT(a,b) a[(b)/8] |= (1 << ((b)&7))
72    
73  /* Maximum length value to check against when making sure that the integer that  /* Maximum length value to check against when making sure that the integer that
74  holds the compiled pattern length does not overflow. We make it a bit less than  holds the compiled pattern length does not overflow. We make it a bit less than
# Line 77  to check them every time. */ Line 77  to check them every time. */
77    
78  #define OFLOW_MAX (INT_MAX - 20)  #define OFLOW_MAX (INT_MAX - 20)
79    
80    /* Definitions to allow mutual recursion */
81    
82    static int
83      add_list_to_class(pcre_uint8 *, pcre_uchar **, int, compile_data *,
84        const pcre_uint32 *, unsigned int);
85    
86    static BOOL
87      compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int,
88        pcre_uint32 *, pcre_int32 *, pcre_uint32 *, pcre_int32 *, branch_chain *,
89        compile_data *, int *);
90    
91    
92    
93  /*************************************************  /*************************************************
94  *      Code parameters and static tables         *  *      Code parameters and static tables         *
# Line 110  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    /* 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 490  static const char error_texts[] = Line 505  static const char error_texts[] =
505    "disallowed Unicode code point (>= 0xd800 && <= 0xdfff)\0"    "disallowed Unicode code point (>= 0xd800 && <= 0xdfff)\0"
506    "invalid UTF-16 string\0"    "invalid UTF-16 string\0"
507    /* 75 */    /* 75 */
508    "name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)\0"    "name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)\0"
509      "character value in \\u.... sequence is too large\0"
510      "invalid UTF-32 string\0"
511    ;    ;
512    
513  /* Table to identify digits and hex digits. This is used when compiling  /* Table to identify digits and hex digits. This is used when compiling
# Line 630  static const pcre_uint8 ebcdic_chartab[] Line 647  static const pcre_uint8 ebcdic_chartab[]
647  #endif  #endif
648    
649    
 /* Definition to allow mutual recursion */  
   
 static BOOL  
   compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int,  
     int *, int *, branch_chain *, compile_data *, int *);  
   
   
650    
651  /*************************************************  /*************************************************
652  *            Find an error text                  *  *            Find an error text                  *
# Line 741  return (*p == CHAR_RIGHT_CURLY_BRACKET); Line 751  return (*p == CHAR_RIGHT_CURLY_BRACKET);
751  *************************************************/  *************************************************/
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 a negative value which  positive value for a simple escape such as \n, or 0 for a data character
755  encodes one of the more complicated things such as \d. 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. On entry,  be returned in chptr.
758  ptr is pointing at the \. On exit, it is on the final character of the escape  On entry,ptr is pointing at the \. On exit, it is on the final character of the
759  sequence.  escape sequence.
760    
761  Arguments:  Arguments:
762    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
763      chptr          points to the data character
764    errorcodeptr   points to the errorcode variable    errorcodeptr   points to the errorcode variable
765    bracount       number of previous extracting brackets    bracount       number of previous extracting brackets
766    options        the options bits    options        the options bits
767    isclass        TRUE if inside a character class    isclass        TRUE if inside a character class
768    
769  Returns:         zero or positive => a data character  Returns:         zero => a data character
770                   negative => 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 *errorcodeptr, int bracount,  check_escape(const pcre_uchar **ptrptr, pcre_uint32 *chptr, int *errorcodeptr,
777    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;
784  int i;  int i;
785    
786  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */
# Line 784  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) c = 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 < '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)  c = 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 797  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 822  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 831  else Line 845  else
845            c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
846  #endif  #endif
847            }            }
848    
849    #if defined COMPILE_PCRE8
850            if (c > (utf ? 0x10ffff : 0xff))
851    #elif defined COMPILE_PCRE16
852            if (c > (utf ? 0x10ffff : 0xffff))
853    #elif defined COMPILE_PCRE32
854            if (utf && c > 0x10ffff)
855    #endif
856              {
857              *errorcodeptr = ERR76;
858              }
859            else if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR73;
860          }          }
861        }        }
862      else      else
# Line 857  else Line 883  else
883      (3) For Oniguruma compatibility we also support \g followed by a name or a      (3) For Oniguruma compatibility we also support \g followed by a name or a
884      number either in angle brackets or in single quotes. However, these are      number either in angle brackets or in single quotes. However, these are
885      (possibly recursive) subroutine calls, _not_ backreferences. Just return      (possibly recursive) subroutine calls, _not_ backreferences. Just return
886      the -ESC_g code (cf \k). */      the ESC_g code (cf \k). */
887    
888      case CHAR_g:      case CHAR_g:
889      if (isclass) break;      if (isclass) break;
890      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)
891        {        {
892        c = -ESC_g;        escape = ESC_g;
893        break;        break;
894        }        }
895    
# Line 876  else Line 902  else
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 != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)
904          {          {
905          c = -ESC_k;          escape = ESC_k;
906          break;          break;
907          }          }
908        braced = TRUE;        braced = TRUE;
# Line 892  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 916  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 924  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      c = -(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 954  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          c = -(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 1020  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 1038  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 1051  else Line 1084  else
1084          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
1085  #endif  #endif
1086    
1087  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1088          if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }          if (c > (utf ? 0x10ffff : 0xff)) { overflow = TRUE; break; }
1089  #else  #elif defined COMPILE_PCRE16
1090  #ifdef COMPILE_PCRE16          if (c > (utf ? 0x10ffff : 0xffff)) { overflow = TRUE; break; }
1091          if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }  #elif defined COMPILE_PCRE32
1092  #endif          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 1082  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 1141  else Line 1174  else
1174  newline". PCRE does not support \N{name}. However, it does support  newline". PCRE does not support \N{name}. However, it does support
1175  quantification such as \N{2,3}. */  quantification such as \N{2,3}. */
1176    
1177  if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&  if (escape == ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&
1178       !is_counted_repeat(ptr+2))       !is_counted_repeat(ptr+2))
1179    *errorcodeptr = ERR37;    *errorcodeptr = ERR37;
1180    
1181  /* If PCRE_UCP is set, we change the values for \d etc. */  /* If PCRE_UCP is set, we change the values for \d etc. */
1182    
1183  if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w)  if ((options & PCRE_UCP) != 0 && escape >= ESC_D && escape <= ESC_w)
1184    c -= (ESC_DU - ESC_D);    escape += (ESC_DU - ESC_D);
1185    
1186  /* Set the pointer to the final character before returning. */  /* Set the pointer to the final character before returning. */
1187    
1188  *ptrptr = ptr;  *ptrptr = ptr;
1189  return c;  *chptr = c;
1190    return escape;
1191  }  }
1192    
   
   
1193  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1194  /*************************************************  /*************************************************
1195  *               Handle \P and \p                 *  *               Handle \P and \p                 *
# Line 1171  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 1227  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 1279  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 1294  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 1349  Arguments: Line 1386  Arguments:
1386    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1387    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1388    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1389    utf          TRUE if we are in UTF-8 / UTF-16 mode    utf          TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
1390    count        pointer to the current capturing subpattern number (updated)    count        pointer to the current capturing subpattern number (updated)
1391    
1392  Returns:       the number of the named subpattern, or -1 if not found  Returns:       the number of the named subpattern, or -1 if not found
# Line 1425  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 1433  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 1583  Arguments: Line 1620  Arguments:
1620    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1621    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1622    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1623    utf          TRUE if we are in UTF-8 / UTF-16 mode    utf          TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
1624    
1625  Returns:       the number of the found subpattern, or -1 if not found  Returns:       the number of the found subpattern, or -1 if not found
1626  */  */
# Line 1686  and doing the check at the end; a flag s Line 1723  and doing the check at the end; a flag s
1723    
1724  Arguments:  Arguments:
1725    code     points to the start of the pattern (the bracket)    code     points to the start of the pattern (the bracket)
1726    utf      TRUE in UTF-8 / UTF-16 mode    utf      TRUE in UTF-8 / UTF-16 / UTF-32 mode
1727    atend    TRUE if called when the pattern is complete    atend    TRUE if called when the pattern is complete
1728    cd       the "compile data" structure    cd       the "compile data" structure
1729    
# Line 1712  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 1832  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  #ifdef SUPPORT_UTF  #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]);
# Line 1841  for (;;) Line 1878  for (;;)
1878    
1879      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1880      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1881      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP) cc += 2;      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP)
1882          cc += 2;
1883      cc += 1 + IMM2_SIZE + 1;      cc += 1 + IMM2_SIZE + 1;
1884      break;      break;
1885    
# Line 1876  for (;;) Line 1914  for (;;)
1914    
1915      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1916    
1917  #if defined SUPPORT_UTF || defined COMPILE_PCRE16  #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32
1918      case OP_XCLASS:      case OP_XCLASS:
1919      cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];      cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];
1920      /* Fall through */      /* Fall through */
# Line 1899  for (;;) Line 1937  for (;;)
1937        case OP_CRRANGE:        case OP_CRRANGE:
1938        case OP_CRMINRANGE:        case OP_CRMINRANGE:
1939        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;
1940        branchlength += GET2(cc,1);        branchlength += (int)GET2(cc,1);
1941        cc += 1 + 2 * IMM2_SIZE;        cc += 1 + 2 * IMM2_SIZE;
1942        break;        break;
1943    
# Line 2015  length. Line 2053  length.
2053    
2054  Arguments:  Arguments:
2055    code        points to start of expression    code        points to start of expression
2056    utf         TRUE in UTF-8 / UTF-16 mode    utf         TRUE in UTF-8 / UTF-16 / UTF-32 mode
2057    number      the required bracket number or negative to find a lookbehind    number      the required bracket number or negative to find a lookbehind
2058    
2059  Returns:      pointer to the opcode for the bracket, or NULL if not found  Returns:      pointer to the opcode for the bracket, or NULL if not found
# Line 2026  PRIV(find_bracket)(const pcre_uchar *cod Line 2064  PRIV(find_bracket)(const pcre_uchar *cod
2064  {  {
2065  for (;;)  for (;;)
2066    {    {
2067    register int c = *code;    register pcre_uchar c = *code;
2068    
2069    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2070    
# Line 2049  for (;;) Line 2087  for (;;)
2087    else if (c == OP_CBRA || c == OP_SCBRA ||    else if (c == OP_CBRA || c == OP_SCBRA ||
2088             c == OP_CBRAPOS || c == OP_SCBRAPOS)             c == OP_CBRAPOS || c == OP_SCBRAPOS)
2089      {      {
2090      int n = GET2(code, 1+LINK_SIZE);      int n = (int)GET2(code, 1+LINK_SIZE);
2091      if (n == number) return (pcre_uchar *)code;      if (n == number) return (pcre_uchar *)code;
2092      code += PRIV(OP_lengths)[c];      code += PRIV(OP_lengths)[c];
2093      }      }
# Line 2079  for (;;) Line 2117  for (;;)
2117        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2118        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2119        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
2120        if (code[1 + IMM2_SIZE] == OP_PROP        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2121          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;          code += 2;
2122        break;        break;
2123    
2124        case OP_MARK:        case OP_MARK:
# Line 2102  for (;;) Line 2140  for (;;)
2140    a multi-byte character. The length in the table is a minimum, so we have to    a multi-byte character. The length in the table is a minimum, so we have to
2141    arrange to skip the extra bytes. */    arrange to skip the extra bytes. */
2142    
2143  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2144      if (utf) switch(c)      if (utf) switch(c)
2145        {        {
2146        case OP_CHAR:        case OP_CHAR:
# Line 2154  instance of OP_RECURSE. Line 2192  instance of OP_RECURSE.
2192    
2193  Arguments:  Arguments:
2194    code        points to start of expression    code        points to start of expression
2195    utf         TRUE in UTF-8 / UTF-16 mode    utf         TRUE in UTF-8 / UTF-16 / UTF-32 mode
2196    
2197  Returns:      pointer to the opcode for OP_RECURSE, or NULL if not found  Returns:      pointer to the opcode for OP_RECURSE, or NULL if not found
2198  */  */
# Line 2164  find_recurse(const pcre_uchar *code, BOO Line 2202  find_recurse(const pcre_uchar *code, BOO
2202  {  {
2203  for (;;)  for (;;)
2204    {    {
2205    register int c = *code;    register pcre_uchar c = *code;
2206    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2207    if (c == OP_RECURSE) return code;    if (c == OP_RECURSE) return code;
2208    
# Line 2199  for (;;) Line 2237  for (;;)
2237        case OP_TYPEUPTO:        case OP_TYPEUPTO:
2238        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2239        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2240        if (code[1 + IMM2_SIZE] == OP_PROP        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2241          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;          code += 2;
2242        break;        break;
2243    
2244        case OP_MARK:        case OP_MARK:
# Line 2222  for (;;) Line 2260  for (;;)
2260      by a multi-byte character. The length in the table is a minimum, so we have      by a multi-byte character. The length in the table is a minimum, so we have
2261      to arrange to skip the extra bytes. */      to arrange to skip the extra bytes. */
2262    
2263  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2264      if (utf) switch(c)      if (utf) switch(c)
2265        {        {
2266        case OP_CHAR:        case OP_CHAR:
# Line 2308  bracket whose current branch will alread Line 2346  bracket whose current branch will alread
2346  Arguments:  Arguments:
2347    code        points to start of search    code        points to start of search
2348    endcode     points to where to stop    endcode     points to where to stop
2349    utf         TRUE if in UTF-8 / UTF-16 mode    utf         TRUE if in UTF-8 / UTF-16 / UTF-32 mode
2350    cd          contains pointers to tables etc.    cd          contains pointers to tables etc.
2351    
2352  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2318  static BOOL Line 2356  static BOOL
2356  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,
2357    BOOL utf, compile_data *cd)    BOOL utf, compile_data *cd)
2358  {  {
2359  register int c;  register pcre_uchar c;
2360  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
2361       code < endcode;       code < endcode;
2362       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))
# Line 2352  for (code = first_significant_code(code Line 2390  for (code = first_significant_code(code
2390      /* Test for forward reference */      /* Test for forward reference */
2391    
2392      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)
2393        if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE;        if ((int)GET(scode, 0) == (int)(code + 1 - cd->start_code)) return TRUE;
2394    
2395      /* Not a forward reference, test for completed backward reference */      /* Not a forward reference, test for completed backward reference */
2396    
# Line 2525  for (code = first_significant_code(code Line 2563  for (code = first_significant_code(code
2563      case OP_TYPEUPTO:      case OP_TYPEUPTO:
2564      case OP_TYPEMINUPTO:      case OP_TYPEMINUPTO:
2565      case OP_TYPEPOSUPTO:      case OP_TYPEPOSUPTO:
2566      if (code[1 + IMM2_SIZE] == OP_PROP      if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2567        || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;        code += 2;
2568      break;      break;
2569    
2570      /* End of branch */      /* End of branch */
# Line 2541  for (code = first_significant_code(code Line 2579  for (code = first_significant_code(code
2579      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
2580      MINUPTO, and POSUPTO may be followed by a multibyte character */      MINUPTO, and POSUPTO may be followed by a multibyte character */
2581    
2582  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2583      case OP_STAR:      case OP_STAR:
2584      case OP_STARI:      case OP_STARI:
2585      case OP_MINSTAR:      case OP_MINSTAR:
# Line 2607  Arguments: Line 2645  Arguments:
2645    code        points to start of the recursion    code        points to start of the recursion
2646    endcode     points to where to stop (current RECURSE item)    endcode     points to where to stop (current RECURSE item)
2647    bcptr       points to the chain of current (unclosed) branch starts    bcptr       points to the chain of current (unclosed) branch starts
2648    utf         TRUE if in UTF-8 / UTF-16 mode    utf         TRUE if in UTF-8 / UTF-16 / UTF-32 mode
2649    cd          pointers to tables etc    cd          pointers to tables etc
2650    
2651  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2673  Returns:   TRUE or FALSE Line 2711  Returns:   TRUE or FALSE
2711  static BOOL  static BOOL
2712  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)
2713  {  {
2714  int terminator;          /* Don't combine these lines; the Solaris cc */  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */
2715  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
2716  for (++ptr; *ptr != 0; ptr++)  for (++ptr; *ptr != 0; ptr++)
2717    {    {
# Line 2722  register int yield = 0; Line 2760  register int yield = 0;
2760  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
2761    {    {
2762    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
2763      STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;      STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield;
2764    pn += posix_name_lengths[yield] + 1;    pn += posix_name_lengths[yield] + 1;
2765    yield++;    yield++;
2766    }    }
# Line 2754  value in the reference (which is a group Line 2792  value in the reference (which is a group
2792  Arguments:  Arguments:
2793    group      points to the start of the group    group      points to the start of the group
2794    adjust     the amount by which the group is to be moved    adjust     the amount by which the group is to be moved
2795    utf        TRUE in UTF-8 / UTF-16 mode    utf        TRUE in UTF-8 / UTF-16 / UTF-32 mode
2796    cd         contains pointers to tables etc.    cd         contains pointers to tables etc.
2797    save_hwm   the hwm forward reference pointer at the start of the group    save_hwm   the hwm forward reference pointer at the start of the group
2798    
# Line 2777  while ((ptr = (pcre_uchar *)find_recurse Line 2815  while ((ptr = (pcre_uchar *)find_recurse
2815    
2816    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
2817      {      {
2818      offset = GET(hc, 0);      offset = (int)GET(hc, 0);
2819      if (cd->start_code + offset == ptr + 1)      if (cd->start_code + offset == ptr + 1)
2820        {        {
2821        PUT(hc, 0, offset + adjust);        PUT(hc, 0, offset + adjust);
# Line 2790  while ((ptr = (pcre_uchar *)find_recurse Line 2828  while ((ptr = (pcre_uchar *)find_recurse
2828    
2829    if (hc >= cd->hwm)    if (hc >= cd->hwm)
2830      {      {
2831      offset = GET(ptr, 1);      offset = (int)GET(ptr, 1);
2832      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
2833      }      }
2834    
# Line 2858  PUT(previous_callout, 2 + LINK_SIZE, len Line 2896  PUT(previous_callout, 2 + LINK_SIZE, len
2896  *************************************************/  *************************************************/
2897    
2898  /* This function is passed the start and end of a class range, in UTF-8 mode  /* This function is passed the start and end of a class range, in UTF-8 mode
2899  with UCP support. It searches up the characters, looking for internal ranges of  with UCP support. It searches up the characters, looking for ranges of
2900  characters in the "other" case. Each call returns the next one, updating the  characters in the "other" case. Each call returns the next one, updating the
2901  start address.  start address. A character with multiple other cases is returned on its own
2902    with a special return value.
2903    
2904  Arguments:  Arguments:
2905    cptr        points to starting character value; updated    cptr        points to starting character value; updated
# Line 2868  Arguments: Line 2907  Arguments:
2907    ocptr       where to put start of othercase range    ocptr       where to put start of othercase range
2908    odptr       where to put end of othercase range    odptr       where to put end of othercase range
2909    
2910  Yield:        TRUE when range returned; FALSE when no more  Yield:        -1 when no more
2911                   0 when a range is returned
2912                  >0 the CASESET offset for char with multiple other cases
2913                    in this case, ocptr contains the original
2914  */  */
2915    
2916  static BOOL  static int
2917  get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr,  get_othercase_range(pcre_uint32 *cptr, pcre_uint32 d, pcre_uint32 *ocptr,
2918    unsigned int *odptr)    pcre_uint32 *odptr)
2919  {  {
2920  unsigned int c, othercase, next;  pcre_uint32 c, othercase, next;
2921    unsigned int co;
2922    
2923    /* Find the first character that has an other case. If it has multiple other
2924    cases, return its case offset value. */
2925    
2926  for (c = *cptr; c <= d; c++)  for (c = *cptr; c <= d; c++)
2927    { if ((othercase = UCD_OTHERCASE(c)) != c) break; }    {
2928      if ((co = UCD_CASESET(c)) != 0)
2929        {
2930        *ocptr = c++;   /* Character that has the set */
2931        *cptr = c;      /* Rest of input range */
2932        return (int)co;
2933        }
2934      if ((othercase = UCD_OTHERCASE(c)) != c) break;
2935      }
2936    
2937  if (c > d) return FALSE;  if (c > d) return -1;  /* Reached end of range */
2938    
2939  *ocptr = othercase;  *ocptr = othercase;
2940  next = othercase + 1;  next = othercase + 1;
# Line 2891  for (++c; c <= d; c++) Line 2945  for (++c; c <= d; c++)
2945    next++;    next++;
2946    }    }
2947    
2948  *odptr = next - 1;  *odptr = next - 1;     /* End of othercase range */
2949  *cptr = c;  *cptr = c;             /* Rest of input range */
2950    return 0;
 return TRUE;  
2951  }  }
2952    
2953    
# Line 2916  Returns:       TRUE if auto-possessifyin Line 2969  Returns:       TRUE if auto-possessifyin
2969  */  */
2970    
2971  static BOOL  static BOOL
2972  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)
2973  {  {
2974    #ifdef SUPPORT_UCP
2975    const pcre_uint32 *p;
2976    #endif
2977    
2978  const ucd_record *prop = GET_UCD(c);  const ucd_record *prop = GET_UCD(c);
2979    
2980  switch(ptype)  switch(ptype)
2981    {    {
2982    case PT_LAMP:    case PT_LAMP:
# Line 2956  switch(ptype) Line 3014  switch(ptype)
3014    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
3015            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
3016            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
3017    
3018    #ifdef SUPPORT_UCP
3019      case PT_CLIST:
3020      p = PRIV(ucd_caseless_sets) + prop->caseset;
3021      for (;;)
3022        {
3023        if ((unsigned int)c < *p) return !negated;
3024        if ((unsigned int)c == *p++) return negated;
3025        }
3026      break;  /* Control never reaches here */
3027    #endif
3028    }    }
3029    
3030  return FALSE;  return FALSE;
3031  }  }
3032  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
# Line 2973  sense to automatically possessify the re Line 3043  sense to automatically possessify the re
3043    
3044  Arguments:  Arguments:
3045    previous      pointer to the repeated opcode    previous      pointer to the repeated opcode
3046    utf           TRUE in UTF-8 / UTF-16 mode    utf           TRUE in UTF-8 / UTF-16 / UTF-32 mode
3047    ptr           next character in pattern    ptr           next character in pattern
3048    options       options bits    options       options bits
3049    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
# Line 2985  static BOOL Line 3055  static BOOL
3055  check_auto_possessive(const pcre_uchar *previous, BOOL utf,  check_auto_possessive(const pcre_uchar *previous, BOOL utf,
3056    const pcre_uchar *ptr, int options, compile_data *cd)    const pcre_uchar *ptr, int options, compile_data *cd)
3057  {  {
3058  pcre_int32 c, next;  pcre_uint32 c = NOTACHAR;
3059  int op_code = *previous++;  pcre_uint32 next;
3060    int escape;
3061    pcre_uchar op_code = *previous++;
3062    
3063  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
3064    
# Line 3017  value is a character, a negative value i Line 3089  value is a character, a negative value i
3089  if (*ptr == CHAR_BACKSLASH)  if (*ptr == CHAR_BACKSLASH)
3090    {    {
3091    int temperrorcode = 0;    int temperrorcode = 0;
3092    next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE);    escape = check_escape(&ptr, &next, &temperrorcode, cd->bracount, options, FALSE);
3093    if (temperrorcode != 0) return FALSE;    if (temperrorcode != 0) return FALSE;
3094    ptr++;    /* Point after the escape sequence */    ptr++;    /* Point after the escape sequence */
3095    }    }
3096  else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)  else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)
3097    {    {
3098      escape = 0;
3099  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3100    if (utf) { GETCHARINC(next, ptr); } else    if (utf) { GETCHARINC(next, ptr); } else
3101  #endif  #endif
# Line 3059  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3132  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3132    STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)    STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
3133      return FALSE;      return FALSE;
3134    
3135  /* Now compare the next item with the previous opcode. First, handle cases when  /* If the previous item is a character, get its value. */
 the next item is a character. */  
3136    
3137  if (next >= 0) switch(op_code)  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3138        op_code == OP_NOT || op_code == OP_NOTI)
3139    {    {
   case OP_CHAR:  
3140  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3141    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
3142  #else  #else
3143    c = *previous;    c = *previous;
3144  #endif  #endif
3145    return c != next;    }
3146    
3147    /* For CHARI (caseless character) we must check the other case. If we have  /* Now compare the next item with the previous opcode. First, handle cases when
3148    Unicode property support, we can use it to test the other case of  the next item is a character. */
   high-valued characters. */  
3149    
3150    case OP_CHARI:  if (escape == 0)
3151  #ifdef SUPPORT_UTF    {
3152    GETCHARTEST(c, previous);    /* For a caseless UTF match, the next character may have more than one other
3153  #else    case, which maps to the special PT_CLIST property. Check this first. */
3154    c = *previous;  
 #endif  
   if (c == next) return FALSE;  
 #ifdef SUPPORT_UTF  
   if (utf)  
     {  
     unsigned int othercase;  
     if (next < 128) othercase = cd->fcc[next]; else  
3155  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3156      othercase = UCD_OTHERCASE((unsigned int)next);    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3157  #else      {
3158      othercase = NOTACHAR;      unsigned int ocs = UCD_CASESET(next);
3159  #endif      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);
     return (unsigned int)c != othercase;  
3160      }      }
   else  
 #endif  /* SUPPORT_UTF */  
   return (c != TABLE_GET((unsigned int)next, cd->fcc, next));  /* Non-UTF-8 mode */  
   
