/[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 1045 by ph10, Sun Sep 23 16:50:00 2012 UTC revision 1352 by ph10, Mon Jul 29 15:49:21 2013 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2012 University of Cambridge             Copyright (c) 1997-2013 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 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 80  to check them every time. */ Line 80  to check them every time. */
80  /* Definitions to allow mutual recursion */  /* Definitions to allow mutual recursion */
81    
82  static int  static int
83    add_list_to_class(pcre_uint8 *, pcre_uchar **, int, compile_data *,    add_list_to_class(pcre_uint8 *, pcre_uchar **, int, compile_data *,
84      const pcre_uint32 *, unsigned int);      const pcre_uint32 *, unsigned int);
85    
86  static BOOL  static BOOL
87    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL,    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int,
88      int, int, int *, int *, branch_chain *, compile_data *, int *);      pcre_uint32 *, pcre_int32 *, pcre_uint32 *, pcre_int32 *, branch_chain *,
89        compile_data *, int *);
90    
91    
92    
# Line 121  overrun before it actually does run off Line 122  overrun before it actually does run off
122    
123  /* Private flags added to firstchar and reqchar. */  /* Private flags added to firstchar and reqchar. */
124    
125  #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */  #define REQ_CASELESS    (1 << 0)        /* Indicates caselessness */
126  #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */  #define REQ_VARY        (1 << 1)        /* Reqchar followed non-literal item */
127    /* 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 483  static const char error_texts[] = Line 487  static const char error_texts[] =
487    "a numbered reference must not be zero\0"    "a numbered reference must not be zero\0"
488    "an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)\0"    "an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)\0"
489    /* 60 */    /* 60 */
490    "(*VERB) not recognized\0"    "(*VERB) not recognized or malformed\0"
491    "number is too big\0"    "number is too big\0"
492    "subpattern name expected\0"    "subpattern name expected\0"
493    "digit expected after (?+\0"    "digit expected after (?+\0"
# Line 503  static const char error_texts[] = Line 507  static const char error_texts[] =
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"    "character value in \\u.... sequence is too large\0"
510      "invalid UTF-32 string\0"
511      "setting UTF is disabled by the application\0"
512    ;    ;
513    
514  /* 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 643  static const pcre_uint8 ebcdic_chartab[] Line 649  static const pcre_uint8 ebcdic_chartab[]
649    
650    
651    
   
652  /*************************************************  /*************************************************
653  *            Find an error text                  *  *            Find an error text                  *
654  *************************************************/  *************************************************/
# Line 663  find_error_text(int n) Line 668  find_error_text(int n)
668  const char *s = error_texts;  const char *s = error_texts;
669  for (; n > 0; n--)  for (; n > 0; n--)
670    {    {
671    while (*s++ != 0) {};    while (*s++ != CHAR_NULL) {};
672    if (*s == 0) return "Error text not found (please report)";    if (*s == CHAR_NULL) return "Error text not found (please report)";
673    }    }
674  return s;  return s;
675  }  }
# Line 747  return (*p == CHAR_RIGHT_CURLY_BRACKET); Line 752  return (*p == CHAR_RIGHT_CURLY_BRACKET);
752  *************************************************/  *************************************************/
753    
754  /* 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
755  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
756  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
757  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
758  UTF-8 is enabled, a positive value greater than 255 may be returned. On entry,  be returned in chptr.
759  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
760  sequence.  escape sequence.
761    
762  Arguments:  Arguments:
763    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
764      chptr          points to the data character
765    errorcodeptr   points to the errorcode variable    errorcodeptr   points to the errorcode variable
766    bracount       number of previous extracting brackets    bracount       number of previous extracting brackets
767    options        the options bits    options        the options bits
768    isclass        TRUE if inside a character class    isclass        TRUE if inside a character class
769    
770  Returns:         zero or positive => a data character  Returns:         zero => a data character
771                   negative => a special escape sequence                   positive => a special escape sequence
772                     negative => a back reference
773                   on error, errorcodeptr is set                   on error, errorcodeptr is set
774  */  */
775    
776  static int  static int
777  check_escape(const pcre_uchar **ptrptr, int *errorcodeptr, int bracount,  check_escape(const pcre_uchar **ptrptr, pcre_uint32 *chptr, int *errorcodeptr,
778    int options, BOOL isclass)    int bracount, int options, BOOL isclass)
779  {  {
780  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
781  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
782  const pcre_uchar *ptr = *ptrptr + 1;  const pcre_uchar *ptr = *ptrptr + 1;
783  pcre_int32 c;  pcre_uint32 c;
784    int escape = 0;
785  int i;  int i;
786    
787  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */
# Line 781  ptr--;                            /* Set Line 789  ptr--;                            /* Set
789    
790  /* If backslash is at the end of the pattern, it's an error. */  /* If backslash is at the end of the pattern, it's an error. */
791    
792  if (c == 0) *errorcodeptr = ERR1;  if (c == CHAR_NULL) *errorcodeptr = ERR1;
793    
794  /* Non-alphanumerics are literals. For digits or letters, do an initial lookup  /* Non-alphanumerics are literals. For digits or letters, do an initial lookup
795  in a table. A non-zero result is something that can be returned immediately.  in a table. A non-zero result is something that can be returned immediately.