   case OP_NOT:  
 #ifdef SUPPORT_UTF  
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
3161  #endif  #endif
   return c == next;  
3162    
3163    case OP_NOTI:    switch(op_code)
3164        {
3165        case OP_CHAR:
3166        return c != next;
3167    
3168        /* For CHARI (caseless character) we must check the other case. If we have
3169        Unicode property support, we can use it to test the other case of
3170        high-valued characters. We know that next can have only one other case,
3171        because multi-other-case characters are dealt with above. */
3172    
3173        case OP_CHARI:
3174        if (c == next) return FALSE;
3175  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3176    GETCHARTEST(c, previous);      if (utf)
3177          {
3178          pcre_uint32 othercase;
3179          if (next < 128) othercase = cd->fcc[next]; else
3180    #ifdef SUPPORT_UCP
3181          othercase = UCD_OTHERCASE(next);
3182  #else  #else
3183    c = *previous;        othercase = NOTACHAR;
3184  #endif  #endif
3185    if (c == next) return TRUE;        return c != othercase;
3186          }
3187        else
3188    #endif  /* SUPPORT_UTF */
3189        return (c != TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3190    
3191        case OP_NOT:
3192        return c == next;
3193    
3194        case OP_NOTI:
3195        if (c == next) return TRUE;
3196  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3197    if (utf)      if (utf)
3198      {        {
3199      unsigned int othercase;        pcre_uint32 othercase;
3200      if (next < 128) othercase = cd->fcc[next]; else        if (next < 128) othercase = cd->fcc[next]; else
3201  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3202      othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE(next);
3203  #else  #else
3204      othercase = NOTACHAR;        othercase = NOTACHAR;
3205  #endif  #endif
3206      return (unsigned int)c == othercase;        return c == othercase;
3207      }        }
3208    else      else
3209  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3210    return (c == TABLE_GET((unsigned int)next, cd->fcc, next));  /* Non-UTF-8 mode */      return (c == TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3211    
3212    /* 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.
3213    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. */
3214    
3215    case OP_DIGIT:      case OP_DIGIT:
3216    return next > 255 || (cd->ctypes[next] & ctype_digit) == 0;      return next > 255 || (cd->ctypes[next] & ctype_digit) == 0;
3217    
3218    case OP_NOT_DIGIT:      case OP_NOT_DIGIT:
3219    return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;
3220    
3221    case OP_WHITESPACE:      case OP_WHITESPACE:
3222    return next > 255 || (cd->ctypes[next] & ctype_space) == 0;      return next > 255 || (cd->ctypes[next] & ctype_space) == 0;
3223    
3224    case OP_NOT_WHITESPACE:      case OP_NOT_WHITESPACE:
3225    return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;
3226    
3227    case OP_WORDCHAR:      case OP_WORDCHAR:
3228    return next > 255 || (cd->ctypes[next] & ctype_word) == 0;      return next > 255 || (cd->ctypes[next] & ctype_word) == 0;
3229    
3230    case OP_NOT_WORDCHAR:      case OP_NOT_WORDCHAR:
3231    return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;
3232    
3233    case OP_HSPACE:      case OP_HSPACE:
3234    case OP_NOT_HSPACE:      case OP_NOT_HSPACE:
3235    switch(next)      switch(next)
3236      {        {
3237      case 0x09:        HSPACE_CASES:
3238      case 0x20:        return op_code == OP_NOT_HSPACE;
     case 0xa0:  
     case 0x1680:  
     case 0x180e:  
     case 0x2000:  
     case 0x2001:  
     case 0x2002:  
     case 0x2003:  
     case 0x2004:  
     case 0x2005:  
     case 0x2006:  
     case 0x2007:  
     case 0x2008:  
     case 0x2009:  
     case 0x200A:  
     case 0x202f:  
     case 0x205f:  
     case 0x3000:  
     return op_code == OP_NOT_HSPACE;  
     default:  
     return op_code != OP_NOT_HSPACE;  
     }  
3239    
3240    case OP_ANYNL:        default:
3241    case OP_VSPACE:        return op_code != OP_NOT_HSPACE;
3242    case OP_NOT_VSPACE:        }
3243    switch(next)  
3244      {      case OP_ANYNL:
3245      case 0x0a:      case OP_VSPACE:
3246      case 0x0b:      case OP_NOT_VSPACE:
3247      case 0x0c:      switch(next)
3248      case 0x0d:        {
3249      case 0x85:        VSPACE_CASES:
3250      case 0x2028:        return op_code == OP_NOT_VSPACE;
3251      case 0x2029:  
3252      return op_code == OP_NOT_VSPACE;        default:
3253      default:        return op_code != OP_NOT_VSPACE;
3254      return op_code != OP_NOT_VSPACE;        }
     }  
3255    
3256  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3257    case OP_PROP:      case OP_PROP:
3258    return check_char_prop(next, previous[0], previous[1], FALSE);      return check_char_prop(next, previous[0], previous[1], FALSE);
3259    
3260    case OP_NOTPROP:      case OP_NOTPROP:
3261    return check_char_prop(next, previous[0], previous[1], TRUE);      return check_char_prop(next, previous[0], previous[1], TRUE);
3262  #endif  #endif
3263    
3264    default:      default:
3265    return FALSE;      return FALSE;
3266        }
3267    }    }
3268    
   
3269  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
3270  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
3271  generated only when PCRE_UCP is *not* set, that is, when only ASCII  generated only when PCRE_UCP is *not* set, that is, when only ASCII
# Line 3219  switch(op_code) Line 3276  switch(op_code)
3276    {    {
3277    case OP_CHAR:    case OP_CHAR:
3278    case OP_CHARI:    case OP_CHARI:
3279  #ifdef SUPPORT_UTF    switch(escape)
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
 #endif  
   switch(-next)  
3280      {      {
3281      case ESC_d:      case ESC_d:
3282      return c > 255 || (cd->ctypes[c] & ctype_digit) == 0;      return c > 255 || (cd->ctypes[c] & ctype_digit) == 0;
# Line 3248  switch(op_code) Line 3300  switch(op_code)
3300      case ESC_H:      case ESC_H:
3301      switch(c)      switch(c)
3302        {        {
3303        case 0x09:        HSPACE_CASES:
3304        case 0x20:        return escape != ESC_h;
3305        case 0xa0:  
       case 0x1680:  
       case 0x180e:  
       case 0x2000:  
       case 0x2001:  
       case 0x2002:  
       case 0x2003:  
       case 0x2004:  
       case 0x2005:  
       case 0x2006:  
       case 0x2007:  
       case 0x2008:  
       case 0x2009:  
       case 0x200A:  
       case 0x202f:  
       case 0x205f:  
       case 0x3000:  
       return -next != ESC_h;  
3306        default:        default:
3307        return -next == ESC_h;        return escape == ESC_h;
3308        }        }
3309    
3310      case ESC_v:      case ESC_v:
3311      case ESC_V:      case ESC_V:
3312      switch(c)      switch(c)
3313        {        {
3314        case 0x0a:        VSPACE_CASES:
3315        case 0x0b:        return escape != ESC_v;
3316        case 0x0c:  
       case 0x0d:  
       case 0x85:  
       case 0x2028:  
       case 0x2029:  
       return -next != ESC_v;  
3317        default:        default:
3318        return -next == ESC_v;        return escape == ESC_v;
3319        }        }
3320    
3321      /* When PCRE_UCP is set, these values get generated for \d etc. Find      /* When PCRE_UCP is set, these values get generated for \d etc. Find
3322      their substitutions and process them. The result will always be either      their substitutions and process them. The result will always be either
3323      -ESC_p or -ESC_P. Then fall through to process those values. */      ESC_p or ESC_P. Then fall through to process those values. */
3324    
3325  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3326      case ESC_du:      case ESC_du:
# Line 3301  switch(op_code) Line 3331  switch(op_code)
3331      case ESC_SU:      case ESC_SU:
3332        {        {
3333        int temperrorcode = 0;        int temperrorcode = 0;
3334        ptr = substitutes[-next - ESC_DU];        ptr = substitutes[escape - ESC_DU];
3335        next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);        escape = check_escape(&ptr, &next, &temperrorcode, 0, options, FALSE);
3336        if (temperrorcode != 0) return FALSE;        if (temperrorcode != 0) return FALSE;
3337        ptr++;    /* For compatibility */        ptr++;    /* For compatibility */
3338        }        }
# Line 3311  switch(op_code) Line 3341  switch(op_code)
3341      case ESC_p:      case ESC_p:
3342      case ESC_P:      case ESC_P:
3343        {        {
3344        int ptype, pdata, errorcodeptr;        unsigned int ptype = 0, pdata = 0;
3345          int errorcodeptr;
3346        BOOL negated;        BOOL negated;
3347    
3348        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
3349        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
3350        if (ptype < 0) return FALSE;          return FALSE;
3351        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
3352    
3353        /* 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 3329  switch(op_code) Line 3360  switch(op_code)
3360    
3361        /* Do the property check. */        /* Do the property check. */
3362    
3363        return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);        return check_char_prop(c, ptype, pdata, (escape == ESC_P) != negated);
3364        }        }
3365  #endif  #endif
3366    
# Line 3344  switch(op_code) Line 3375  switch(op_code)
3375    these op-codes are never generated.) */    these op-codes are never generated.) */
3376    
3377    case OP_DIGIT:    case OP_DIGIT:
3378    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||    return escape == ESC_D || escape == ESC_s || escape == ESC_W ||
3379           next == -ESC_h || next == -ESC_v || next == -ESC_R;           escape == ESC_h || escape == ESC_v || escape == ESC_R;
3380    
3381    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
3382    return next == -ESC_d;    return escape == ESC_d;
3383    
3384    case OP_WHITESPACE:    case OP_WHITESPACE:
3385    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return escape == ESC_S || escape == ESC_d || escape == ESC_w;
3386    
3387    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
3388    return next == -ESC_s || next == -ESC_h || next == -ESC_v || next == -ESC_R;    return escape == ESC_s || escape == ESC_h || escape == ESC_v || escape == ESC_R;
3389    
3390    case OP_HSPACE:    case OP_HSPACE:
3391    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||    return escape == ESC_S || escape == ESC_H || escape == ESC_d ||
3392           next == -ESC_w || next == -ESC_v || next == -ESC_R;           escape == ESC_w || escape == ESC_v || escape == ESC_R;
3393    
3394    case OP_NOT_HSPACE:    case OP_NOT_HSPACE:
3395    return next == -ESC_h;    return escape == ESC_h;
3396    
3397    /* Can't have \S in here because VT matches \S (Perl anomaly) */    /* Can't have \S in here because VT matches \S (Perl anomaly) */
3398    case OP_ANYNL:    case OP_ANYNL:
3399    case OP_VSPACE:    case OP_VSPACE:
3400    return next == -ESC_V || next == -ESC_d || next == -ESC_w;    return escape == ESC_V || escape == ESC_d || escape == ESC_w;
3401    
3402    case OP_NOT_VSPACE:    case OP_NOT_VSPACE:
3403    return next == -ESC_v || next == -ESC_R;    return escape == ESC_v || escape == ESC_R;
3404    
3405    case OP_WORDCHAR:    case OP_WORDCHAR:
3406    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||    return escape == ESC_W || escape == ESC_s || escape == ESC_h ||
3407           next == -ESC_v || next == -ESC_R;           escape == ESC_v || escape == ESC_R;
3408    
3409    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
3410    return next == -ESC_w || next == -ESC_d;    return escape == ESC_w || escape == ESC_d;
3411    
3412    default:    default:
3413    return FALSE;    return FALSE;
# Line 3388  switch(op_code) Line 3419  switch(op_code)
3419    
3420    
3421  /*************************************************  /*************************************************
3422    *        Add a character or range to a class     *
3423    *************************************************/
3424    
3425    /* This function packages up the logic of adding a character or range of
3426    characters to a class. The character values in the arguments will be within the
3427    valid values for the current mode (8-bit, 16-bit, UTF, etc). This function is
3428    mutually recursive with the function immediately below.
3429    
3430    Arguments:
3431      classbits     the bit map for characters < 256
3432      uchardptr     points to the pointer for extra data
3433      options       the options word
3434      cd            contains pointers to tables etc.
3435      start         start of range character
3436      end           end of range character
3437    
3438    Returns:        the number of < 256 characters added
3439                    the pointer to extra data is updated
3440    */
3441    
3442    static int
3443    add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,
3444      compile_data *cd, pcre_uint32 start, pcre_uint32 end)
3445    {
3446    pcre_uint32 c;
3447    int n8 = 0;
3448    
3449    /* If caseless matching is required, scan the range and process alternate
3450    cases. In Unicode, there are 8-bit characters that have alternate cases that
3451    are greater than 255 and vice-versa. Sometimes we can just extend the original
3452    range. */
3453    
3454    if ((options & PCRE_CASELESS) != 0)
3455      {
3456    #ifdef SUPPORT_UCP
3457      if ((options & PCRE_UTF8) != 0)
3458        {
3459        int rc;
3460        pcre_uint32 oc, od;
3461    
3462        options &= ~PCRE_CASELESS;   /* Remove for recursive calls */
3463        c = start;
3464    
3465        while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)
3466          {
3467          /* Handle a single character that has more than one other case. */
3468    
3469          if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,
3470            PRIV(ucd_caseless_sets) + rc, oc);
3471    
3472          /* Do nothing if the other case range is within the original range. */
3473    
3474          else if (oc >= start && od <= end) continue;
3475    
3476          /* Extend the original range if there is overlap, noting that if oc < c, we
3477          can't have od > end because a subrange is always shorter than the basic
3478          range. Otherwise, use a recursive call to add the additional range. */
3479    
3480          else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */
3481          else if (od > end && oc <= end + 1) end = od;       /* Extend upwards */
3482          else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);
3483          }
3484        }
3485      else
3486    #endif  /* SUPPORT_UCP */
3487    
3488      /* Not UTF-mode, or no UCP */
3489    
3490      for (c = start; c <= end && c < 256; c++)
3491        {
3492        SETBIT(classbits, cd->fcc[c]);
3493        n8++;
3494        }
3495      }
3496    
3497    /* Now handle the original range. Adjust the final value according to the bit
3498    length - this means that the same lists of (e.g.) horizontal spaces can be used
3499    in all cases. */
3500    
3501    #if defined COMPILE_PCRE8
3502    #ifdef SUPPORT_UTF
3503      if ((options & PCRE_UTF8) == 0)
3504    #endif
3505      if (end > 0xff) end = 0xff;
3506    
3507    #elif defined COMPILE_PCRE16
3508    #ifdef SUPPORT_UTF
3509      if ((options & PCRE_UTF16) == 0)
3510    #endif
3511      if (end > 0xffff) end = 0xffff;
3512    
3513    #endif /* COMPILE_PCRE[8|16] */
3514    
3515    /* If all characters are less than 256, use the bit map. Otherwise use extra
3516    data. */
3517    
3518    if (end < 0x100)
3519      {
3520      for (c = start; c <= end; c++)
3521        {
3522        n8++;
3523        SETBIT(classbits, c);
3524        }
3525      }
3526    
3527    else
3528      {
3529      pcre_uchar *uchardata = *uchardptr;
3530    
3531    #ifdef SUPPORT_UTF
3532      if ((options & PCRE_UTF8) != 0)  /* All UTFs use the same flag bit */
3533        {
3534        if (start < end)
3535          {
3536          *uchardata++ = XCL_RANGE;
3537          uchardata += PRIV(ord2utf)(start, uchardata);
3538          uchardata += PRIV(ord2utf)(end, uchardata);
3539          }
3540        else if (start == end)
3541          {
3542          *uchardata++ = XCL_SINGLE;
3543          uchardata += PRIV(ord2utf)(start, uchardata);
3544          }
3545        }
3546      else
3547    #endif  /* SUPPORT_UTF */
3548    
3549      /* Without UTF support, character values are constrained by the bit length,
3550      and can only be > 256 for 16-bit and 32-bit libraries. */
3551    
3552    #ifdef COMPILE_PCRE8
3553        {}
3554    #else
3555      if (start < end)
3556        {
3557        *uchardata++ = XCL_RANGE;
3558        *uchardata++ = start;
3559        *uchardata++ = end;
3560        }
3561      else if (start == end)
3562        {
3563        *uchardata++ = XCL_SINGLE;
3564        *uchardata++ = start;
3565        }
3566    #endif
3567    
3568      *uchardptr = uchardata;   /* Updata extra data pointer */
3569      }
3570    
3571    return n8;    /* Number of 8-bit characters */
3572    }
3573    
3574    
3575    
3576    
3577    /*************************************************
3578    *        Add a list of characters to a class     *
3579    *************************************************/
3580    
3581    /* This function is used for adding a list of case-equivalent characters to a
3582    class, and also for adding a list of horizontal or vertical whitespace. If the
3583    list is in order (which it should be), ranges of characters are detected and
3584    handled appropriately. This function is mutually recursive with the function
3585    above.
3586    
3587    Arguments:
3588      classbits     the bit map for characters < 256
3589      uchardptr     points to the pointer for extra data
3590      options       the options word
3591      cd            contains pointers to tables etc.
3592      p             points to row of 32-bit values, terminated by NOTACHAR
3593      except        character to omit; this is used when adding lists of
3594                      case-equivalent characters to avoid including the one we
3595                      already know about
3596    
3597    Returns:        the number of < 256 characters added
3598                    the pointer to extra data is updated
3599    */
3600    
3601    static int
3602    add_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,
3603      compile_data *cd, const pcre_uint32 *p, unsigned int except)
3604    {
3605    int n8 = 0;
3606    while (p[0] < NOTACHAR)
3607      {
3608      int n = 0;
3609      if (p[0] != except)
3610        {
3611        while(p[n+1] == p[0] + n + 1) n++;
3612        n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);
3613        }
3614      p += n + 1;
3615      }
3616    return n8;
3617    }
3618    
3619    
3620    
3621    /*************************************************
3622    *    Add characters not in a list to a class     *
3623    *************************************************/
3624    
3625    /* This function is used for adding the complement of a list of horizontal or
3626    vertical whitespace to a class. The list must be in order.
3627    
3628    Arguments:
3629      classbits     the bit map for characters < 256
3630      uchardptr     points to the pointer for extra data
3631      options       the options word
3632      cd            contains pointers to tables etc.
3633      p             points to row of 32-bit values, terminated by NOTACHAR
3634    
3635    Returns:        the number of < 256 characters added
3636                    the pointer to extra data is updated
3637    */
3638    
3639    static int
3640    add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,
3641      int options, compile_data *cd, const pcre_uint32 *p)
3642    {
3643    BOOL utf = (options & PCRE_UTF8) != 0;
3644    int n8 = 0;
3645    if (p[0] > 0)
3646      n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);
3647    while (p[0] < NOTACHAR)
3648      {
3649      while (p[1] == p[0] + 1) p++;
3650      n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,
3651        (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);
3652      p++;
3653      }
3654    return n8;
3655    }
3656    
3657    
3658    
3659    /*************************************************
3660  *           Compile one branch                   *  *           Compile one branch                   *
3661  *************************************************/  *************************************************/
3662    
# Line 3402  Arguments: Line 3671  Arguments:
3671    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3672    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3673    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3674    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr    place to put the first required character
3675    reqcharptr     set to the last literal character required, else < 0    firstcharflagsptr place to put the first character flags, or a negative number
3676      reqcharptr     place to put the last required character
3677      reqcharflagsptr place to put the last required character flags, or a negative number
3678    bcptr          points to current branch chain    bcptr          points to current branch chain
3679    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3680    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3416  Returns:         TRUE on success Line 3687  Returns:         TRUE on success
3687    
3688  static BOOL  static BOOL
3689  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3690    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,    const pcre_uchar **ptrptr, int *errorcodeptr,
3691    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,    pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
3692      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
3693      branch_chain *bcptr, int cond_depth,
3694    compile_data *cd, int *lengthptr)    compile_data *cd, int *lengthptr)
3695  {  {
3696  int repeat_type, op_type;  int repeat_type, op_type;
3697  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3698  int bravalue = 0;  int bravalue = 0;
3699  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3700  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
3701  pcre_int32 zeroreqchar, zerofirstchar;  pcre_int32 firstcharflags, reqcharflags;
3702    pcre_uint32 zeroreqchar, zerofirstchar;
3703    pcre_int32 zeroreqcharflags, zerofirstcharflags;
3704  pcre_int32 req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3705  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3706  int after_manual_callout = 0;  int after_manual_callout = 0;
3707  int length_prevgroup = 0;  int length_prevgroup = 0;
3708  register int c;  register pcre_uint32 c;
3709    int escape;
3710  register pcre_uchar *code = *codeptr;  register pcre_uchar *code = *codeptr;
3711  pcre_uchar *last_code = code;  pcre_uchar *last_code = code;
3712  pcre_uchar *orig_code = code;  pcre_uchar *orig_code = code;
# Line 3450  must not do this for other options (e.g. Line 3726  must not do this for other options (e.g.
3726  dynamically as we process the pattern. */  dynamically as we process the pattern. */
3727    
3728  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3729  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3730  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3731    #ifndef COMPILE_PCRE32
3732  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3733    #endif
3734  #else  #else
3735  BOOL utf = FALSE;  BOOL utf = FALSE;
3736  #endif  #endif
3737    
3738  /* Helper variables for OP_XCLASS opcode (for characters > 255). */  /* Helper variables for OP_XCLASS opcode (for characters > 255). We define
3739    class_uchardata always so that it can be passed to add_to_class() always,
3740    though it will not be used in non-UTF 8-bit cases. This avoids having to supply
3741    alternative calls for the different cases. */
3742    
3743    pcre_uchar *class_uchardata;
3744  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3745  BOOL xclass;  BOOL xclass;
 pcre_uchar *class_uchardata;  
3746  pcre_uchar *class_uchardata_base;  pcre_uchar *class_uchardata_base;
3747  #endif  #endif
3748    
# Line 3484  to take the zero repeat into account. Th Line 3765  to take the zero repeat into account. Th
3765  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3766  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3767    