# Line 790  Otherwise further processing may be requ Line 798  Otherwise further processing may be requ
798  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
799  /* Not alphanumeric */  /* Not alphanumeric */
800  else if (c < CHAR_0 || c > CHAR_z) {}  else if (c < CHAR_0 || c > CHAR_z) {}
801  else if ((i = escapes[c - CHAR_0]) != 0) c = i;  else if ((i = escapes[c - CHAR_0]) != 0)
802      { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
803    
804  #else           /* EBCDIC coding */  #else           /* EBCDIC coding */
805  /* Not alphanumeric */  /* Not alphanumeric */
806  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}
807  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; }
808  #endif  #endif
809    
810  /* Escapes that need further processing, or are illegal. */  /* Escapes that need further processing, or are illegal. */
# Line 803  else if ((i = escapes[c - 0x48]) != 0) Line 812  else if ((i = escapes[c - 0x48]) != 0)
812  else  else
813    {    {
814    const pcre_uchar *oldptr;    const pcre_uchar *oldptr;
815    BOOL braced, negated;    BOOL braced, negated, overflow;
816      int s;
817    
818    switch (c)    switch (c)
819      {      {
# Line 828  else Line 838  else
838          c = 0;          c = 0;
839          for (i = 0; i < 4; ++i)          for (i = 0; i < 4; ++i)
840            {            {
841            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
842  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
843            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
844            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 838  else Line 848  else
848  #endif  #endif
849            }            }
850    
851  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
852          if (c > (utf ? 0x10ffff : 0xff))          if (c > (utf ? 0x10ffffU : 0xffU))
853  #else  #elif defined COMPILE_PCRE16
854  #ifdef COMPILE_PCRE16          if (c > (utf ? 0x10ffffU : 0xffffU))
855          if (c > (utf ? 0x10ffff : 0xffff))  #elif defined COMPILE_PCRE32
856  #endif          if (utf && c > 0x10ffffU)
857  #endif  #endif
858            {            {
859            *errorcodeptr = ERR76;            *errorcodeptr = ERR76;
# Line 875  else Line 885  else
885      (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
886      number either in angle brackets or in single quotes. However, these are      number either in angle brackets or in single quotes. However, these are
887      (possibly recursive) subroutine calls, _not_ backreferences. Just return      (possibly recursive) subroutine calls, _not_ backreferences. Just return
888      the -ESC_g code (cf \k). */      the ESC_g code (cf \k). */
889    
890      case CHAR_g:      case CHAR_g:
891      if (isclass) break;      if (isclass) break;
892      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)
893        {        {
894        c = -ESC_g;        escape = ESC_g;
895        break;        break;
896        }        }
897    
# Line 890  else Line 900  else
900      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
901        {        {
902        const pcre_uchar *p;        const pcre_uchar *p;
903        for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++)        for (p = ptr+2; *p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET; p++)
904          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;
905        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)        if (*p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET)
906          {          {
907          c = -ESC_k;          escape = ESC_k;
908          break;          break;
909          }          }
910        braced = TRUE;        braced = TRUE;
# Line 910  else Line 920  else
920      else negated = FALSE;      else negated = FALSE;
921    
922      /* The integer range is limited by the machine's int representation. */      /* The integer range is limited by the machine's int representation. */
923      c = 0;      s = 0;
924        overflow = FALSE;
925      while (IS_DIGIT(ptr[1]))      while (IS_DIGIT(ptr[1]))
926        {        {
927        if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */        if (s > INT_MAX / 10 - 1) /* Integer overflow */
928          {          {
929          c = -1;          overflow = TRUE;
930          break;          break;
931          }          }
932        c = c * 10 + *(++ptr) - CHAR_0;        s = s * 10 + (int)(*(++ptr) - CHAR_0);
933        }        }
934      if (((unsigned int)c) > INT_MAX) /* Integer overflow */      if (overflow) /* Integer overflow */
935        {        {
936        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
937          ptr++;          ptr++;
# Line 934  else Line 945  else
945        break;        break;
946        }        }
947    
948      if (c == 0)      if (s == 0)
949        {        {
950        *errorcodeptr = ERR58;        *errorcodeptr = ERR58;
951        break;        break;
# Line 942  else Line 953  else
953    
954      if (negated)      if (negated)
955        {        {
956        if (c > bracount)        if (s > bracount)
957          {          {
958          *errorcodeptr = ERR15;          *errorcodeptr = ERR15;
959          break;          break;
960          }          }
961        c = bracount - (c - 1);        s = bracount - (s - 1);
962        }        }
963    
964      c = -(ESC_REF + c);      escape = -s;
965      break;      break;
966    
967      /* The handling of escape sequences consisting of a string of digits      /* The handling of escape sequences consisting of a string of digits
# Line 972  else Line 983  else
983        {        {
984        oldptr = ptr;        oldptr = ptr;
985        /* The integer range is limited by the machine's int representation. */        /* The integer range is limited by the machine's int representation. */
986        c -= CHAR_0;        s = (int)(c -CHAR_0);
987          overflow = FALSE;
988        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
989          {          {
990          if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */          if (s > INT_MAX / 10 - 1) /* Integer overflow */
991            {            {
992            c = -1;            overflow = TRUE;
993            break;            break;
994            }            }
995          c = c * 10 + *(++ptr) - CHAR_0;          s = s * 10 + (int)(*(++ptr) - CHAR_0);
996          }          }
997        if (((unsigned int)c) > INT_MAX) /* Integer overflow */        if (overflow) /* Integer overflow */
998          {          {
999          while (IS_DIGIT(ptr[1]))          while (IS_DIGIT(ptr[1]))
1000            ptr++;            ptr++;
1001          *errorcodeptr = ERR61;          *errorcodeptr = ERR61;
1002          break;          break;
1003          }          }
1004        if (c < 10 || c <= bracount)        if (s < 10 || s <= bracount)
1005          {          {
1006          c = -(ESC_REF + c);          escape = -s;
1007          break;          break;
1008          }          }
1009        ptr = oldptr;      /* Put the pointer back and fall through */        ptr = oldptr;      /* Put the pointer back and fall through */
# Line 1038  else Line 1050  else
1050          c = 0;          c = 0;
1051          for (i = 0; i < 2; ++i)          for (i = 0; i < 2; ++i)
1052            {            {
1053            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
1054  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1055            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1056            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 1056  else Line 1068  else
1068        const pcre_uchar *pt = ptr + 2;        const pcre_uchar *pt = ptr + 2;
1069    
1070        c = 0;        c = 0;
1071          overflow = FALSE;
1072        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)
1073          {          {
1074          register int cc = *pt++;          register pcre_uint32 cc = *pt++;
1075          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */
1076    
1077    #ifdef COMPILE_PCRE32
1078            if (c >= 0x10000000l) { overflow = TRUE; break; }
1079    #endif
1080    
1081  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1082          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1083          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 1069  else Line 1086  else
1086          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
1087  #endif  #endif
1088    
1089  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1090          if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }          if (c > (utf ? 0x10ffffU : 0xffU)) { overflow = TRUE; break; }
1091  #else  #elif defined COMPILE_PCRE16
1092  #ifdef COMPILE_PCRE16          if (c > (utf ? 0x10ffffU : 0xffffU)) { overflow = TRUE; break; }
1093          if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }  #elif defined COMPILE_PCRE32
1094  #endif          if (utf && c > 0x10ffffU) { overflow = TRUE; break; }
1095  #endif  #endif
1096          }          }
1097    
1098        if (c < 0)        if (overflow)
1099          {          {
1100          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;
1101          *errorcodeptr = ERR34;          *errorcodeptr = ERR34;
# Line 1100  else Line 1117  else
1117      c = 0;      c = 0;
1118      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)
1119        {        {
1120        int cc;                                  /* Some compilers don't like */        pcre_uint32 cc;                          /* Some compilers don't like */
1121        cc = *(++ptr);                           /* ++ in initializers */        cc = *(++ptr);                           /* ++ in initializers */
1122  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1123        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */
# Line 1119  else Line 1136  else
1136    
1137      case CHAR_c:      case CHAR_c:
1138      c = *(++ptr);      c = *(++ptr);
1139      if (c == 0)      if (c == CHAR_NULL)
1140        {        {
1141        *errorcodeptr = ERR2;        *errorcodeptr = ERR2;
1142        break;        break;
# Line 1159  else Line 1176  else
1176  newline". PCRE does not support \N{name}. However, it does support  newline". PCRE does not support \N{name}. However, it does support
1177  quantification such as \N{2,3}. */  quantification such as \N{2,3}. */
1178    
1179  if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&  if (escape == ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&
1180       !is_counted_repeat(ptr+2))       !is_counted_repeat(ptr+2))
1181    *errorcodeptr = ERR37;    *errorcodeptr = ERR37;
1182    
1183  /* If PCRE_UCP is set, we change the values for \d etc. */  /* If PCRE_UCP is set, we change the values for \d etc. */
1184    
1185  if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w)  if ((options & PCRE_UCP) != 0 && escape >= ESC_D && escape <= ESC_w)
1186    c -= (ESC_DU - ESC_D);    escape += (ESC_DU - ESC_D);
1187    
1188  /* Set the pointer to the final character before returning. */  /* Set the pointer to the final character before returning. */
1189    
1190  *ptrptr = ptr;  *ptrptr = ptr;
1191  return c;  *chptr = c;
1192    return escape;
1193  }  }
1194    
   
   
1195  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1196  /*************************************************  /*************************************************
1197  *               Handle \P and \p                 *  *               Handle \P and \p                 *