3768  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
3769    firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;
3770    
3771  /* The variable req_caseopt contains either the REQ_CASELESS value  /* The variable req_caseopt contains either the REQ_CASELESS value
3772  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 3505  for (;; ptr++) Line 3787  for (;; ptr++)
3787    BOOL is_recurse;    BOOL is_recurse;
3788    BOOL reset_bracount;    BOOL reset_bracount;
3789    int class_has_8bitchar;    int class_has_8bitchar;
3790    int class_single_char;    int class_one_char;
3791    int newoptions;    int newoptions;
3792    int recno;    int recno;
3793    int refsign;    int refsign;
3794    int skipbytes;    int skipbytes;
3795    int subreqchar;    pcre_uint32 subreqchar, subfirstchar;
3796    int subfirstchar;    pcre_int32 subreqcharflags, subfirstcharflags;
3797    int terminator;    int terminator;
3798    int mclength;    unsigned int mclength;
3799    int tempbracount;    int tempbracount;
3800      pcre_uint32 ec;
3801    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3802    
3803    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 3677  for (;; ptr++) Line 3960  for (;; ptr++)
3960      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3961      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3962      *firstcharptr = firstchar;      *firstcharptr = firstchar;
3963        *firstcharflagsptr = firstcharflags;
3964      *reqcharptr = reqchar;      *reqcharptr = reqchar;
3965        *reqcharflagsptr = reqcharflags;
3966      *codeptr = code;      *codeptr = code;
3967      *ptrptr = ptr;      *ptrptr = ptr;
3968      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3701  for (;; ptr++) Line 3986  for (;; ptr++)
3986      previous = NULL;      previous = NULL;
3987      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3988        {        {
3989        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
3990        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3991        }        }
3992      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3716  for (;; ptr++) Line 4001  for (;; ptr++)
4001      repeats. The value of reqchar doesn't change either. */      repeats. The value of reqchar doesn't change either. */
4002    
4003      case CHAR_DOT:      case CHAR_DOT:
4004      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4005      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4006        zerofirstcharflags = firstcharflags;
4007      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4008        zeroreqcharflags = reqcharflags;
4009      previous = code;      previous = code;
4010      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
4011      break;      break;
# Line 3792  for (;; ptr++) Line 4079  for (;; ptr++)
4079          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
4080        {        {
4081        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
4082        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4083        zerofirstchar = firstchar;        zerofirstchar = firstchar;
4084          zerofirstcharflags = firstcharflags;
4085        break;        break;
4086        }        }
4087    
# Line 3803  for (;; ptr++) Line 4091  for (;; ptr++)
4091    
4092      should_flip_negation = FALSE;      should_flip_negation = FALSE;
4093    
4094      /* For optimization purposes, we track some properties of the class.      /* For optimization purposes, we track some properties of the class:
4095      class_has_8bitchar will be non-zero, if the class contains at least one      class_has_8bitchar will be non-zero if the class contains at least one <
4096      < 256 character. class_single_char will be 1 if the class contains only      256 character; class_one_char will be 1 if the class contains just one
4097      a single character. */      character. */
4098    
4099      class_has_8bitchar = 0;      class_has_8bitchar = 0;
4100      class_single_char = 0;      class_one_char = 0;
4101    
4102      /* Initialize the 32-char bit map to all zeros. We build the map in a      /* Initialize the 32-char bit map to all zeros. We build the map in a
4103      temporary bit of memory, in case the class contains only 1 character (less      temporary bit of memory, in case the class contains fewer than two
4104      than 256), because in that case the compiled code doesn't use the bit map.      8-bit characters because in that case the compiled code doesn't use the bit
4105      */      map. */
4106    
4107      memset(classbits, 0, 32 * sizeof(pcre_uint8));      memset(classbits, 0, 32 * sizeof(pcre_uint8));
4108    
4109  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
4110      xclass = FALSE;                           /* No chars >= 256 */      xclass = FALSE;
4111      class_uchardata = code + LINK_SIZE + 2;   /* For UTF-8 items */      class_uchardata = code + LINK_SIZE + 2;   /* For XCLASS items */
4112      class_uchardata_base = class_uchardata;   /* For resetting in pass 1 */      class_uchardata_base = class_uchardata;   /* Save the start */
4113  #endif  #endif
4114    
4115      /* Process characters until ] is reached. By writing this as a "do" it      /* Process characters until ] is reached. By writing this as a "do" it
# Line 3843  for (;; ptr++) Line 4131  for (;; ptr++)
4131        /* In the pre-compile phase, accumulate the length of any extra        /* In the pre-compile phase, accumulate the length of any extra
4132        data and reset the pointer. This is so that very large classes that        data and reset the pointer. This is so that very large classes that
4133        contain a zillion > 255 characters no longer overwrite the work space        contain a zillion > 255 characters no longer overwrite the work space
4134        (which is on the stack). */        (which is on the stack). We have to remember that there was XCLASS data,
4135          however. */
4136    
4137        if (lengthptr != NULL)        if (lengthptr != NULL && class_uchardata > class_uchardata_base)
4138          {          {
4139            xclass = TRUE;
4140          *lengthptr += class_uchardata - class_uchardata_base;          *lengthptr += class_uchardata - class_uchardata_base;
4141          class_uchardata = class_uchardata_base;          class_uchardata = class_uchardata_base;
4142          }          }
# Line 3905  for (;; ptr++) Line 4195  for (;; ptr++)
4195          alpha. This relies on the fact that the class table starts with          alpha. This relies on the fact that the class table starts with
4196          alpha, lower, upper as the first 3 entries. */          alpha, lower, upper as the first 3 entries. */
4197    
4198          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
4199            posix_class = 0;            posix_class = 0;
4200    
4201          /* When PCRE_UCP is set, some of the POSIX classes are converted to          /* When PCRE_UCP is set, some of the POSIX classes are converted to
# Line 3948  for (;; ptr++) Line 4238  for (;; ptr++)
4238              for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset];              for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset];
4239            }            }
4240    
4241          /* Not see if we need to remove any special characters. An option          /* Now see if we need to remove any special characters. An option
4242          value of 1 removes vertical space and 2 removes underscore. */          value of 1 removes vertical space and 2 removes underscore. */
4243    
4244          if (tabopt < 0) tabopt = -tabopt;          if (tabopt < 0) tabopt = -tabopt;
# Line 3964  for (;; ptr++) Line 4254  for (;; ptr++)
4254            for (c = 0; c < 32; c++) classbits[c] |= pbits[c];            for (c = 0; c < 32; c++) classbits[c] |= pbits[c];
4255    
4256          ptr = tempptr + 1;          ptr = tempptr + 1;
4257          /* Every class contains at least one < 256 characters. */          /* Every class contains at least one < 256 character. */
4258          class_has_8bitchar = 1;          class_has_8bitchar = 1;
4259          /* Every class contains at least two characters. */          /* Every class contains at least two characters. */
4260          class_single_char = 2;          class_one_char = 2;
4261          continue;    /* End of POSIX syntax handling */          continue;    /* End of POSIX syntax handling */
4262          }          }
4263    
# Line 3975  for (;; ptr++) Line 4265  for (;; ptr++)
4265        of the specials, which just set a flag. The sequence \b is a special        of the specials, which just set a flag. The sequence \b is a special
4266        case. Inside a class (and only there) it is treated as backspace. We        case. Inside a class (and only there) it is treated as backspace. We
4267        assume that other escapes have more than one character in them, so        assume that other escapes have more than one character in them, so
4268        speculatively set both class_has_8bitchar and class_single_char bigger        speculatively set both class_has_8bitchar and class_one_char bigger
4269        than one. Unrecognized escapes fall through and are either treated        than one. Unrecognized escapes fall through and are either treated
4270        as literal characters (by default), or are faulted if        as literal characters (by default), or are faulted if
4271        PCRE_EXTRA is set. */        PCRE_EXTRA is set. */
4272    
4273        if (c == CHAR_BACKSLASH)        if (c == CHAR_BACKSLASH)
4274          {          {
4275          c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);          escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, TRUE);
4276    
4277          if (*errorcodeptr != 0) goto FAILED;          if (*errorcodeptr != 0) goto FAILED;
4278    
4279          if (-c == ESC_b) c = CHAR_BS;    /* \b is backspace in a class */          if (escape == 0)
4280          else if (-c == ESC_N)            /* \N is not supported in a class */            c = ec;
4281            else if (escape == ESC_b) c = CHAR_BS; /* \b is backspace in a class */
4282            else if (escape == ESC_N)            /* \N is not supported in a class */
4283            {            {
4284            *errorcodeptr = ERR71;            *errorcodeptr = ERR71;
4285            goto FAILED;            goto FAILED;
4286            }            }
4287          else if (-c == ESC_Q)            /* Handle start of quoted string */          else if (escape == ESC_Q)            /* Handle start of quoted string */
4288            {            {
4289            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
4290              {              {
# Line 4000  for (;; ptr++) Line 4293  for (;; ptr++)
4293            else inescq = TRUE;            else inescq = TRUE;
4294            continue;            continue;
4295            }            }
4296          else if (-c == ESC_E) continue;  /* Ignore orphan \E */          else if (escape == ESC_E) continue;  /* Ignore orphan \E */
4297    
4298          if (c < 0)          else
4299            {            {
4300            register const pcre_uint8 *cbits = cd->cbits;            register const pcre_uint8 *cbits = cd->cbits;
4301            /* Every class contains at least two < 256 characters. */            /* Every class contains at least two < 256 characters. */
4302            class_has_8bitchar++;            class_has_8bitchar++;
4303            /* Every class contains at least two characters. */            /* Every class contains at least two characters. */
4304            class_single_char += 2;            class_one_char += 2;
4305    
4306            switch (-c)            switch (escape)
4307              {              {
4308  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4309              case ESC_du:     /* These are the values given for \d etc */              case ESC_du:     /* These are the values given for \d etc */
# Line 4020  for (;; ptr++) Line 4313  for (;; ptr++)
4313              case ESC_su:     /* of the default ASCII testing. */              case ESC_su:     /* of the default ASCII testing. */
4314              case ESC_SU:              case ESC_SU:
4315              nestptr = ptr;              nestptr = ptr;
4316              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */              ptr = substitutes[escape - ESC_DU] - 1;  /* Just before substitute */
4317              class_has_8bitchar--;                /* Undo! */              class_has_8bitchar--;                /* Undo! */
4318              continue;              continue;
4319  #endif  #endif
# Line 4044  for (;; ptr++) Line 4337  for (;; ptr++)
4337    
4338              /* Perl 5.004 onwards omits VT from \s, but we must preserve it              /* Perl 5.004 onwards omits VT from \s, but we must preserve it
4339              if it was previously set by something earlier in the character              if it was previously set by something earlier in the character
4340              class. */              class. Luckily, the value of CHAR_VT is 0x0b in both ASCII and
4341                EBCDIC, so we lazily just adjust the appropriate bit. */
4342    
4343              case ESC_s:              case ESC_s:
4344              classbits[0] |= cbits[cbit_space];              classbits[0] |= cbits[cbit_space];
# Line 4057  for (;; ptr++) Line 4351  for (;; ptr++)
4351              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];
4352              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */
4353              continue;              continue;
4354    
4355                /* The rest apply in both UCP and non-UCP cases. */
4356    
4357              case ESC_h:              case ESC_h:
4358              SETBIT(classbits, 0x09); /* VT */              (void)add_list_to_class(classbits, &class_uchardata, options, cd,
4359              SETBIT(classbits, 0x20); /* SPACE */                PRIV(hspace_list), NOTACHAR);
             SETBIT(classbits, 0xa0); /* NSBP */  
 #ifndef COMPILE_PCRE8  
             xclass = TRUE;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x1680;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x180e;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x2000;  
             *class_uchardata++ = 0x200a;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x202f;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x205f;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x3000;  
 #elif defined SUPPORT_UTF  
             if (utf)  
               {  
               xclass = TRUE;  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x1680, class_uchardata);  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x180e, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x2000, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x200a, class_uchardata);  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x202f, class_uchardata);  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x205f, class_uchardata);  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x3000, class_uchardata);  
               }  
 #endif  
4360              continue;              continue;
4361    
4362              case ESC_H:              case ESC_H:
4363              for (c = 0; c < 32; c++)              (void)add_not_list_to_class(classbits, &class_uchardata, options,
4364                {                cd, PRIV(hspace_list));
               int x = 0xff;  
               switch (c)  
                 {  
                 case 0x09/8: x ^= 1 << (0x09%8); break;  
                 case 0x20/8: x ^= 1 << (0x20%8); break;  
                 case 0xa0/8: x ^= 1 << (0xa0%8); break;  
                 default: break;  
                 }  
               classbits[c] |= x;  
               }  
 #ifndef COMPILE_PCRE8  
             xclass = TRUE;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x0100;  
             *class_uchardata++ = 0x167f;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x1681;  
             *class_uchardata++ = 0x180d;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x180f;  
             *class_uchardata++ = 0x1fff;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x200b;  
             *class_uchardata++ = 0x202e;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x2030;  
             *class_uchardata++ = 0x205e;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x2060;  
             *class_uchardata++ = 0x2fff;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x3001;  
 #ifdef SUPPORT_UTF  
             if (utf)  
               class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);  
             else  
 #endif  
               *class_uchardata++ = 0xffff;  
 #elif defined SUPPORT_UTF  
             if (utf)  
               {  
               xclass = TRUE;  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x167f, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x1681, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x180d, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x180f, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x1fff, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x200b, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x202e, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x2030, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x205e, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x2060, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x2fff, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x3001, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);  
               }  
 #endif  
4365              continue;              continue;
4366    
4367              case ESC_v:              case ESC_v:
4368              SETBIT(classbits, 0x0a); /* LF */              (void)add_list_to_class(classbits, &class_uchardata, options, cd,
4369              SETBIT(classbits, 0x0b); /* VT */                PRIV(vspace_list), NOTACHAR);
             SETBIT(classbits, 0x0c); /* FF */  
             SETBIT(classbits, 0x0d); /* CR */  
             SETBIT(classbits, 0x85); /* NEL */  
 #ifndef COMPILE_PCRE8  
             xclass = TRUE;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x2028;  
             *class_uchardata++ = 0x2029;  
 #elif defined SUPPORT_UTF  
             if (utf)  
               {  
               xclass = TRUE;  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x2028, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x2029, class_uchardata);  
               }  
 #endif  
4370              continue;              continue;
4371    
4372              case ESC_V:              case ESC_V:
4373              for (c = 0; c < 32; c++)              (void)add_not_list_to_class(classbits, &class_uchardata, options,
4374                {                cd, PRIV(vspace_list));
               int x = 0xff;  
               switch (c)  
                 {  
                 case 0x0a/8: x ^= 1 << (0x0a%8);  
                              x ^= 1 << (0x0b%8);  
                              x ^= 1 << (0x0c%8);  
                              x ^= 1 << (0x0d%8);  
                              break;  
                 case 0x85/8: x ^= 1 << (0x85%8); break;  
                 default: break;  
                 }  
               classbits[c] |= x;  
               }  
   