# Line 1189  escape sequence. Line 1205  escape sequence.
1205  Argument:  Argument:
1206    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
1207    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
1208    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
1209      pdataptr       points to an unsigned int that is set to the detailed property value
1210    errorcodeptr   points to the error code variable    errorcodeptr   points to the error code variable
1211    
1212  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
1213  */  */
1214    
1215  static int  static BOOL
1216  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr)  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, unsigned int *ptypeptr,
1217      unsigned int *pdataptr, int *errorcodeptr)
1218  {  {
1219  int c, i, bot, top;  pcre_uchar c;
1220    int i, bot, top;
1221  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
1222  pcre_uchar name[32];  pcre_uchar name[32];
1223    
1224  c = *(++ptr);  c = *(++ptr);
1225  if (c == 0) goto ERROR_RETURN;  if (c == CHAR_NULL) goto ERROR_RETURN;
1226    
1227  *negptr = FALSE;  *negptr = FALSE;
1228    
# Line 1220  if (c == CHAR_LEFT_CURLY_BRACKET) Line 1239  if (c == CHAR_LEFT_CURLY_BRACKET)
1239    for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)    for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)
1240      {      {
1241      c = *(++ptr);      c = *(++ptr);
1242      if (c == 0) goto ERROR_RETURN;      if (c == CHAR_NULL) goto ERROR_RETURN;
1243      if (c == CHAR_RIGHT_CURLY_BRACKET) break;      if (c == CHAR_RIGHT_CURLY_BRACKET) break;
1244      name[i] = c;      name[i] = c;
1245      }      }
# Line 1245  top = PRIV(utt_size); Line 1264  top = PRIV(utt_size);
1264    
1265  while (bot < top)  while (bot < top)
1266    {    {
1267      int r;
1268    i = (bot + top) >> 1;    i = (bot + top) >> 1;
1269    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);
1270    if (c == 0)    if (r == 0)
1271      {      {
1272      *dptr = PRIV(utt)[i].value;      *ptypeptr = PRIV(utt)[i].type;
1273      return PRIV(utt)[i].type;      *pdataptr = PRIV(utt)[i].value;
1274        return TRUE;
1275      }      }
1276    if (c > 0) bot = i + 1; else top = i;    if (r > 0) bot = i + 1; else top = i;
1277    }    }
1278    
1279  *errorcodeptr = ERR47;  *errorcodeptr = ERR47;
1280  *ptrptr = ptr;  *ptrptr = ptr;
1281  return -1;  return FALSE;
1282    
1283  ERROR_RETURN:  ERROR_RETURN:
1284  *errorcodeptr = ERR46;  *errorcodeptr = ERR46;
1285  *ptrptr = ptr;  *ptrptr = ptr;
1286  return -1;  return FALSE;
1287  }  }
1288  #endif  #endif
1289    
# Line 1297  int max = -1; Line 1318  int max = -1;
1318  /* 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
1319  an integer overflow. */  an integer overflow. */
1320    
1321  while (IS_DIGIT(*p)) min = min * 10 + *p++ - CHAR_0;  while (IS_DIGIT(*p)) min = min * 10 + (int)(*p++ - CHAR_0);
1322  if (min < 0 || min > 65535)  if (min < 0 || min > 65535)
1323    {    {
1324    *errorcodeptr = ERR5;    *errorcodeptr = ERR5;
# Line 1312  if (*p == CHAR_RIGHT_CURLY_BRACKET) max Line 1333  if (*p == CHAR_RIGHT_CURLY_BRACKET) max
1333    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
1334      {      {
1335      max = 0;      max = 0;
1336      while(IS_DIGIT(*p)) max = max * 10 + *p++ - CHAR_0;      while(IS_DIGIT(*p)) max = max * 10 + (int)(*p++ - CHAR_0);
1337      if (max < 0 || max > 65535)      if (max < 0 || max > 65535)
1338        {        {
1339        *errorcodeptr = ERR5;        *errorcodeptr = ERR5;
# Line 1367  Arguments: Line 1388  Arguments:
1388    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1389    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1390    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1391    utf          TRUE if we are in UTF-8 / UTF-16 mode    utf          TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
1392    count        pointer to the current capturing subpattern number (updated)    count        pointer to the current capturing subpattern number (updated)
1393    
1394  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 1389  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1410  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1410    {    {
1411    /* Handle specials such as (*SKIP) or (*UTF8) etc. */    /* Handle specials such as (*SKIP) or (*UTF8) etc. */
1412    
1413    if (ptr[1] == CHAR_ASTERISK) ptr += 2;    if (ptr[1] == CHAR_ASTERISK)
1414        {
1415        ptr += 2;
1416        while (ptr < cd->end_pattern && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
1417        }
1418    
1419    /* Handle a normal, unnamed capturing parenthesis. */    /* Handle a normal, unnamed capturing parenthesis. */
1420    
# Line 1413  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1438  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1438    
1439    else if (ptr[2] == CHAR_NUMBER_SIGN)    else if (ptr[2] == CHAR_NUMBER_SIGN)
1440      {      {
1441      for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break;      for (ptr += 3; *ptr != CHAR_NULL; ptr++)
1442          if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
1443      goto FAIL_EXIT;      goto FAIL_EXIT;
1444      }      }
1445    
# Line 1426  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1452  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1452      ptr += 2;      ptr += 2;
1453      if (ptr[1] != CHAR_QUESTION_MARK)      if (ptr[1] != CHAR_QUESTION_MARK)
1454        {        {
1455        while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;        while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
1456        if (*ptr != 0) ptr++;        if (*ptr != CHAR_NULL) ptr++;
1457        }        }
1458      }      }
1459    
# Line 1443  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1469  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1469      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&
1470          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)
1471        {        {
1472        int term;        pcre_uchar term;
1473        const pcre_uchar *thisname;        const pcre_uchar *thisname;
1474        *count += 1;        *count += 1;
1475        if (name == NULL && *count == lorn) return *count;        if (name == NULL && *count == lorn) return *count;
# Line 1451  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1477  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1477        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;
1478        thisname = ptr;        thisname = ptr;
1479        while (*ptr != term) ptr++;        while (*ptr != term) ptr++;
1480        if (name != NULL && lorn == ptr - thisname &&        if (name != NULL && lorn == (int)(ptr - thisname) &&
1481            STRNCMP_UC_UC(name, thisname, lorn) == 0)            STRNCMP_UC_UC(name, thisname, (unsigned int)lorn) == 0)
1482          return *count;          return *count;
1483        term++;        term++;
1484        }        }
# Line 1470  for (; ptr < cd->end_pattern; ptr++) Line 1496  for (; ptr < cd->end_pattern; ptr++)
1496    
1497    if (*ptr == CHAR_BACKSLASH)    if (*ptr == CHAR_BACKSLASH)
1498      {      {
1499      if (*(++ptr) == 0) goto FAIL_EXIT;      if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
1500      if (*ptr == CHAR_Q) for (;;)      if (*ptr == CHAR_Q) for (;;)
1501        {        {
1502        while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};        while (*(++ptr) != CHAR_NULL && *ptr != CHAR_BACKSLASH) {};
1503        if (*ptr == 0) goto FAIL_EXIT;        if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1504        if (*(++ptr) == CHAR_E) break;        if (*(++ptr) == CHAR_E) break;
1505        }        }
1506      continue;      continue;
# Line 1518  for (; ptr < cd->end_pattern; ptr++) Line 1544  for (; ptr < cd->end_pattern; ptr++)
1544    
1545      while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)      while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)
1546        {        {
1547        if (*ptr == 0) return -1;        if (*ptr == CHAR_NULL) return -1;
1548        if (*ptr == CHAR_BACKSLASH)        if (*ptr == CHAR_BACKSLASH)
1549          {          {
1550          if (*(++ptr) == 0) goto FAIL_EXIT;          if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
1551          if (*ptr == CHAR_Q) for (;;)          if (*ptr == CHAR_Q) for (;;)
1552            {            {
1553            while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};            while (*(++ptr) != CHAR_NULL && *ptr != CHAR_BACKSLASH) {};
1554            if (*ptr == 0) goto FAIL_EXIT;            if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1555            if (*(++ptr) == CHAR_E) break;            if (*(++ptr) == CHAR_E) break;
1556            }            }
1557          continue;          continue;
# Line 1539  for (; ptr < cd->end_pattern; ptr++) Line 1565  for (; ptr < cd->end_pattern; ptr++)
1565    if (xmode && *ptr == CHAR_NUMBER_SIGN)    if (xmode && *ptr == CHAR_NUMBER_SIGN)
1566      {      {
1567      ptr++;      ptr++;
1568      while (*ptr != 0)      while (*ptr != CHAR_NULL)
1569        {        {
1570        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
1571        ptr++;        ptr++;
# Line 1547  for (; ptr < cd->end_pattern; ptr++) Line 1573  for (; ptr < cd->end_pattern; ptr++)
1573        if (utf) FORWARDCHAR(ptr);        if (utf) FORWARDCHAR(ptr);
1574  #endif  #endif
1575        }        }
1576      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1577      continue;      continue;
1578      }      }
1579    
# Line 1557  for (; ptr < cd->end_pattern; ptr++) Line 1583  for (; ptr < cd->end_pattern; ptr++)
1583      {      {
1584      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);
1585      if (rc > 0) return rc;      if (rc > 0) return rc;
1586      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1587      }      }
1588    
1589    else if (*ptr == CHAR_RIGHT_PARENTHESIS)    else if (*ptr == CHAR_RIGHT_PARENTHESIS)
# Line 1601  Arguments: Line 1627  Arguments:
1627    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1628    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1629    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1630    utf          TRUE if we are in UTF-8 / UTF-16 mode    utf          TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
1631    
1632  Returns:       the number of the found subpattern, or -1 if not found  Returns:       the number of the found subpattern, or -1 if not found
1633  */  */
# Line 1622  matching closing parens. That is why we Line 1648  matching closing parens. That is why we
1648  for (;;)  for (;;)
1649    {    {
1650    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);
1651    if (rc > 0 || *ptr++ == 0) break;    if (rc > 0 || *ptr++ == CHAR_NULL) break;
1652    }    }
1653    
1654  return rc;  return rc;
# Line 1704  and doing the check at the end; a flag s Line 1730  and doing the check at the end; a flag s
1730    
1731  Arguments:  Arguments:
1732    code     points to the start of the pattern (the bracket)    code     points to the start of the pattern (the bracket)
1733    utf      TRUE in UTF-8 / UTF-16 mode    utf      TRUE in UTF-8 / UTF-16 / UTF-32 mode
1734    atend    TRUE if called when the pattern is complete    atend    TRUE if called when the pattern is complete
1735    cd       the "compile data" structure    cd       the "compile data" structure
1736    
# Line 1730  for (;;) Line 1756  for (;;)
1756    {    {
1757    int d;    int d;
1758    pcre_uchar *ce, *cs;    pcre_uchar *ce, *cs;
1759    register int op = *cc;    register pcre_uchar op = *cc;
1760    
1761    switch (op)    switch (op)
1762      {      {
# Line 1850  for (;;) Line 1876  for (;;)
1876      case OP_EXACTI:      case OP_EXACTI:
1877      case OP_NOTEXACT:      case OP_NOTEXACT:
1878      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1879      branchlength += GET2(cc,1);      branchlength += (int)GET2(cc,1);
1880      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1881  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
1882      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
# Line 1859  for (;;) Line 1885  for (;;)
1885    
1886      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1887      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1888      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)
1889          cc += 2;
1890      cc += 1 + IMM2_SIZE + 1;      cc += 1 + IMM2_SIZE + 1;
1891      break;      break;
1892    
# Line 1894  for (;;) Line 1921  for (;;)
1921    
1922      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1923    
 #if defined SUPPORT_UTF || defined COMPILE_PCRE16  
     case OP_XCLASS:  
     cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];  
     /* Fall through */  
 #endif  
   
1924      case OP_CLASS:      case OP_CLASS:
1925      case OP_NCLASS:      case OP_NCLASS:
1926    #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32
1927        case OP_XCLASS:
1928        /* The original code caused an unsigned overflow in 64 bit systems,
1929        so now we use a conditional statement. */
1930        if (op == OP_XCLASS)
1931          cc += GET(cc, 1);
1932        else
1933          cc += PRIV(OP_lengths)[OP_CLASS];
1934    #else
1935      cc += PRIV(OP_lengths)[OP_CLASS];      cc += PRIV(OP_lengths)[OP_CLASS];
1936    #endif
1937    
1938      switch (*cc)      switch (*cc)
1939        {        {
# Line 1917  for (;;) Line 1948  for (;;)
1948        case OP_CRRANGE:        case OP_CRRANGE:
1949        case OP_CRMINRANGE:        case OP_CRMINRANGE:
1950        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;
1951        branchlength += GET2(cc,1);        branchlength += (int)GET2(cc,1);
1952        cc += 1 + 2 * IMM2_SIZE;        cc += 1 + 2 * IMM2_SIZE;
1953        break;        break;