 #ifndef COMPILE_PCRE8  
             xclass = TRUE;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x0100;  
             *class_uchardata++ = 0x2027;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x202a;  
 #ifdef SUPPORT_UTF  
             if (utf)  
               class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);  
             else  
 #endif  
               *class_uchardata++ = 0xffff;  
 #elif defined SUPPORT_UTF  
             if (utf)  
               {  
               xclass = TRUE;  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x2027, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x202a, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);  
               }  
 #endif  
4375              continue;              continue;
4376    
4377  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 4239  for (;; ptr++) Line 4379  for (;; ptr++)
4379              case ESC_P:              case ESC_P:
4380                {                {
4381                BOOL negated;                BOOL negated;
4382                int pdata;                unsigned int ptype = 0, pdata = 0;
4383                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
4384                if (ptype < 0) goto FAILED;                  goto FAILED;
4385                xclass = TRUE;                *class_uchardata++ = ((escape == ESC_p) != negated)?
               *class_uchardata++ = ((-c == ESC_p) != negated)?  
4386                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4387                *class_uchardata++ = ptype;                *class_uchardata++ = ptype;
4388                *class_uchardata++ = pdata;                *class_uchardata++ = pdata;
# Line 4262  for (;; ptr++) Line 4401  for (;; ptr++)
4401                goto FAILED;                goto FAILED;
4402                }                }
4403              class_has_8bitchar--;    /* Undo the speculative increase. */              class_has_8bitchar--;    /* Undo the speculative increase. */
4404              class_single_char -= 2;  /* Undo the speculative increase. */              class_one_char -= 2;     /* Undo the speculative increase. */
4405              c = *ptr;                /* Get the final character and fall through */              c = *ptr;                /* Get the final character and fall through */
4406              break;              break;
4407              }              }
4408            }            }
4409    
4410          /* Fall through if we have a single character (c >= 0). This may be          /* Fall through if the escape just defined a single character (c >= 0).
4411          greater than 256. */          This may be greater than 256. */
4412    
4413            escape = 0;
4414    
4415          }   /* End of backslash handling */          }   /* End of backslash handling */
4416    
4417        /* A single character may be followed by '-' to form a range. However,        /* A character may be followed by '-' to form a range. However, Perl does
4418        Perl does not permit ']' to be the end of the range. A '-' character        not permit ']' to be the end of the range. A '-' character at the end is
4419        at the end is treated as a literal. Perl ignores orphaned \E sequences        treated as a literal. Perl ignores orphaned \E sequences entirely. The
4420        entirely. The code for handling \Q and \E is messy. */        code for handling \Q and \E is messy. */
4421    
4422        CHECK_RANGE:        CHECK_RANGE:
4423        while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)        while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
# Line 4284  for (;; ptr++) Line 4425  for (;; ptr++)
4425          inescq = FALSE;          inescq = FALSE;
4426          ptr += 2;          ptr += 2;
4427          }          }
   
4428        oldptr = ptr;        oldptr = ptr;
4429    
4430        /* Remember \r or \n */        /* Remember if \r or \n were explicitly used */
4431    
4432        if (c == CHAR_CR || c == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;        if (c == CHAR_CR || c == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;
4433    
# Line 4295  for (;; ptr++) Line 4435  for (;; ptr++)
4435    
4436        if (!inescq && ptr[1] == CHAR_MINUS)        if (!inescq && ptr[1] == CHAR_MINUS)
4437          {          {
4438          int d;          pcre_uint32 d;
4439          ptr += 2;          ptr += 2;
4440          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;
4441    
# Line 4310  for (;; ptr++) Line 4450  for (;; ptr++)
4450            inescq = TRUE;            inescq = TRUE;
4451            break;            break;
4452            }            }
4453    
4454            /* Minus (hyphen) at the end of a class is treated as a literal, so put
4455            back the pointer and jump to handle the character that preceded it. */
4456    
4457          if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))          if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))
4458            {            {
4459            ptr = oldptr;            ptr = oldptr;
4460            goto LONE_SINGLE_CHARACTER;            goto CLASS_SINGLE_CHARACTER;
4461            }            }
4462    
4463            /* Otherwise, we have a potential range; pick up the next character */
4464    
4465  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
4466          if (utf)          if (utf)
# Line 4332  for (;; ptr++) Line 4477  for (;; ptr++)
4477    
4478          if (!inescq && d == CHAR_BACKSLASH)          if (!inescq && d == CHAR_BACKSLASH)
4479            {            {
4480            d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);            int descape;
4481              descape = check_escape(&ptr, &d, errorcodeptr, cd->bracount, options, TRUE);
4482            if (*errorcodeptr != 0) goto FAILED;            if (*errorcodeptr != 0) goto FAILED;
4483    
4484            /* \b is backspace; any other special means the '-' was literal */            /* \b is backspace; any other special means the '-' was literal. */
4485    
4486            if (d < 0)            if (descape != 0)
4487              {              {
4488              if (d == -ESC_b) d = CHAR_BS; else              if (descape == ESC_b) d = CHAR_BS; else
4489                {                {
4490                ptr = oldptr;                ptr = oldptr;
4491                goto LONE_SINGLE_CHARACTER;  /* A few lines below */                goto CLASS_SINGLE_CHARACTER;  /* A few lines below */
4492                }                }
4493              }              }
4494            }            }
4495    
4496          /* Check that the two values are in the correct order. Optimize          /* Check that the two values are in the correct order. Optimize
4497          one-character ranges */          one-character ranges. */
4498    
4499          if (d < c)          if (d < c)
4500            {            {
4501            *errorcodeptr = ERR8;            *errorcodeptr = ERR8;
4502            goto FAILED;            goto FAILED;
4503            }            }
4504            if (d == c) goto CLASS_SINGLE_CHARACTER;  /* A few lines below */
4505    
4506          if (d == c) goto LONE_SINGLE_CHARACTER;  /* A few lines below */          /* We have found a character range, so single character optimizations
4507            cannot be done anymore. Any value greater than 1 indicates that there
4508            is more than one character. */
4509    
4510            class_one_char = 2;
4511    
4512          /* Remember \r or \n */          /* Remember an explicit \r or \n, and add the range to the class. */
4513    
4514          if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;          if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;
4515    
4516          /* Since we found a character range, single character optimizations          class_has_8bitchar +=
4517          cannot be done anymore. */            add_to_class(classbits, &class_uchardata, options, cd, c, d);
4518          class_single_char = 2;  
   
         /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless  
         matching, we have to use an XCLASS with extra data items. Caseless  
         matching for characters > 127 is available only if UCP support is  
         available. */  
   
 #if defined SUPPORT_UTF && !(defined COMPILE_PCRE8)  
         if ((d > 255) || (utf && ((options & PCRE_CASELESS) != 0 && d > 127)))  
 #elif defined  SUPPORT_UTF  
         if (utf && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))  
 #elif !(defined COMPILE_PCRE8)  
         if (d > 255)  
 #endif  
 #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)  
           {  
           xclass = TRUE;  
   
           /* With UCP support, we can find the other case equivalents of  
           the relevant characters. There may be several ranges. Optimize how  
           they fit with the basic range. */  
   
 #ifdef SUPPORT_UCP  
 #ifndef COMPILE_PCRE8  
           if (utf && (options & PCRE_CASELESS) != 0)  
 #else  
           if ((options & PCRE_CASELESS) != 0)  
 #endif  
             {  
             unsigned int occ, ocd;  
             unsigned int cc = c;  
             unsigned int origd = d;  
             while (get_othercase_range(&cc, origd, &occ, &ocd))  
               {  
               if (occ >= (unsigned int)c &&  
                   ocd <= (unsigned int)d)  
                 continue;                          /* Skip embedded ranges */  
   
               if (occ < (unsigned int)c  &&  
                   ocd >= (unsigned int)c - 1)      /* Extend the basic range */  
                 {                                  /* if there is overlap,   */  
                 c = occ;                           /* noting that if occ < c */  
                 continue;                          /* we can't have ocd > d  */  
                 }                                  /* because a subrange is  */  
               if (ocd > (unsigned int)d &&  
                   occ <= (unsigned int)d + 1)      /* always shorter than    */  
                 {                                  /* the basic range.       */  
                 d = ocd;  
                 continue;  
                 }  
   
               if (occ == ocd)  
                 {  
                 *class_uchardata++ = XCL_SINGLE;  
                 }  
               else  
                 {  
                 *class_uchardata++ = XCL_RANGE;  
                 class_uchardata += PRIV(ord2utf)(occ, class_uchardata);  
                 }  
               class_uchardata += PRIV(ord2utf)(ocd, class_uchardata);  
               }  
             }  
 #endif  /* SUPPORT_UCP */  
   
           /* Now record the original range, possibly modified for UCP caseless  
           overlapping ranges. */  
   
           *class_uchardata++ = XCL_RANGE;  
 #ifdef SUPPORT_UTF  
 #ifndef COMPILE_PCRE8  
           if (utf)  
             {  
             class_uchardata += PRIV(ord2utf)(c, class_uchardata);  
             class_uchardata += PRIV(ord2utf)(d, class_uchardata);  
             }  
           else  
             {  
             *class_uchardata++ = c;  
             *class_uchardata++ = d;  
             }  
 #else  
           class_uchardata += PRIV(ord2utf)(c, class_uchardata);  
           class_uchardata += PRIV(ord2utf)(d, class_uchardata);  
 #endif  
 #else /* SUPPORT_UTF */  
           *class_uchardata++ = c;  
           *class_uchardata++ = d;  
 #endif /* SUPPORT_UTF */  
   
           /* With UCP support, we are done. Without UCP support, there is no  
           caseless matching for UTF characters > 127; we can use the bit map  
           for the smaller ones. As for 16 bit characters without UTF, we  
           can still use  */  
   
 #ifdef SUPPORT_UCP  
 #ifndef COMPILE_PCRE8  
           if (utf)  
 #endif  
             continue;    /* With next character in the class */  
 #endif  /* SUPPORT_UCP */  
   
 #if defined SUPPORT_UTF && !defined(SUPPORT_UCP) && !(defined COMPILE_PCRE8)  
           if (utf)  
             {  
             if ((options & PCRE_CASELESS) == 0 || c > 127) continue;  
             /* Adjust upper limit and fall through to set up the map */  
             d = 127;  
             }  
           else  
             {  
             if (c > 255) continue;  
             /* Adjust upper limit and fall through to set up the map */  
             d = 255;  
             }  
 #elif defined SUPPORT_UTF && !defined(SUPPORT_UCP)  
           if ((options & PCRE_CASELESS) == 0 || c > 127) continue;  
           /* Adjust upper limit and fall through to set up the map */  
           d = 127;  
 #else  
           if (c > 255) continue;  
           /* Adjust upper limit and fall through to set up the map */  
           d = 255;  
 #endif  /* SUPPORT_UTF && !SUPPORT_UCP && !COMPILE_PCRE8 */  
           }  
 #endif  /* SUPPORT_UTF || !COMPILE_PCRE8 */  
   
         /* We use the bit map for 8 bit mode, or when the characters fall  
         partially or entirely to [0-255] ([0-127] for UCP) ranges. */  
   
         class_has_8bitchar = 1;  
   
         /* We can save a bit of time by skipping this in the pre-compile. */  
   
         if (lengthptr == NULL) for (; c <= d; c++)  
           {  
           classbits[c/8] |= (1 << (c&7));  
           if ((options & PCRE_CASELESS) != 0)  
             {  
             int uc = cd->fcc[c]; /* flip case */  
             classbits[uc/8] |= (1 << (uc&7));  
             }  
           }  
   
4519          continue;   /* Go get the next char in the class */          continue;   /* Go get the next char in the class */
4520          }          }
4521    
4522        /* Handle a lone single character - we can get here for a normal        /* Handle a single character - we can get here for a normal non-escape
4523        non-escape char, or after \ that introduces a single character or for an        char, or after \ that introduces a single character or for an apparent
4524        apparent range that isn't. */        range that isn't. Only the value 1 matters for class_one_char, so don't
4525          increase it if it is already 2 or more ... just in case there's a class
4526          with a zillion characters in it. */
4527    
4528          CLASS_SINGLE_CHARACTER:
4529          if (class_one_char < 2) class_one_char++;
4530    
4531          /* If class_one_char is 1, we have the first single character in the
4532          class, and there have been no prior ranges, or XCLASS items generated by
4533          escapes. If this is the final character in the class, we can optimize by
4534          turning the item into a 1-character OP_CHAR[I] if it's positive, or
4535          OP_NOT[I] if it's negative. In the positive case, it can cause firstchar
4536          to be set. Otherwise, there can be no first char if this item is first,
4537          whatever repeat count may follow. In the case of reqchar, save the
4538          previous value for reinstating. */
4539    
4540        LONE_SINGLE_CHARACTER:        if (class_one_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
   
       /* Only the value of 1 matters for class_single_char. */  
   
       if (class_single_char < 2) class_single_char++;  
   
       /* If class_charcount is 1, we saw precisely one character. As long as  
       there was no use of \p or \P, in other words, no use of any XCLASS  
       features, we can optimize.  
   
       The optimization throws away the bit map. We turn the item into a  
       1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative.  
       In the positive case, it can cause firstchar to be set. Otherwise, there  
       can be no first char if this item is first, whatever repeat count may  
       follow. In the case of reqchar, save the previous value for reinstating. */  
   
       if (class_single_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)  
4541          {          {
4542          ptr++;          ptr++;
4543          zeroreqchar = reqchar;          zeroreqchar = reqchar;
4544            zeroreqcharflags = reqcharflags;
4545    
4546          if (negate_class)          if (negate_class)
4547            {            {
4548            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;  #ifdef SUPPORT_UCP
4549              int d;
4550    #endif
4551              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4552            zerofirstchar = firstchar;            zerofirstchar = firstchar;
4553            *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;            zerofirstcharflags = firstcharflags;
4554  #ifdef SUPPORT_UTF  
4555            if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)            /* For caseless UTF-8 mode when UCP support is available, check
4556              code += PRIV(ord2utf)(c, code);            whether this character has more than one other case. If so, generate
4557            else            a special OP_NOTPROP item instead of OP_NOTI. */
4558    
4559    #ifdef SUPPORT_UCP
4560              if (utf && (options & PCRE_CASELESS) != 0 &&
4561                  (d = UCD_CASESET(c)) != 0)
4562                {
4563                *code++ = OP_NOTPROP;
4564                *code++ = PT_CLIST;
4565                *code++ = d;
4566                }
4567              else
4568  #endif  #endif
4569              *code++ = c;            /* Char has only one other case, or UCP not available */
4570            goto NOT_CHAR;  
4571                {
4572                *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
4573    #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4574                if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
4575                  code += PRIV(ord2utf)(c, code);
4576                else
4577    #endif
4578                  *code++ = c;
4579                }
4580    
4581              /* We are finished with this character class */
4582    
4583              goto END_CLASS;
4584            }            }
4585    
4586          /* For a single, positive character, get the value into mcbuffer, and          /* For a single, positive character, get the value into mcbuffer, and
4587          then we can handle this with the normal one-character code. */          then we can handle this with the normal one-character code. */
4588    
4589  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4590          if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)          if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
4591            mclength = PRIV(ord2utf)(c, mcbuffer);            mclength = PRIV(ord2utf)(c, mcbuffer);
4592          else          else
# Line 4564  for (;; ptr++) Line 4597  for (;; ptr++)
4597            }            }
4598          goto ONE_CHAR;          goto ONE_CHAR;
4599          }       /* End of 1-char optimization */          }       /* End of 1-char optimization */
4600    
4601        /* Handle a character that cannot go in the bit map. */        /* There is more than one character in the class, or an XCLASS item
4602          has been generated. Add this character to the class. */
4603  #if defined SUPPORT_UTF && !(defined COMPILE_PCRE8)  
4604        if ((c > 255) || (utf && ((options & PCRE_CASELESS) != 0 && c > 127)))        class_has_8bitchar +=
4605  #elif defined SUPPORT_UTF          add_to_class(classbits, &class_uchardata, options, cd, c, c);
       if (utf && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))  
 #elif !(defined COMPILE_PCRE8)  
       if (c > 255)  
 #endif  
   
 #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)  
         {  
         xclass = TRUE;  
         *class_uchardata++ = XCL_SINGLE;  
 #ifdef SUPPORT_UTF  
 #ifndef COMPILE_PCRE8  
         /* In non 8 bit mode, we can get here even if we are not in UTF mode. */  
         if (!utf)  
           *class_uchardata++ = c;  
         else  
 #endif  
           class_uchardata += PRIV(ord2utf)(c, class_uchardata);  
 #else /* SUPPORT_UTF */  
         *class_uchardata++ = c;  
 #endif /* SUPPORT_UTF */  
   
 #ifdef SUPPORT_UCP  
 #ifdef COMPILE_PCRE8  
         if ((options & PCRE_CASELESS) != 0)  
 #else  
         /* In non 8 bit mode, we can get here even if we are not in UTF mode. */  
         if (utf && (options & PCRE_CASELESS) != 0)  
 #endif  
           {  
           unsigned int othercase;  
           if ((int)(othercase = UCD_OTHERCASE(c)) != c)  
             {  
             *class_uchardata++ = XCL_SINGLE;  
             class_uchardata += PRIV(ord2utf)(othercase, class_uchardata);  
             }  
           }  
 #endif  /* SUPPORT_UCP */  
   