1954    
# Line 2033  length. Line 2064  length.
2064    
2065  Arguments:  Arguments:
2066    code        points to start of expression    code        points to start of expression
2067    utf         TRUE in UTF-8 / UTF-16 mode    utf         TRUE in UTF-8 / UTF-16 / UTF-32 mode
2068    number      the required bracket number or negative to find a lookbehind    number      the required bracket number or negative to find a lookbehind
2069    
2070  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 2044  PRIV(find_bracket)(const pcre_uchar *cod Line 2075  PRIV(find_bracket)(const pcre_uchar *cod
2075  {  {
2076  for (;;)  for (;;)
2077    {    {
2078    register int c = *code;    register pcre_uchar c = *code;
2079    
2080    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2081    
# Line 2067  for (;;) Line 2098  for (;;)
2098    else if (c == OP_CBRA || c == OP_SCBRA ||    else if (c == OP_CBRA || c == OP_SCBRA ||
2099             c == OP_CBRAPOS || c == OP_SCBRAPOS)             c == OP_CBRAPOS || c == OP_SCBRAPOS)
2100      {      {
2101      int n = GET2(code, 1+LINK_SIZE);      int n = (int)GET2(code, 1+LINK_SIZE);
2102      if (n == number) return (pcre_uchar *)code;      if (n == number) return (pcre_uchar *)code;
2103      code += PRIV(OP_lengths)[c];      code += PRIV(OP_lengths)[c];
2104      }      }
# Line 2097  for (;;) Line 2128  for (;;)
2128        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2129        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2130        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
2131        if (code[1 + IMM2_SIZE] == OP_PROP        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2132          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;          code += 2;
2133        break;        break;
2134    
2135        case OP_MARK:        case OP_MARK:
2136        case OP_PRUNE_ARG:        case OP_PRUNE_ARG:
2137        case OP_SKIP_ARG:        case OP_SKIP_ARG:
       code += code[1];  
       break;  
   
2138        case OP_THEN_ARG:        case OP_THEN_ARG:
2139        code += code[1];        code += code[1];
2140        break;        break;
# Line 2120  for (;;) Line 2148  for (;;)
2148    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
2149    arrange to skip the extra bytes. */    arrange to skip the extra bytes. */
2150    
2151  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2152      if (utf) switch(c)      if (utf) switch(c)
2153        {        {
2154        case OP_CHAR:        case OP_CHAR:
# Line 2172  instance of OP_RECURSE. Line 2200  instance of OP_RECURSE.
2200    
2201  Arguments:  Arguments:
2202    code        points to start of expression    code        points to start of expression
2203    utf         TRUE in UTF-8 / UTF-16 mode    utf         TRUE in UTF-8 / UTF-16 / UTF-32 mode
2204    
2205  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
2206  */  */
# Line 2182  find_recurse(const pcre_uchar *code, BOO Line 2210  find_recurse(const pcre_uchar *code, BOO
2210  {  {
2211  for (;;)  for (;;)
2212    {    {
2213    register int c = *code;    register pcre_uchar c = *code;
2214    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2215    if (c == OP_RECURSE) return code;    if (c == OP_RECURSE) return code;
2216    
# Line 2217  for (;;) Line 2245  for (;;)
2245        case OP_TYPEUPTO:        case OP_TYPEUPTO:
2246        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2247        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2248        if (code[1 + IMM2_SIZE] == OP_PROP        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2249          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;          code += 2;
2250        break;        break;
2251    
2252        case OP_MARK:        case OP_MARK:
2253        case OP_PRUNE_ARG:        case OP_PRUNE_ARG:
2254        case OP_SKIP_ARG:        case OP_SKIP_ARG:
       code += code[1];  
       break;  
   
2255        case OP_THEN_ARG:        case OP_THEN_ARG:
2256        code += code[1];        code += code[1];
2257        break;        break;
# Line 2240  for (;;) Line 2265  for (;;)
2265      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
2266      to arrange to skip the extra bytes. */      to arrange to skip the extra bytes. */
2267    
2268  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2269      if (utf) switch(c)      if (utf) switch(c)
2270        {        {
2271        case OP_CHAR:        case OP_CHAR:
# Line 2326  bracket whose current branch will alread Line 2351  bracket whose current branch will alread
2351  Arguments:  Arguments:
2352    code        points to start of search    code        points to start of search
2353    endcode     points to where to stop    endcode     points to where to stop
2354    utf         TRUE if in UTF-8 / UTF-16 mode    utf         TRUE if in UTF-8 / UTF-16 / UTF-32 mode
2355    cd          contains pointers to tables etc.    cd          contains pointers to tables etc.