         }  
       else  
 #endif  /* SUPPORT_UTF || COMPILE_PCRE16 */  
   
       /* Handle a single-byte character */  
         {  
         class_has_8bitchar = 1;  
         classbits[c/8] |= (1 << (c&7));  
         if ((options & PCRE_CASELESS) != 0)  
           {  
           c = cd->fcc[c]; /* flip case */  
           classbits[c/8] |= (1 << (c&7));  
           }  
         }  
4606        }        }
4607    
4608      /* Loop until ']' reached. This "while" is the end of the "do" far above.      /* Loop until ']' reached. This "while" is the end of the "do" far above.
# Line 4641  for (;; ptr++) Line 4622  for (;; ptr++)
4622        goto FAILED;        goto FAILED;
4623        }        }
4624    
4625        /* We will need an XCLASS if data has been placed in class_uchardata. In
4626        the second phase this is a sufficient test. However, in the pre-compile
4627        phase, class_uchardata gets emptied to prevent workspace overflow, so it
4628        only if the very last character in the class needs XCLASS will it contain
4629        anything at this point. For this reason, xclass gets set TRUE above when
4630        uchar_classdata is emptied, and that's why this code is the way it is here
4631        instead of just doing a test on class_uchardata below. */
4632    
4633    #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
4634        if (class_uchardata > class_uchardata_base) xclass = TRUE;
4635    #endif
4636    
4637      /* If this is the first thing in the branch, there can be no first char      /* If this is the first thing in the branch, there can be no first char
4638      setting, whatever the repeat count. Any reqchar setting must remain      setting, whatever the repeat count. Any reqchar setting must remain
4639      unchanged after any kind of repeat. */      unchanged after any kind of repeat. */
4640    
4641      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4642      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4643        zerofirstcharflags = firstcharflags;
4644      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4645        zeroreqcharflags = reqcharflags;
4646    
4647      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4648      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 4703  for (;; ptr++) Line 4698  for (;; ptr++)
4698        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
4699        }        }
4700      code += 32 / sizeof(pcre_uchar);      code += 32 / sizeof(pcre_uchar);
4701      NOT_CHAR:  
4702        END_CLASS:
4703      break;      break;
4704    
4705    
# Line 4741  for (;; ptr++) Line 4737  for (;; ptr++)
4737      if (repeat_min == 0)      if (repeat_min == 0)
4738        {        {
4739        firstchar = zerofirstchar;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4740          firstcharflags = zerofirstcharflags;
4741        reqchar = zeroreqchar;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4742          reqcharflags = zeroreqcharflags;
4743        }        }
4744    
4745      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4813  for (;; ptr++) Line 4811  for (;; ptr++)
4811      if (*previous == OP_CHAR || *previous == OP_CHARI      if (*previous == OP_CHAR || *previous == OP_CHARI
4812          || *previous == OP_NOT || *previous == OP_NOTI)          || *previous == OP_NOT || *previous == OP_NOTI)
4813        {        {
4814        switch (*previous)        switch (*previous)
4815          {          {
4816          default: /* Make compiler happy. */          default: /* Make compiler happy. */
4817          case OP_CHAR:  op_type = OP_STAR - OP_STAR; break;          case OP_CHAR:  op_type = OP_STAR - OP_STAR; break;
# Line 4827  for (;; ptr++) Line 4825  for (;; ptr++)
4825        hold the length of the character in bytes, plus UTF_LENGTH to flag that        hold the length of the character in bytes, plus UTF_LENGTH to flag that
4826        it's a length rather than a small character. */        it's a length rather than a small character. */
4827    
4828  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4829        if (utf && NOT_FIRSTCHAR(code[-1]))        if (utf && NOT_FIRSTCHAR(code[-1]))
4830          {          {
4831          pcre_uchar *lastchar = code - 1;          pcre_uchar *lastchar = code - 1;
# Line 4844  for (;; ptr++) Line 4842  for (;; ptr++)
4842          {          {
4843          c = code[-1];          c = code[-1];
4844          if (*previous <= OP_CHARI && repeat_min > 1)          if (*previous <= OP_CHARI && repeat_min > 1)
4845            reqchar = c | req_caseopt | cd->req_varyopt;            {
4846              reqchar = c;
4847              reqcharflags = req_caseopt | cd->req_varyopt;
4848              }
4849          }          }
4850    
4851        /* 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 4963  for (;; ptr++) Line 4964  for (;; ptr++)
4964    
4965          if (repeat_max < 0)          if (repeat_max < 0)
4966            {            {
4967  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4968            if (utf && (c & UTF_LENGTH) != 0)            if (utf && (c & UTF_LENGTH) != 0)
4969              {              {
4970              memcpy(code, utf_chars, IN_UCHARS(c & 7));              memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 4988  for (;; ptr++) Line 4989  for (;; ptr++)
4989    
4990          else if (repeat_max != repeat_min)          else if (repeat_max != repeat_min)
4991            {            {
4992  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4993            if (utf && (c & UTF_LENGTH) != 0)            if (utf && (c & UTF_LENGTH) != 0)
4994              {              {
4995              memcpy(code, utf_chars, IN_UCHARS(c & 7));              memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 5018  for (;; ptr++) Line 5019  for (;; ptr++)
5019    
5020        /* The character or character type itself comes last in all cases. */        /* The character or character type itself comes last in all cases. */
5021    
5022  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
5023        if (utf && (c & UTF_LENGTH) != 0)        if (utf && (c & UTF_LENGTH) != 0)
5024          {          {
5025          memcpy(code, utf_chars, IN_UCHARS(c & 7));          memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 5226  for (;; ptr++) Line 5227  for (;; ptr++)
5227    
5228            else            else
5229              {              {
5230              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;              if (groupsetfirstchar && reqcharflags < 0)
5231                  {
5232                  reqchar = firstchar;
5233                  reqcharflags = firstcharflags;
5234                  }
5235    
5236              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5237                {                {
# Line 5593  for (;; ptr++) Line 5598  for (;; ptr++)
5598        ptr++;        ptr++;
5599        while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_letter) != 0) ptr++;        while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_letter) != 0) ptr++;
5600        namelen = (int)(ptr - name);        namelen = (int)(ptr - name);
5601    
5602        /* It appears that Perl allows any characters whatsoever, other than        /* It appears that Perl allows any characters whatsoever, other than
5603        a closing parenthesis, to appear in arguments, so we no longer insist on        a closing parenthesis, to appear in arguments, so we no longer insist on
5604        letters, digits, and underscores. */        letters, digits, and underscores. */
# Line 5603  for (;; ptr++) Line 5608  for (;; ptr++)
5608          arg = ++ptr;          arg = ++ptr;
5609          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5610          arglen = (int)(ptr - arg);          arglen = (int)(ptr - arg);
5611          if (arglen > (int)MAX_MARK)          if ((unsigned int)arglen > MAX_MARK)
5612            {            {
5613            *errorcodeptr = ERR75;            *errorcodeptr = ERR75;
5614            goto FAILED;            goto FAILED;
5615            }            }
5616          }          }
5617    
5618        if (*ptr != CHAR_RIGHT_PARENTHESIS)        if (*ptr != CHAR_RIGHT_PARENTHESIS)
# Line 5623  for (;; ptr++) Line 5628  for (;; ptr++)
5628          if (namelen == verbs[i].len &&          if (namelen == verbs[i].len &&
5629              STRNCMP_UC_C8(name, vn, namelen) == 0)              STRNCMP_UC_C8(name, vn, namelen) == 0)
5630            {            {
5631              int setverb;
5632    
5633            /* Check for open captures before ACCEPT and convert it to            /* Check for open captures before ACCEPT and convert it to
5634            ASSERT_ACCEPT if in an assertion. */            ASSERT_ACCEPT if in an assertion. */
5635    
# Line 5640  for (;; ptr++) Line 5647  for (;; ptr++)
5647                *code++ = OP_CLOSE;                *code++ = OP_CLOSE;
5648                PUT2INC(code, 0, oc->number);                PUT2INC(code, 0, oc->number);
5649                }                }
5650              *code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;              setverb = *code++ =
5651                  (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5652    
5653              /* Do not set firstchar after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5654              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
5655              }              }
5656    
5657            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5655  for (;; ptr++) Line 5663  for (;; ptr++)
5663                *errorcodeptr = ERR66;                *errorcodeptr = ERR66;
5664                goto FAILED;                goto FAILED;
5665                }                }
5666              *code = verbs[i].op;              setverb = *code++ = verbs[i].op;
             if (*code++ == OP_THEN) cd->external_flags |= PCRE_HASTHEN;  
5667              }              }
5668    
5669            else            else
# Line 5666  for (;; ptr++) Line 5673  for (;; ptr++)
5673                *errorcodeptr = ERR59;                *errorcodeptr = ERR59;
5674                goto FAILED;                goto FAILED;
5675                }                }
5676              *code = verbs[i].op_arg;              setverb = *code++ = verbs[i].op_arg;
             if (*code++ == OP_THEN_ARG) cd->external_flags |= PCRE_HASTHEN;  
5677              *code++ = arglen;              *code++ = arglen;
5678              memcpy(code, arg, IN_UCHARS(arglen));              memcpy(code, arg, IN_UCHARS(arglen));
5679              code += arglen;              code += arglen;
5680              *code++ = 0;              *code++ = 0;
5681              }              }
5682    
5683              switch (setverb)
5684                {
5685                case OP_THEN:
5686                case OP_THEN_ARG:
5687                cd->external_flags |= PCRE_HASTHEN;
5688                break;
5689    
5690                case OP_PRUNE:
5691                case OP_PRUNE_ARG:
5692                case OP_SKIP:
5693                case OP_SKIP_ARG:
5694                cd->had_pruneorskip = TRUE;
5695                break;
5696                }
5697    
5698            break;  /* Found verb, exit loop */            break;  /* Found verb, exit loop */
5699            }            }
5700    
# Line 5796  for (;; ptr++) Line 5817  for (;; ptr++)
5817          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)
5818            {            {
5819            if (recno >= 0)            if (recno >= 0)
5820              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;              recno = (IS_DIGIT(*ptr))? recno * 10 + (int)(*ptr - CHAR_0) : -1;
5821            ptr++;            ptr++;
5822            }            }
5823          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5824    
5825          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) ||
5826              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
5827            {            {
5828            ptr--;      /* Error offset */            ptr--;      /* Error offset */
# Line 6041  for (;; ptr++) Line 6062  for (;; ptr++)
6062    
6063            if (lengthptr != NULL)            if (lengthptr != NULL)
6064              {              {
6065              if (*ptr != terminator)              if (*ptr != (pcre_uchar)terminator)
6066                {                {
6067                *errorcodeptr = ERR42;                *errorcodeptr = ERR42;
6068                goto FAILED;                goto FAILED;
# Line 6183  for (;; ptr++) Line 6204  for (;; ptr++)
6204              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
6205              goto FAILED;              goto FAILED;
6206              }              }
6207            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6208              {              {
6209              *errorcodeptr = ERR42;              *errorcodeptr = ERR42;
6210              goto FAILED;              goto FAILED;
# Line 6289  for (;; ptr++) Line 6310  for (;; ptr++)
6310            while(IS_DIGIT(*ptr))            while(IS_DIGIT(*ptr))
6311              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6312    
6313            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6314              {              {
6315              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
6316              goto FAILED;              goto FAILED;
# Line 6393  for (;; ptr++) Line 6414  for (;; ptr++)
6414    
6415          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6416    
6417          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6418          continue;          continue;
6419    
6420    
# Line 6527  for (;; ptr++) Line 6548  for (;; ptr++)
6548           cond_depth +           cond_depth +
6549             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6550           &subfirstchar,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6551             &subfirstcharflags,
6552           &subreqchar,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6553             &subreqcharflags,
6554           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6555           cd,                              /* Tables block */           cd,                              /* Tables block */
6556           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6588  for (;; ptr++) Line 6611  for (;; ptr++)
6611            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6612            goto FAILED;            goto FAILED;
6613            }            }
6614          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;          if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
6615          }          }
6616        }        }
6617    
# Line 6637  for (;; ptr++) Line 6660  for (;; ptr++)
6660      back off. */      back off. */
6661    
6662      zeroreqchar = reqchar;      zeroreqchar = reqchar;
6663        zeroreqcharflags = reqcharflags;
6664      zerofirstchar = firstchar;      zerofirstchar = firstchar;
6665        zerofirstcharflags = firstcharflags;
6666      groupsetfirstchar = FALSE;      groupsetfirstchar = FALSE;
6667    
6668      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
# Line 6648  for (;; ptr++) Line 6673  for (;; ptr++)
6673        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
6674        repeat forces firstchar to "none". */        repeat forces firstchar to "none". */
6675    
6676        if (firstchar == REQ_UNSET)        if (firstcharflags == REQ_UNSET)
6677          {          {
6678          if (subfirstchar >= 0)          if (subfirstcharflags >= 0)
6679            {            {
6680            firstchar = subfirstchar;            firstchar = subfirstchar;
6681              firstcharflags = subfirstcharflags;
6682            groupsetfirstchar = TRUE;            groupsetfirstchar = TRUE;
6683            }            }
6684          else firstchar = REQ_NONE;          else firstcharflags = REQ_NONE;
6685          zerofirstchar = REQ_NONE;          zerofirstcharflags = REQ_NONE;
6686          }          }
6687    
6688        /* If firstchar was previously set, convert the subpattern's firstchar        /* If firstchar was previously set, convert the subpattern's firstchar
6689        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
6690        existence beforehand. */        existence beforehand. */
6691    
6692        else if (subfirstchar >= 0 && subreqchar < 0)        else if (subfirstcharflags >= 0 && subreqcharflags < 0)
6693          subreqchar = subfirstchar | tempreqvary;          {
6694            subreqchar = subfirstchar;
6695            subreqcharflags = subfirstcharflags | tempreqvary;
6696            }
6697    
6698        /* 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
6699        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6700    
6701        if (subreqchar >= 0) reqchar = subreqchar;        if (subreqcharflags >= 0)
6702            {
6703            reqchar = subreqchar;
6704            reqcharflags = subreqcharflags;
6705            }
6706        }        }
6707    
6708      /* 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 6680  for (;; ptr++) Line 6713  for (;; ptr++)
6713      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
6714      firstchar, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6715    
6716      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;      else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
6717          {
6718          reqchar = subreqchar;
6719          reqcharflags = subreqcharflags;
6720          }
6721      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6722    
6723    
# Line 6688  for (;; ptr++) Line 6725  for (;; ptr++)
6725      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
6726      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
6727      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
6728      are ESC_REF plus the reference number. Only back references and those types      are negative the reference number. Only back references and those types
6729      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
6730      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
6731      ever created. */      ever created. */
6732    
6733      case CHAR_BACKSLASH:      case CHAR_BACKSLASH:
6734      tempptr = ptr;      tempptr = ptr;
6735      c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE);      escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, FALSE);
6736    
6737      if (*errorcodeptr != 0) goto FAILED;      if (*errorcodeptr != 0) goto FAILED;
6738    
6739      if (c < 0)      if (escape == 0)
6740          c = ec;
6741        else
6742        {        {
6743        if (-c == ESC_Q)            /* Handle start of quoted string */        if (escape == ESC_Q)            /* Handle start of quoted string */
6744          {          {
6745          if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)          if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
6746            ptr += 2;               /* avoid empty string */            ptr += 2;               /* avoid empty string */
# Line 6708  for (;; ptr++) Line 6748  for (;; ptr++)
6748          continue;          continue;
6749          }          }
6750    
6751        if (-c == ESC_E) continue;  /* Perl ignores an orphan \E */        if (escape == ESC_E) continue;  /* Perl ignores an orphan \E */
6752    
6753        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6754        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6755    
6756        if (firstchar == REQ_UNSET && -c > ESC_b && -c < ESC_Z)        if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6757          firstchar = REQ_NONE;          firstcharflags = REQ_NONE;
6758    
6759        /* 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. */
6760    
6761        zerofirstchar = firstchar;        zerofirstchar = firstchar;
6762          zerofirstcharflags = firstcharflags;
6763        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6764          zeroreqcharflags = reqcharflags;
6765    
6766        /* \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'
6767        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
6768        -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 <
6769        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
6770        -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
6771        that is a synonym for a named back reference). */        that is a synonym for a named back reference). */
6772    
6773        if (-c == ESC_g)        if (escape == ESC_g)
6774          {          {
6775          const pcre_uchar *p;          const pcre_uchar *p;
6776          save_hwm = cd->hwm;   /* Normally this is set when '(' is read */          save_hwm = cd->hwm;   /* Normally this is set when '(' is read */
# Line 6748  for (;; ptr++) Line 6790  for (;; ptr++)
6790          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
6791            {            {
6792            BOOL is_a_number = TRUE;            BOOL is_a_number = TRUE;
6793            for (p = ptr + 1; *p != 0 && *p != terminator; p++)            for (p = ptr + 1; *p != 0 && *p != (pcre_uchar)terminator; p++)
6794              {              {
6795              if (!MAX_255(*p)) { is_a_number = FALSE; break; }              if (!MAX_255(*p)) { is_a_number = FALSE; break; }
6796              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
6797              if ((cd->ctypes[*p] & ctype_word) == 0) break;              if ((cd->ctypes[*p] & ctype_word) == 0) break;
6798              }              }
6799            if (*p != terminator)            if (*p != (pcre_uchar)terminator)
6800              {              {
6801              *errorcodeptr = ERR57;              *errorcodeptr = ERR57;
6802              break;              break;
# Line 6772  for (;; ptr++) Line 6814  for (;; ptr++)
6814    
6815          p = ptr + 2;          p = ptr + 2;
6816          while (IS_DIGIT(*p)) p++;          while (IS_DIGIT(*p)) p++;
6817          if (*p != terminator)          if (*p != (pcre_uchar)terminator)
6818            {            {
6819            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
6820            break;            break;
# Line 6784  for (;; ptr++) Line 6826  for (;; ptr++)
6826        /* \k<name> or \k'name' is a back reference by name (Perl syntax).        /* \k<name> or \k'name' is a back reference by name (Perl syntax).
6827        We also support \k{name} (.NET syntax).  */        We also support \k{name} (.NET syntax).  */
6828    
6829        if (-c == ESC_k)        if (escape == ESC_k)
6830          {          {
6831          if ((ptr[1] != CHAR_LESS_THAN_SIGN &&          if ((ptr[1] != CHAR_LESS_THAN_SIGN &&
6832            ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))            ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))
# Line 6803  for (;; ptr++) Line 6845  for (;; ptr++)
6845        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
6846        ':' later. */        ':' later. */
6847    
6848        if (-c >= ESC_REF)        if (escape < 0)
6849          {          {
6850          open_capitem *oc;          open_capitem *oc;
6851          recno = -c - ESC_REF;          recno = -escape;
6852    
6853          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6854          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6855          previous = code;          previous = code;
6856          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6857          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6833  for (;; ptr++) Line 6875  for (;; ptr++)
6875        /* So are Unicode property matches, if supported. */        /* So are Unicode property matches, if supported. */
6876    
6877  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6878        else if (-c == ESC_P || -c == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6879          {          {
6880          BOOL negated;          BOOL negated;
6881          int pdata;          unsigned int ptype = 0, pdata = 0;
6882          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
6883          if (ptype < 0) goto FAILED;            goto FAILED;
6884          previous = code;          previous = code;
6885          *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6886          *code++ = ptype;          *code++ = ptype;
6887          *code++ = pdata;          *code++ = pdata;
6888          }          }
# Line 6849  for (;; ptr++) Line 6891  for (;; ptr++)
6891        /* If Unicode properties are not supported, \X, \P, and \p are not        /* If Unicode properties are not supported, \X, \P, and \p are not
6892        allowed. */        allowed. */
6893    
6894        else if (-c == ESC_X || -c == ESC_P || -c == ESC_p)        else if (escape == ESC_X || escape == ESC_P || escape == ESC_p)
6895          {          {
6896          *errorcodeptr = ERR45;          *errorcodeptr = ERR45;
6897          goto FAILED;          goto FAILED;
# Line 6859  for (;; ptr++) Line 6901  for (;; ptr++)
6901        /* For the rest (including \X when Unicode properties are supported), we        /* For the rest (including \X when Unicode properties are supported), we
6902        can obtain the OP value by negating the escape value in the default        can obtain the OP value by negating the escape value in the default
6903        situation when PCRE_UCP is not set. When it *is* set, we substitute        situation when PCRE_UCP is not set. When it *is* set, we substitute
6904        Unicode property tests. Note that \b and \B do a one-character        Unicode property tests. Note that \b and \B do a one-character
6905        lookbehind. */        lookbehind. */
6906    
6907        else        else
6908          {          {
6909          if ((-c == ESC_b || -c == ESC_B) && cd->max_lookbehind == 0)          if ((escape == ESC_b || escape == ESC_B) && cd->max_lookbehind == 0)
6910            cd->max_lookbehind = 1;            cd->max_lookbehind = 1;
6911  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6912          if (-c >= ESC_DU && -c <= ESC_wu)          if (escape >= ESC_DU && escape <= ESC_wu)
6913            {            {
6914            nestptr = ptr + 1;                   /* Where to resume */            nestptr = ptr + 1;                   /* Where to resume */
6915            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */            ptr = substitutes[escape - ESC_DU] - 1;  /* Just before substitute */
6916            }            }
6917          else          else
6918  #endif  #endif
# Line 6878  for (;; ptr++) Line 6920  for (;; ptr++)
6920          so that it works in DFA mode and in lookbehinds. */          so that it works in DFA mode and in lookbehinds. */
6921    
6922            {            {
6923            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (escape > ESC_b && escape < ESC_Z)? code : NULL;
6924            *code++ = (!utf && c == -ESC_C)? OP_ALLANY : -c;            *code++ = (!utf && escape == ESC_C)? OP_ALLANY : escape;
6925            }            }
6926          }          }
6927        continue;        continue;
# Line 6889  for (;; ptr++) Line 6931  for (;; ptr++)
6931      a value > 127. We set its representation in the length/buffer, and then      a value > 127. We set its representation in the length/buffer, and then
6932      handle it as a data character. */      handle it as a data character. */
6933    
6934  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
6935      if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)      if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
6936        mclength = PRIV(ord2utf)(c, mcbuffer);        mclength = PRIV(ord2utf)(c, mcbuffer);
6937      else      else
# Line 6922  for (;; ptr++) Line 6964  for (;; ptr++)
6964    
6965      ONE_CHAR:      ONE_CHAR:
6966      previous = code;      previous = code;
6967    
6968        /* For caseless UTF-8 mode when UCP support is available, check whether
6969        this character has more than one other case. If so, generate a special
6970        OP_PROP item instead of OP_CHARI. */
6971    
6972    #ifdef SUPPORT_UCP
6973        if (utf && (options & PCRE_CASELESS) != 0)
6974          {
6975          GETCHAR(c, mcbuffer);
6976          if ((c = UCD_CASESET(c)) != 0)
6977            {
6978            *code++ = OP_PROP;
6979            *code++ = PT_CLIST;
6980            *code++ = c;
6981            if (firstcharflags == REQ_UNSET) firstcharflags = zerofirstcharflags = REQ_NONE;
6982            break;
6983            }
6984          }
6985    #endif
6986    
6987        /* Caseful matches, or not one of the multicase characters. */
6988    
6989      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;
6990      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
6991    
# Line 6935  for (;; ptr++) Line 6999  for (;; ptr++)
6999      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
7000      repeat. */      repeat. */
7001    
7002      if (firstchar == REQ_UNSET)      if (firstcharflags == REQ_UNSET)
7003        {        {
7004        zerofirstchar = REQ_NONE;        zerofirstcharflags = REQ_NONE;
7005        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7006          zeroreqcharflags = reqcharflags;
7007    
7008        /* 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
7009        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
# Line 6946  for (;; ptr++) Line 7011  for (;; ptr++)
7011        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7012          {          {
7013          firstchar = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
7014          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;          firstchar = mcbuffer[0];
7015            firstcharflags = req_caseopt;
7016    
7017            if (mclength != 1)
7018              {
7019              reqchar = code[-1];
7020              reqcharflags = cd->req_varyopt;
7021              }
7022          }          }
7023        else firstchar = reqchar = REQ_NONE;        else firstcharflags = reqcharflags = REQ_NONE;
7024        }        }
7025    
7026      /* 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 6957  for (;; ptr++) Line 7029  for (;; ptr++)
7029      else      else
7030        {        {
7031        zerofirstchar = firstchar;        zerofirstchar = firstchar;
7032          zerofirstcharflags = firstcharflags;
7033        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7034          zeroreqcharflags = reqcharflags;
7035        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7036          reqchar = code[-1] | req_caseopt | cd->req_varyopt;          {
7037            reqchar = code[-1];
7038            reqcharflags = req_caseopt | cd->req_varyopt;
7039            }
7040        }        }
7041    
7042      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6978  return FALSE; Line 7055  return FALSE;
7055    
7056    
7057    
   