2356      recurses    chain of recurse_check to catch mutual recursion
2357    
2358  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
2359  */  */
2360    
2361    typedef struct recurse_check {
2362      struct recurse_check *prev;
2363      const pcre_uchar *group;
2364    } recurse_check;
2365    
2366  static BOOL  static BOOL
2367  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,
2368    BOOL utf, compile_data *cd)    BOOL utf, compile_data *cd, recurse_check *recurses)
2369  {  {
2370  register int c;  register pcre_uchar c;
2371    recurse_check this_recurse;
2372    
2373  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
2374       code < endcode;       code < endcode;
2375       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))
# Line 2364  for (code = first_significant_code(code Line 2397  for (code = first_significant_code(code
2397    
2398    if (c == OP_RECURSE)    if (c == OP_RECURSE)
2399      {      {
2400      const pcre_uchar *scode;      const pcre_uchar *scode = cd->start_code + GET(code, 1);
2401      BOOL empty_branch;      BOOL empty_branch;
2402    
2403      /* Test for forward reference */      /* Test for forward reference or uncompleted reference. This is disabled
2404        when called to scan a completed pattern by setting cd->start_workspace to
2405        NULL. */
2406    
2407        if (cd->start_workspace != NULL)
2408          {
2409          const pcre_uchar *tcode;
2410          for (tcode = cd->start_workspace; tcode < cd->hwm; tcode += LINK_SIZE)
2411            if ((int)GET(tcode, 0) == (int)(code + 1 - cd->start_code)) return TRUE;
2412          if (GET(scode, 1) == 0) return TRUE;    /* Unclosed */
2413          }
2414    
2415        /* If we are scanning a completed pattern, there are no forward references
2416        and all groups are complete. We need to detect whether this is a recursive
2417        call, as otherwise there will be an infinite loop. If it is a recursion,
2418        just skip over it. Simple recursions are easily detected. For mutual
2419        recursions we keep a chain on the stack. */
2420    
2421      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)      else
2422        if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE;        {
2423          recurse_check *r = recurses;
2424          const pcre_uchar *endgroup = scode;
2425    
2426      /* Not a forward reference, test for completed backward reference */        do endgroup += GET(endgroup, 1); while (*endgroup == OP_ALT);
2427          if (code >= scode && code <= endgroup) continue;  /* Simple recursion */
2428    
2429      empty_branch = FALSE;        for (r = recurses; r != NULL; r = r->prev)
2430      scode = cd->start_code + GET(code, 1);          if (r->group == scode) break;
2431      if (GET(scode, 1) == 0) return TRUE;    /* Unclosed */        if (r != NULL) continue;   /* Mutual recursion */
2432          }
2433    
2434        /* Completed reference; scan the referenced group, remembering it on the
2435        stack chain to detect mutual recursions. */
2436    
2437      /* Completed backwards reference */      empty_branch = FALSE;
2438        this_recurse.prev = recurses;
2439        this_recurse.group = scode;
2440    
2441      do      do
2442        {        {
2443        if (could_be_empty_branch(scode, endcode, utf, cd))        if (could_be_empty_branch(scode, endcode, utf, cd, &this_recurse))
2444          {          {
2445          empty_branch = TRUE;          empty_branch = TRUE;
2446          break;          break;
# Line 2438  for (code = first_significant_code(code Line 2496  for (code = first_significant_code(code
2496        empty_branch = FALSE;        empty_branch = FALSE;
2497        do        do
2498          {          {
2499          if (!empty_branch && could_be_empty_branch(code, endcode, utf, cd))          if (!empty_branch && could_be_empty_branch(code, endcode, utf, cd, NULL))
2500            empty_branch = TRUE;            empty_branch = TRUE;
2501          code += GET(code, 1);          code += GET(code, 1);
2502          }          }
# Line 2496  for (code = first_significant_code(code Line 2554  for (code = first_significant_code(code
2554    
2555      /* Opcodes that must match a character */      /* Opcodes that must match a character */
2556    
2557        case OP_ANY:
2558        case OP_ALLANY:
2559        case OP_ANYBYTE:
2560    
2561      case OP_PROP:      case OP_PROP:
2562      case OP_NOTPROP:      case OP_NOTPROP:
2563        case OP_ANYNL:
2564    
2565        case OP_NOT_HSPACE:
2566        case OP_HSPACE:
2567        case OP_NOT_VSPACE:
2568        case OP_VSPACE:
2569      case OP_EXTUNI:      case OP_EXTUNI:
2570    
2571      case OP_NOT_DIGIT:      case OP_NOT_DIGIT:
2572      case OP_DIGIT:      case OP_DIGIT:
2573      case OP_NOT_WHITESPACE:      case OP_NOT_WHITESPACE:
2574      case OP_WHITESPACE:      case OP_WHITESPACE:
2575      case OP_NOT_WORDCHAR:      case OP_NOT_WORDCHAR:
2576      case OP_WORDCHAR:      case OP_WORDCHAR:
2577      case OP_ANY:  
     case OP_ALLANY:  
     case OP_ANYBYTE:  
2578      case OP_CHAR:      case OP_CHAR:
2579      case OP_CHARI:      case OP_CHARI:
2580      case OP_NOT:      case OP_NOT:
2581      case OP_NOTI:      case OP_NOTI:
2582    
2583      case OP_PLUS:      case OP_PLUS:
2584        case OP_PLUSI:
2585      case OP_MINPLUS:      case OP_MINPLUS:
2586      case OP_POSPLUS:      case OP_MINPLUSI:
2587      case OP_EXACT:  
2588      case OP_NOTPLUS:      case OP_NOTPLUS:
2589        case OP_NOTPLUSI:
2590      case OP_NOTMINPLUS:      case OP_NOTMINPLUS:
2591        case OP_NOTMINPLUSI:
2592    
2593        case OP_POSPLUS:
2594        case OP_POSPLUSI:
2595      case OP_NOTPOSPLUS:      case OP_NOTPOSPLUS:
2596        case OP_NOTPOSPLUSI:
2597    
2598        case OP_EXACT:
2599        case OP_EXACTI:
2600      case OP_NOTEXACT:      case OP_NOTEXACT:
2601        case OP_NOTEXACTI:
2602    
2603      case OP_TYPEPLUS:      case OP_TYPEPLUS:
2604      case OP_TYPEMINPLUS:      case OP_TYPEMINPLUS:
2605      case OP_TYPEPOSPLUS:      case OP_TYPEPOSPLUS:
2606      case OP_TYPEEXACT:      case OP_TYPEEXACT:
2607    
2608      return FALSE;      return FALSE;
2609    
2610      /* These are going to continue, as they may be empty, but we have to      /* These are going to continue, as they may be empty, but we have to
# Line 2543  for (code = first_significant_code(code Line 2624  for (code = first_significant_code(code
2624      case OP_TYPEUPTO:      case OP_TYPEUPTO:
2625      case OP_TYPEMINUPTO:      case OP_TYPEMINUPTO:
2626      case OP_TYPEPOSUPTO:      case OP_TYPEPOSUPTO:
2627      if (code[1 + IMM2_SIZE] == OP_PROP      if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2628        || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;        code += 2;
2629      break;      break;
2630    
2631      /* End of branch */      /* End of branch */
# Line 2557  for (code = first_significant_code(code Line 2638  for (code = first_significant_code(code
2638      return TRUE;      return TRUE;