7058  /*************************************************  /*************************************************
7059  *     Compile sequence of alternatives           *  *     Compile sequence of alternatives           *
7060  *************************************************/  *************************************************/
# Line 6999  Arguments: Line 7075  Arguments:
7075    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
7076    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
7077    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
7078    firstcharptr   place to put the first required character, or a negative number    firstcharptr    place to put the first required character
7079    reqcharptr     place to put the last required character, or a negative number    firstcharflagsptr place to put the first character flags, or a negative number
7080      reqcharptr     place to put the last required character
7081      reqcharflagsptr place to put the last required character flags, or a negative number
7082    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
7083    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
7084    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 7012  Returns:         TRUE on success Line 7090  Returns:         TRUE on success
7090  static BOOL  static BOOL
7091  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
7092    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
7093    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,    int cond_depth,
7094      pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
7095      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
7096    branch_chain *bcptr, compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
7097  {  {
7098  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
# Line 7022  pcre_uchar *start_bracket = code; Line 7102  pcre_uchar *start_bracket = code;
7102  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
7103  open_capitem capitem;  open_capitem capitem;
7104  int capnumber = 0;  int capnumber = 0;
7105  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7106  pcre_int32 branchfirstchar, branchreqchar;  pcre_int32 firstcharflags, reqcharflags;
7107    pcre_uint32 branchfirstchar, branchreqchar;
7108    pcre_int32 branchfirstcharflags, branchreqcharflags;
7109  int length;  int length;
7110  int orig_bracount;  int orig_bracount;
7111  int max_bracount;  int max_bracount;
# Line 7032  branch_chain bc; Line 7114  branch_chain bc;
7114  bc.outer = bcptr;  bc.outer = bcptr;
7115  bc.current_branch = code;  bc.current_branch = code;
7116    
7117  firstchar = reqchar = REQ_UNSET;  firstchar = reqchar = 0;
7118    firstcharflags = reqcharflags = REQ_UNSET;
7119    
7120  /* 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
7121  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 7092  for (;;) Line 7175  for (;;)
7175    into the length. */    into the length. */
7176    
7177    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
7178          &branchreqchar, &bc, cond_depth, cd,          &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
7179          (lengthptr == NULL)? NULL : &length))          cond_depth, cd, (lengthptr == NULL)? NULL : &length))
7180      {      {
7181      *ptrptr = ptr;      *ptrptr = ptr;
7182      return FALSE;      return FALSE;
# Line 7114  for (;;) Line 7197  for (;;)
7197      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
7198        {        {
7199        firstchar = branchfirstchar;        firstchar = branchfirstchar;
7200          firstcharflags = branchfirstcharflags;
7201        reqchar = branchreqchar;        reqchar = branchreqchar;
7202          reqcharflags = branchreqcharflags;
7203        }        }
7204    
7205      /* 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 7128  for (;;) Line 7213  for (;;)
7213        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
7214        previously no reqchar, it takes on the value of the old firstchar. */        previously no reqchar, it takes on the value of the old firstchar. */
7215    
7216        if (firstchar >= 0 && firstchar != branchfirstchar)        if (firstcharflags >= 0 &&
7217              (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar))
7218          {          {
7219          if (reqchar < 0) reqchar = firstchar;          if (reqcharflags < 0)
7220          firstchar = REQ_NONE;            {
7221              reqchar = firstchar;
7222              reqcharflags = firstcharflags;
7223              }
7224            firstcharflags = REQ_NONE;
7225          }          }
7226    
7227        /* 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
7228        branch becomes a reqchar if there isn't a branch reqchar. */        branch becomes a reqchar if there isn't a branch reqchar. */
7229    
7230        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)        if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
7231            branchreqchar = branchfirstchar;          {
7232            branchreqchar = branchfirstchar;
7233            branchreqcharflags = branchfirstcharflags;
7234            }
7235    
7236        /* Now ensure that the reqchars match */        /* Now ensure that the reqchars match */
7237    
7238        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))        if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) ||
7239          reqchar = REQ_NONE;            reqchar != branchreqchar)
7240        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */          reqcharflags = REQ_NONE;
7241          else
7242            {
7243            reqchar = branchreqchar;
7244            reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */
7245            }
7246        }        }
7247    
7248      /* 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 7173  for (;;) Line 7271  for (;;)
7271          *ptrptr = ptr;          *ptrptr = ptr;
7272          return FALSE;          return FALSE;
7273          }          }
7274        else        else
7275          {          {
7276          if (fixed_length > cd->max_lookbehind)          if (fixed_length > cd->max_lookbehind)
7277            cd->max_lookbehind = fixed_length;            cd->max_lookbehind = fixed_length;
7278          PUT(reverse_count, 0, fixed_length);          PUT(reverse_count, 0, fixed_length);
7279          }          }
7280        }        }
7281      }      }
# Line 7240  for (;;) Line 7338  for (;;)
7338      *codeptr = code;      *codeptr = code;
7339      *ptrptr = ptr;      *ptrptr = ptr;
7340      *firstcharptr = firstchar;      *firstcharptr = firstchar;
7341        *firstcharflagsptr = firstcharflags;
7342      *reqcharptr = reqchar;      *reqcharptr = reqchar;
7343        *reqcharflagsptr = reqcharflags;
7344      if (lengthptr != NULL)      if (lengthptr != NULL)
7345        {        {
7346        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7310  and the highest back reference was great Line 7410  and the highest back reference was great
7410  However, by keeping a bitmap of the first 31 back references, we can catch some  However, by keeping a bitmap of the first 31 back references, we can catch some
7411  of the more common cases more precisely.  of the more common cases more precisely.
7412    
7413    ... A second exception is when the .* appears inside an atomic group, because
7414    this prevents the number of characters it matches from being adjusted.
7415    
7416  Arguments:  Arguments:
7417    code           points to start of expression (the bracket)    code           points to start of expression (the bracket)
7418    bracket_map    a bitmap of which brackets we are inside while testing; this    bracket_map    a bitmap of which brackets we are inside while testing; this
7419                    handles up to substring 31; after that we just have to take                    handles up to substring 31; after that we just have to take
7420                    the less precise approach                    the less precise approach
7421    backref_map    the back reference bitmap    cd             points to the compile data block
7422      atomcount      atomic group level
7423    
7424  Returns:     TRUE or FALSE  Returns:     TRUE or FALSE
7425  */  */
7426    
7427  static BOOL  static BOOL
7428  is_anchored(register const pcre_uchar *code, unsigned int bracket_map,  is_anchored(register const pcre_uchar *code, unsigned int bracket_map,
7429    unsigned int backref_map)    compile_data *cd, int atomcount)
7430  {  {
7431  do {  do {
7432     const pcre_uchar *scode = first_significant_code(     const pcre_uchar *scode = first_significant_code(
# Line 7334  do { Line 7438  do {
7438     if (op == OP_BRA  || op == OP_BRAPOS ||     if (op == OP_BRA  || op == OP_BRAPOS ||
7439         op == OP_SBRA || op == OP_SBRAPOS)         op == OP_SBRA || op == OP_SBRAPOS)
7440       {       {
7441       if (!is_anchored(scode, bracket_map, backref_map)) return FALSE;       if (!is_anchored(scode, bracket_map, cd, atomcount)) return FALSE;
7442       }       }
7443    
7444     /* Capturing brackets */     /* Capturing brackets */
# Line 7344  do { Line 7448  do {
7448       {       {
7449       int n = GET2(scode, 1+LINK_SIZE);       int n = GET2(scode, 1+LINK_SIZE);
7450       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);
7451       if (!is_anchored(scode, new_map, backref_map)) return FALSE;       if (!is_anchored(scode, new_map, cd, atomcount)) return FALSE;
7452       }       }
7453    
7454     /* Other brackets */     /* Positive forward assertions and conditions */
7455    
7456     else if (op == OP_ASSERT || op == OP_ONCE || op == OP_ONCE_NC ||     else if (op == OP_ASSERT || op == OP_COND)
             op == OP_COND)  
7457       {       {
7458       if (!is_anchored(scode, bracket_map, backref_map)) return FALSE;       if (!is_anchored(scode, bracket_map, cd, atomcount)) return FALSE;
7459         }
7460    
7461       /* Atomic groups */
7462    
7463       else if (op == OP_ONCE || op == OP_ONCE_NC)
7464         {
7465         if (!is_anchored(scode, bracket_map, cd, atomcount + 1))
7466           return FALSE;
7467       }       }
7468    
7469     /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and     /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and
7470     it isn't in brackets that are or may be referenced. */     it isn't in brackets that are or may be referenced or inside an atomic
7471       group. */
7472    
7473     else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR ||     else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR ||
7474               op == OP_TYPEPOSSTAR))               op == OP_TYPEPOSSTAR))
7475       {       {
7476       if (scode[1] != OP_ALLANY || (bracket_map & backref_map) != 0)       if (scode[1] != OP_ALLANY || (bracket_map & cd->backref_map) != 0 ||
7477             atomcount > 0 || cd->had_pruneorskip)
7478         return FALSE;         return FALSE;
7479       }       }
7480    
7481     /* Check for explicit anchoring */     /* Check for explicit anchoring */
7482    
7483     else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE;     else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE;
7484    
7485     code += GET(code, 1);     code += GET(code, 1);
7486     }     }
7487  while (*code == OP_ALT);   /* Loop for each alternative */  while (*code == OP_ALT);   /* Loop for each alternative */
# Line 7385  return TRUE; Line 7499  return TRUE;
7499  matching and for non-DOTALL patterns that start with .* (which must start at  matching and for non-DOTALL patterns that start with .* (which must start at
7500  the beginning or after \n). As in the case of is_anchored() (see above), we  the beginning or after \n). As in the case of is_anchored() (see above), we
7501  have to take account of back references to capturing brackets that contain .*  have to take account of back references to capturing brackets that contain .*
7502  because in that case we can't make the assumption.  because in that case we can't make the assumption. Also, the appearance of .*
7503    inside atomic brackets or in a pattern that contains *PRUNE or *SKIP does not
7504    count, because once again the assumption no longer holds.
7505    
7506  Arguments:  Arguments:
7507    code           points to start of expression (the bracket)    code           points to start of expression (the bracket)
7508    bracket_map    a bitmap of which brackets we are inside while testing; this    bracket_map    a bitmap of which brackets we are inside while testing; this
7509                    handles up to substring 31; after that we just have to take                    handles up to substring 31; after that we just have to take
7510                    the less precise approach                    the less precise approach
7511    backref_map    the back reference bitmap    cd             points to the compile data
7512      atomcount      atomic group level
7513    
7514  Returns:         TRUE or FALSE  Returns:         TRUE or FALSE
7515  */  */
7516    
7517  static BOOL  static BOOL
7518  is_startline(const pcre_uchar *code, unsigned int bracket_map,  is_startline(const pcre_uchar *code, unsigned int bracket_map,
7519    unsigned int backref_map)    compile_data *cd, int atomcount)
7520  {  {
7521  do {  do {
7522     const pcre_uchar *scode = first_significant_code(     const pcre_uchar *scode = first_significant_code(
# Line 7425  do { Line 7542  do {
7542         return FALSE;         return FALSE;
7543    
7544         default:     /* Assertion */         default:     /* Assertion */
7545         if (!is_startline(scode, bracket_map, backref_map)) return FALSE;         if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7546         do scode += GET(scode, 1); while (*scode == OP_ALT);         do scode += GET(scode, 1); while (*scode == OP_ALT);
7547         scode += 1 + LINK_SIZE;         scode += 1 + LINK_SIZE;
7548         break;         break;
# Line 7439  do { Line 7556  do {
7556     if (op == OP_BRA  || op == OP_BRAPOS ||     if (op == OP_BRA  || op == OP_BRAPOS ||
7557         op == OP_SBRA || op == OP_SBRAPOS)         op == OP_SBRA || op == OP_SBRAPOS)
7558       {       {
7559       if (!is_startline(scode, bracket_map, backref_map)) return FALSE;       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7560       }       }
7561    
7562     /* Capturing brackets */     /* Capturing brackets */
# Line 7449  do { Line 7566  do {
7566       {       {
7567       int n = GET2(scode, 1+LINK_SIZE);       int n = GET2(scode, 1+LINK_SIZE);
7568       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);
7569       if (!is_startline(scode, new_map, backref_map)) return FALSE;       if (!is_startline(scode, new_map, cd, atomcount)) return FALSE;
7570       }       }
7571    
7572     /* Other brackets */     /* Positive forward assertions */
7573    
7574       else if (op == OP_ASSERT)
7575         {
7576         if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7577         }
7578    
7579       /* Atomic brackets */
7580    
7581     else if (op == OP_ASSERT || op == OP_ONCE || op == OP_ONCE_NC)     else if (op == OP_ONCE || op == OP_ONCE_NC)
7582       {       {
7583       if (!is_startline(scode, bracket_map, backref_map)) return FALSE;       if (!is_startline(scode, bracket_map, cd, atomcount + 1)) return FALSE;
7584       }       }
7585    
7586     /* .* means "start at start or after \n" if it isn't in brackets that     /* .* means "start at start or after \n" if it isn't in atomic brackets or
7587     may be referenced. */     brackets that may be referenced, as long as the pattern does not contain
7588       *PRUNE or *SKIP, because these break the feature. Consider, for example,
7589       /.*?a(*PRUNE)b/ with the subject "aab", which matches "ab", i.e. not at the
7590       start of a line. */
7591    
7592     else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR || op == OP_TYPEPOSSTAR)     else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR || op == OP_TYPEPOSSTAR)
7593       {       {
7594       if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE;       if (scode[1] != OP_ANY || (bracket_map & cd->backref_map) != 0 ||
7595             atomcount > 0 || cd->had_pruneorskip)
7596           return FALSE;
7597       }       }
7598    
7599     /* Check for explicit circumflex */     /* Check for explicit circumflex; anything else gives a FALSE result. Note
7600       in particular that this includes atomic brackets OP_ONCE and OP_ONCE_NC
7601       because the number of characters matched by .* cannot be adjusted inside
7602       them. */
7603    
7604     else if (op != OP_CIRC && op != OP_CIRCM) return FALSE;     else if (op != OP_CIRC && op != OP_CIRCM) return FALSE;
7605    
# Line 7495  we return that char, otherwise -1. Line 7627  we return that char, otherwise -1.
7627    
7628  Arguments:  Arguments:
7629    code       points to start of expression (the bracket)    code       points to start of expression (the bracket)
7630      flags       points to the first char flags, or to REQ_NONE
7631    inassert   TRUE if in an assertion    inassert   TRUE if in an assertion
7632    
7633  Returns:     -1 or the fixed first char  Returns:     the fixed first char, or 0 with REQ_NONE in flags
7634  */  */
7635    
7636  static int  static pcre_uint32
7637  find_firstassertedchar(const pcre_uchar *code, BOOL inassert)  find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags,
7638      BOOL inassert)
7639  {  {
7640  register int c = -1;  register pcre_uint32 c = 0;
7641    int cflags = REQ_NONE;
7642    
7643    *flags = REQ_NONE;
7644  do {  do {
7645     int d;     pcre_uint32 d;
7646       int dflags;
7647     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7648               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7649     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7650       TRUE);       TRUE);
7651     register int op = *scode;     register pcre_uchar op = *scode;
7652    
7653     switch(op)     switch(op)
7654       {       {
7655       default:       default:
7656       return -1;       return 0;
7657    
7658       case OP_BRA:       case OP_BRA:
7659       case OP_BRAPOS:       case OP_BRAPOS:
# Line 7527  do { Line 7665  do {
7665       case OP_ONCE:       case OP_ONCE:
7666       case OP_ONCE_NC:       case OP_ONCE_NC:
7667       case OP_COND:       case OP_COND:
7668       if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)       d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
7669         return -1;       if (dflags < 0)
7670       if (c < 0) c = d; else if (c != d) return -1;         return 0;
7671         if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0;
7672       break;       break;
7673    
7674       case OP_EXACT:       case OP_EXACT:
# Line 7540  do { Line 7679  do {
7679       case OP_PLUS:       case OP_PLUS:
7680       case OP_MINPLUS:       case OP_MINPLUS:
7681       case OP_POSPLUS:       case OP_POSPLUS:
7682       if (!inassert) return -1;       if (!inassert) return 0;
7683       if (c < 0) c = scode[1];       if (cflags < 0) { c = scode[1]; cflags = 0; }
7684         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7685       break;       break;
7686    
7687       case OP_EXACTI:       case OP_EXACTI:
# Line 7553  do { Line 7692  do {
7692       case OP_PLUSI:       case OP_PLUSI:
7693       case OP_MINPLUSI:       case OP_MINPLUSI:
7694       case OP_POSPLUSI:       case OP_POSPLUSI:
7695       if (!inassert) return -1;       if (!inassert) return 0;
7696       if (c < 0) c = scode[1] | REQ_CASELESS;       if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
7697         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7698       break;       break;
7699       }       }
7700    
7701     code += GET(code, 1);     code += GET(code, 1);
7702     }     }
7703  while (*code == OP_ALT);  while (*code == OP_ALT);
7704    
7705    *flags = cflags;
7706  return c;  return c;
7707  }  }
7708    
# Line 7589  Returns:        pointer to compiled data Line 7730  Returns:        pointer to compiled data
7730                  with errorptr and erroroffset set                  with errorptr and erroroffset set
7731  */  */
7732    
7733  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7734  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7735  pcre_compile(const char *pattern, int options, const char **errorptr,  pcre_compile(const char *pattern, int options, const char **errorptr,
7736    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7737  #else  #elif defined COMPILE_PCRE16
7738  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
7739  pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,  pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,
7740    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7741    #elif defined COMPILE_PCRE32
7742    PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
7743    pcre32_compile(PCRE_SPTR32 pattern, int options, const char **errorptr,
7744      int *erroroffset, const unsigned char *tables)
7745  #endif  #endif
7746  {  {
7747  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7748  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7749  #else  #elif defined COMPILE_PCRE16
7750  return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7751    #elif defined COMPILE_PCRE32
7752    return pcre32_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7753  #endif  #endif
7754  }  }
7755    
7756    
7757  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7758  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7759  pcre_compile2(const char *pattern, int options, int *errorcodeptr,  pcre_compile2(const char *pattern, int options, int *errorcodeptr,
7760    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7761  #else  #elif defined COMPILE_PCRE16
7762  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
7763  pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,  pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,
7764    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7765    #elif defined COMPILE_PCRE32
7766    PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
7767    pcre32_compile2(PCRE_SPTR32 pattern, int options, int *errorcodeptr,
7768      const char **errorptr, int *erroroffset, const unsigned char *tables)
7769  #endif  #endif
7770  {  {
7771  REAL_PCRE *re;  REAL_PCRE *re;
7772  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7773  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7774    pcre_int32 firstcharflags, reqcharflags;
7775  int newline;  int newline;
7776  int errorcode = 0;  int errorcode = 0;
7777  int skipatstart = 0;  int skipatstart = 0;
# Line 7700  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7852  while (ptr[skipatstart] == CHAR_LEFT_PAR
7852    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)
7853      { skipatstart += 8; options |= PCRE_UTF16; continue; }      { skipatstart += 8; options |= PCRE_UTF16; continue; }
7854  #endif  #endif
7855    #ifdef COMPILE_PCRE32
7856      if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)
7857        { skipatstart += 8; options |= PCRE_UTF32; continue; }
7858    #endif
7859    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)
7860      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
7861    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 7728  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7884  while (ptr[skipatstart] == CHAR_LEFT_PAR
7884    else break;    else break;
7885    }    }
7886    
7887  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF(16|32) have the same value as PCRE_UTF8. */
7888  utf = (options & PCRE_UTF8) != 0;  utf = (options & PCRE_UTF8) != 0;
7889    
7890  /* Can't support UTF unless PCRE has been compiled to include the code. The  /* Can't support UTF unless PCRE has been compiled to include the code. The
# Line 7740  not used here. */ Line 7896  not used here. */
7896  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&
7897       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)
7898    {    {
7899  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7900    errorcode = ERR44;    errorcode = ERR44;
7901  #else  #elif defined COMPILE_PCRE16
7902    errorcode = ERR74;    errorcode = ERR74;
7903    #elif defined COMPILE_PCRE32
7904      errorcode = ERR77;
7905  #endif  #endif
7906    goto PCRE_EARLY_ERROR_RETURN2;    goto PCRE_EARLY_ERROR_RETURN2;
7907    }    }
# Line 7863  ptr += skipatstart; Line 8021  ptr += skipatstart;
8021  code = cworkspace;  code = cworkspace;
8022  *code = OP_BRA;  *code = OP_BRA;
8023  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8024    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8025      cd, &length);
8026  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
8027    
8028  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 7907  re->name_count = cd->names_found; Line 8066  re->name_count = cd->names_found;
8066  re->ref_count = 0;  re->ref_count = 0;
8067  re->tables = (tables == PRIV(default_tables))? NULL : tables;  re->tables = (tables == PRIV(default_tables))? NULL : tables;
8068  re->nullpad = NULL;  re->nullpad = NULL;
8069    #ifdef COMPILE_PCRE32
8070    re->dummy1 = re->dummy2 = 0;
8071    #endif
8072    
8073  /* The starting points of the name/number translation table and of the code are  /* The starting points of the name/number translation table and of the code are
8074  passed around in the compile data block. The start/end pattern and initial  passed around in the compile data block. The start/end pattern and initial
# Line 7926  cd->start_code = codestart; Line 8088  cd->start_code = codestart;
8088  cd->hwm = (pcre_uchar *)(cd->start_workspace);  cd->hwm = (pcre_uchar *)(cd->start_workspace);
8089  cd->req_varyopt = 0;  cd->req_varyopt = 0;
8090  cd->had_accept = FALSE;  cd->had_accept = FALSE;
8091    cd->had_pruneorskip = FALSE;
8092  cd->check_lookbehind = FALSE;  cd->check_lookbehind = FALSE;
8093  cd->open_caps = NULL;  cd->open_caps = NULL;
8094    
# Line 7937  ptr = (const pcre_uchar *)pattern + skip Line 8100  ptr = (const pcre_uchar *)pattern + skip
8100  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
8101  *code = OP_BRA;  *code = OP_BRA;
8102  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
8103    &firstchar, &reqchar, NULL, cd, NULL);    &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
8104  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
8105  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
8106  re->max_lookbehind = cd->max_lookbehind;  re->max_lookbehind = cd->max_lookbehind;
8107  re->flags = cd->external_flags | PCRE_MODE;  re->flags = cd->external_flags | PCRE_MODE;
8108    
8109  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept)
8110      {
8111      reqchar = 0;              /* Must disable after (*ACCEPT) */
8112      reqcharflags = REQ_NONE;
8113      }
8114    
8115  /* 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. */
8116    
# Line 7993  if (errorcode == 0 && re->top_backref > Line 8160  if (errorcode == 0 && re->top_backref >
8160    
8161  /* If there were any lookbehind assertions that contained OP_RECURSE  /* If there were any lookbehind assertions that contained OP_RECURSE
8162  (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,
8163  because they may contain forward references. Actual recursions can't be fixed  because they may contain forward references. Actual recursions cannot be fixed
8164  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
8165  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
8166  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 8049  if (errorcode != 0) Line 8216  if (errorcode != 0)
8216    }    }
8217    
8218  /* If the anchored option was not passed, set the flag if we can determine that  /* If the anchored option was not passed, set the flag if we can determine that
8219  the pattern is anchored by virtue of ^ characters or \A or anything else (such  the pattern is anchored by virtue of ^ characters or \A or anything else, such
8220  as starting with .* when DOTALL is set).  as starting with non-atomic .* when DOTALL is set and there are no occurrences
8221    of *PRUNE or *SKIP.
8222    
8223  Otherwise, if we know what the first byte has to be, save it, because that  Otherwise, if we know what the first byte has to be, save it, because that
8224  speeds up unanchored matches no end. If not, see if we can set the  speeds up unanchored matches no end. If not, see if we can set the
8225  PCRE_STARTLINE flag. This is helpful for multiline matches when all branches  PCRE_STARTLINE flag. This is helpful for multiline matches when all branches
8226  start with ^. and also when all branches start with .* for non-DOTALL matches.  start with ^. and also when all branches start with non-atomic .* for
8227  */  non-DOTALL matches when *PRUNE and SKIP are not present. */
8228    
8229  if ((re->options & PCRE_ANCHORED) == 0)  if ((re->options & PCRE_ANCHORED) == 0)
8230    {    {
8231    if (is_anchored(codestart, 0, cd->backref_map))    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
     re->options |= PCRE_ANCHORED;  
8232    else    else
8233      {      {
8234      if (firstchar < 0)      if (firstcharflags < 0)
8235        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
8236      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstcharflags >= 0)   /* Remove caseless flag for non-caseable chars */
8237        {        {
8238  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8239        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8240  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
8241        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8242    #elif defined COMPILE_PCRE32
8243          re->first_char = firstchar;
8244  #endif  #endif
8245  #endif        if ((firstcharflags & REQ_CASELESS) != 0)
       if ((firstchar & REQ_CASELESS) != 0)  
8246          {          {
8247  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8248          /* We ignore non-ASCII first chars in 8 bit mode. */          /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8098  if ((re->options & PCRE_ANCHORED) == 0) Line 8265  if ((re->options & PCRE_ANCHORED) == 0)
8265    
8266        re->flags |= PCRE_FIRSTSET;        re->flags |= PCRE_FIRSTSET;
8267        }        }
8268      else if (is_startline(codestart, 0, cd->backref_map))  
8269        re->flags |= PCRE_STARTLINE;      else if (is_startline(codestart, 0, cd, 0)) re->flags |= PCRE_STARTLINE;
8270      }      }
8271    }    }
8272    
# Line 8107  if ((re->options & PCRE_ANCHORED) == 0) Line 8274  if ((re->options & PCRE_ANCHORED) == 0)
8274  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
8275  bytes. */  bytes. */
8276    
8277  if (reqchar >= 0 &&  if (reqcharflags >= 0 &&
8278       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0))
8279    {    {
8280  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8281    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8282  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
8283    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8284    #elif defined COMPILE_PCRE32
8285      re->req_char = reqchar;
8286  #endif  #endif
8287  #endif    if ((reqcharflags & REQ_CASELESS) != 0)
   if ((reqchar & REQ_CASELESS) != 0)  
8288      {      {
8289  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8290      /* We ignore non-ASCII first chars in 8 bit mode. */      /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8167  if ((re->flags & PCRE_REQCHSET) != 0) Line 8334  if ((re->flags & PCRE_REQCHSET) != 0)
8334      else printf("Req char = \\x%02x%s\n", ch, caseless);      else printf("Req char = \\x%02x%s\n", ch, caseless);
8335    }    }
8336    
8337  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8338  pcre_printint((pcre *)re, stdout, TRUE);  pcre_printint((pcre *)re, stdout, TRUE);
8339  #else  #elif defined COMPILE_PCRE16
8340  pcre16_printint((pcre *)re, stdout, TRUE);  pcre16_printint((pcre *)re, stdout, TRUE);
8341    #elif defined COMPILE_PCRE32
8342    pcre32_printint((pcre *)re, stdout, TRUE);
8343  #endif  #endif
8344    
8345  /* This check is done here in the debugging case so that the code that  /* This check is done here in the debugging case so that the code that
# Line 8186  if (code - codestart > length) Line 8355  if (code - codestart > length)
8355    }    }
8356  #endif   /* PCRE_DEBUG */  #endif   /* PCRE_DEBUG */
8357    
8358  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8359  return (pcre *)re;  return (pcre *)re;
8360  #else  #elif defined COMPILE_PCRE16
8361  return (pcre16 *)re;  return (pcre16 *)re;
8362    #elif defined COMPILE_PCRE32
8363    return (pcre32 *)re;
8364  #endif  #endif
8365  }  }
8366    

Legend:
Removed from v.964  
changed lines
  Added in v.1133

  ViewVC Help
Powered by ViewVC 1.1.5