2639    
2640      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
2641      MINUPTO, and POSUPTO may be followed by a multibyte character */      MINUPTO, and POSUPTO and their caseless and negative versions may be
2642        followed by a multibyte character. */
2643    
2644  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2645      case OP_STAR:      case OP_STAR:
2646      case OP_STARI:      case OP_STARI:
2647        case OP_NOTSTAR:
2648        case OP_NOTSTARI:
2649    
2650      case OP_MINSTAR:      case OP_MINSTAR:
2651      case OP_MINSTARI:      case OP_MINSTARI:
2652        case OP_NOTMINSTAR:
2653        case OP_NOTMINSTARI:
2654    
2655      case OP_POSSTAR:      case OP_POSSTAR:
2656      case OP_POSSTARI:      case OP_POSSTARI:
2657        case OP_NOTPOSSTAR:
2658        case OP_NOTPOSSTARI:
2659    
2660      case OP_QUERY:      case OP_QUERY:
2661      case OP_QUERYI:      case OP_QUERYI:
2662        case OP_NOTQUERY:
2663        case OP_NOTQUERYI:
2664    
2665      case OP_MINQUERY:      case OP_MINQUERY:
2666      case OP_MINQUERYI:      case OP_MINQUERYI:
2667        case OP_NOTMINQUERY:
2668        case OP_NOTMINQUERYI:
2669    
2670      case OP_POSQUERY:      case OP_POSQUERY:
2671      case OP_POSQUERYI:      case OP_POSQUERYI:
2672        case OP_NOTPOSQUERY:
2673        case OP_NOTPOSQUERYI:
2674    
2675      if (utf && HAS_EXTRALEN(code[1])) code += GET_EXTRALEN(code[1]);      if (utf && HAS_EXTRALEN(code[1])) code += GET_EXTRALEN(code[1]);
2676      break;      break;
2677    
2678      case OP_UPTO:      case OP_UPTO:
2679      case OP_UPTOI:      case OP_UPTOI:
2680        case OP_NOTUPTO:
2681        case OP_NOTUPTOI:
2682    
2683      case OP_MINUPTO:      case OP_MINUPTO:
2684      case OP_MINUPTOI:      case OP_MINUPTOI:
2685        case OP_NOTMINUPTO:
2686        case OP_NOTMINUPTOI:
2687    
2688      case OP_POSUPTO:      case OP_POSUPTO:
2689      case OP_POSUPTOI:      case OP_POSUPTOI:
2690        case OP_NOTPOSUPTO:
2691        case OP_NOTPOSUPTOI:
2692    
2693      if (utf && HAS_EXTRALEN(code[1 + IMM2_SIZE])) code += GET_EXTRALEN(code[1 + IMM2_SIZE]);      if (utf && HAS_EXTRALEN(code[1 + IMM2_SIZE])) code += GET_EXTRALEN(code[1 + IMM2_SIZE]);
2694      break;      break;
2695  #endif  #endif
# Line 2591  for (code = first_significant_code(code Line 2700  for (code = first_significant_code(code
2700      case OP_MARK:      case OP_MARK:
2701      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
2702      case OP_SKIP_ARG:      case OP_SKIP_ARG:
     code += code[1];  
     break;  
   
2703      case OP_THEN_ARG:      case OP_THEN_ARG:
2704      code += code[1];      code += code[1];
2705      break;      break;
# Line 2625  Arguments: Line 2731  Arguments:
2731    code        points to start of the recursion    code        points to start of the recursion
2732    endcode     points to where to stop (current RECURSE item)    endcode     points to where to stop (current RECURSE item)
2733    bcptr       points to the chain of current (unclosed) branch starts    bcptr       points to the chain of current (unclosed) branch starts
2734    utf         TRUE if in UTF-8 / UTF-16 mode    utf         TRUE if in UTF-8 / UTF-16 / UTF-32 mode
2735    cd          pointers to tables etc    cd          pointers to tables etc
2736    
2737  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2637  could_be_empty(const pcre_uchar *code, c Line 2743  could_be_empty(const pcre_uchar *code, c
2743  {  {
2744  while (bcptr != NULL && bcptr->current_branch >= code)  while (bcptr != NULL && bcptr->current_branch >= code)
2745    {    {
2746    if (!could_be_empty_branch(bcptr->current_branch, endcode, utf, cd))    if (!could_be_empty_branch(bcptr->current_branch, endcode, utf, cd, NULL))
2747      return FALSE;      return FALSE;
2748    bcptr = bcptr->outer;    bcptr = bcptr->outer;
2749    }    }
# Line 2691  Returns:   TRUE or FALSE Line 2797  Returns:   TRUE or FALSE
2797  static BOOL  static BOOL
2798  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)
2799  {  {
2800  int terminator;          /* Don't combine these lines; the Solaris cc */  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */
2801  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
2802  for (++ptr; *ptr != 0; ptr++)  for (++ptr; *ptr != CHAR_NULL; ptr++)
2803    {    {
2804    if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)    if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
2805      ptr++;      ptr++;
# Line 2740  register int yield = 0; Line 2846  register int yield = 0;
2846  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
2847    {    {
2848    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
2849      STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;      STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield;
2850    pn += posix_name_lengths[yield] + 1;    pn += posix_name_lengths[yield] + 1;
2851    yield++;    yield++;
2852    }    }
# Line 2772  value in the reference (which is a group Line 2878  value in the reference (which is a group
2878  Arguments:  Arguments:
2879    group      points to the start of the group    group      points to the start of the group
2880    adjust     the amount by which the group is to be moved    adjust     the amount by which the group is to be moved
2881    utf        TRUE in UTF-8 / UTF-16 mode    utf        TRUE in UTF-8 / UTF-16 / UTF-32 mode
2882    cd         contains pointers to tables etc.    cd         contains pointers to tables etc.
2883    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
2884    
# Line 2795  while ((ptr = (pcre_uchar *)find_recurse Line 2901  while ((ptr = (pcre_uchar *)find_recurse
2901    
2902    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
2903      {      {
2904      offset = GET(hc, 0);      offset = (int)GET(hc, 0);
2905      if (cd->start_code + offset == ptr + 1)      if (cd->start_code + offset == ptr + 1)
2906        {        {
2907        PUT(hc, 0, offset + adjust);        PUT(hc, 0, offset + adjust);
# Line 2808  while ((ptr = (pcre_uchar *)find_recurse Line 2914  while ((ptr = (pcre_uchar *)find_recurse
2914    
2915    if (hc >= cd->hwm)    if (hc >= cd->hwm)
2916      {      {
2917      offset = GET(ptr, 1);      offset = (int)GET(ptr, 1);
2918      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
2919      }      }
2920    
# Line 2878  PUT(previous_callout, 2 + LINK_SIZE, len Line 2984  PUT(previous_callout, 2 + LINK_SIZE, len
2984  /* 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
2985  with UCP support. It searches up the characters, looking for ranges of  with UCP support. It searches up the characters, looking for ranges of
2986  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
2987  start address. A character with multiple other cases is returned on its own  start address. A character with multiple other cases is returned on its own
2988  with a special return value.  with a special return value.
2989    
2990  Arguments:  Arguments:
# Line 2890  Arguments: Line 2996  Arguments:
2996  Yield:        -1 when no more  Yield:        -1 when no more
2997                 0 when a range is returned                 0 when a range is returned
2998                >0 the CASESET offset for char with multiple other cases                >0 the CASESET offset for char with multiple other cases
2999                  in this case, ocptr contains the original                  in this case, ocptr contains the original
3000  */  */
3001    
3002  static int  static int
3003  get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr,  get_othercase_range(pcre_uint32 *cptr, pcre_uint32 d, pcre_uint32 *ocptr,
3004    unsigned int *odptr)    pcre_uint32 *odptr)
3005  {  {
3006  unsigned int c, othercase, next;  pcre_uint32 c, othercase, next;
3007  int co;  unsigned int co;
3008    
3009  /* Find the first character that has an other case. If it has multiple other  /* Find the first character that has an other case. If it has multiple other
3010  cases, return its case offset value. */  cases, return its case offset value. */
3011    
3012  for (c = *cptr; c <= d; c++)  for (c = *cptr; c <= d; c++)
3013    {    {
3014    if ((co = UCD_CASESET(c)) != 0)    if ((co = UCD_CASESET(c)) != 0)
3015      {      {
3016      *ocptr = c++;   /* Character that has the set */      *ocptr = c++;   /* Character that has the set */
3017      *cptr = c;      /* Rest of input range */      *cptr = c;      /* Rest of input range */
3018      return co;      return (int)co;
3019      }      }
3020    if ((othercase = UCD_OTHERCASE(c)) != c) break;    if ((othercase = UCD_OTHERCASE(c)) != c) break;
3021    }    }
3022    
3023  if (c > d) return -1;  /* Reached end of range */  if (c > d) return -1;  /* Reached end of range */
# Line 2949  Returns:       TRUE if auto-possessifyin Line 3055  Returns:       TRUE if auto-possessifyin
3055  */  */
3056    
3057  static BOOL  static BOOL
3058  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)
3059  {  {
3060    #ifdef SUPPORT_UCP
3061    const pcre_uint32 *p;
3062    #endif
3063    
3064  const ucd_record *prop = GET_UCD(c);  const ucd_record *prop = GET_UCD(c);
3065    
3066  switch(ptype)  switch(ptype)
3067    {    {
3068    case PT_LAMP:    case PT_LAMP:
# Line 2989  switch(ptype) Line 3100  switch(ptype)
3100    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
3101            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
3102            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
3103    
3104    #ifdef SUPPORT_UCP
3105      case PT_CLIST:
3106      p = PRIV(ucd_caseless_sets) + prop->caseset;
3107      for (;;)
3108        {
3109        if (c < *p) return !negated;
3110        if (c == *p++) return negated;
3111        }
3112      break;  /* Control never reaches here */
3113    #endif
3114    }    }
3115    
3116  return FALSE;  return FALSE;
3117  }  }
3118  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
# Line 3006  sense to automatically possessify the re Line 3129  sense to automatically possessify the re
3129    
3130  Arguments:  Arguments:
3131    previous      pointer to the repeated opcode    previous      pointer to the repeated opcode
3132    utf           TRUE in UTF-8 / UTF-16 mode    utf           TRUE in UTF-8 / UTF-16 / UTF-32 mode
3133    ptr           next character in pattern    ptr           next character in pattern
3134    options       options bits    options       options bits
3135    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
# Line 3018  static BOOL Line 3141  static BOOL
3141  check_auto_possessive(const pcre_uchar *previous, BOOL utf,  check_auto_possessive(const pcre_uchar *previous, BOOL utf,
3142    const pcre_uchar *ptr, int options, compile_data *cd)    const pcre_uchar *ptr, int options, compile_data *cd)
3143  {  {
3144  pcre_int32 c, next;  pcre_uint32 c = NOTACHAR;
3145  int op_code = *previous++;  pcre_uint32 next;
3146    int escape;
3147    pcre_uchar op_code = *previous++;
3148    
3149  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
3150    
# Line 3031  if ((options & PCRE_EXTENDED) != 0) Line 3156  if ((options & PCRE_EXTENDED) != 0)
3156      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
3157        {        {
3158        ptr++;        ptr++;
3159        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3160          {          {
3161          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
3162          ptr++;          ptr++;
# Line 3050  value is a character, a negative value i Line 3175  value is a character, a negative value i
3175  if (*ptr == CHAR_BACKSLASH)  if (*ptr == CHAR_BACKSLASH)
3176    {    {
3177    int temperrorcode = 0;    int temperrorcode = 0;
3178    next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE);    escape = check_escape(&ptr, &next, &temperrorcode, cd->bracount, options,
3179        FALSE);
3180    if (temperrorcode != 0) return FALSE;    if (temperrorcode != 0) return FALSE;
3181    ptr++;    /* Point after the escape sequence */    ptr++;    /* Point after the escape sequence */
3182    }    }
3183  else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)  else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)
3184    {    {
3185      escape = 0;
3186  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3187    if (utf) { GETCHARINC(next, ptr); } else    if (utf) { GETCHARINC(next, ptr); } else
3188  #endif  #endif
# Line 3073  if ((options & PCRE_EXTENDED) != 0) Line 3200  if ((options & PCRE_EXTENDED) != 0)
3200      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
3201        {        {
3202        ptr++;        ptr++;
3203        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3204          {          {
3205          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
3206          ptr++;          ptr++;
# Line 3092  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3219  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3219    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)
3220      return FALSE;      return FALSE;
3221    
3222  /* 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. */  
3223    
3224  if (next >= 0) switch(op_code)  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3225        op_code == OP_NOT || op_code == OP_NOTI)
3226    {    {
   case OP_CHAR:  
3227  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3228    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
3229  #else  #else
3230    c = *previous;    c = *previous;
3231  #endif  #endif
3232    return c != next;    }
3233    
3234    /* 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
3235    Unicode property support, we can use it to test the other case of  the next item is a character. */
   high-valued characters. */  
3236    
3237    case OP_CHARI:  if (escape == 0)
3238  #ifdef SUPPORT_UTF    {
3239    GETCHARTEST(c, previous);    /* For a caseless UTF match, the next character may have more than one other
3240  #else    case, which maps to the special PT_CLIST property. Check this first. */
3241    c = *previous;  
3242    #ifdef SUPPORT_UCP
3243      if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3244        {
3245        unsigned int ocs = UCD_CASESET(next);
3246        if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);
3247        }
3248  #endif  #endif
3249    if (c == next) return FALSE;  
3250  #ifdef SUPPORT_UTF    switch(op_code)
   if (utf)  
3251      {      {
3252      unsigned int othercase;      case OP_CHAR:
3253      if (next < 128) othercase = cd->fcc[next]; else      return c != next;
3254    
3255        /* For CHARI (caseless character) we must check the other case. If we have
3256        Unicode property support, we can use it to test the other case of
3257        high-valued characters. We know that next can have only one other case,
3258        because multi-other-case characters are dealt with above. */
3259    
3260        case OP_CHARI:
3261        if (c == next) return FALSE;
3262    #ifdef SUPPORT_UTF
3263        if (utf)
3264          {
3265          pcre_uint32 othercase;
3266          if (next < 128) othercase = cd->fcc[next]; else
3267  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3268      othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE(next);
3269  #else  #else
3270      othercase = NOTACHAR;        othercase = NOTACHAR;
3271  #endif  #endif
3272      return (unsigned int)c != othercase;        return c != othercase;
3273      }        }
3274    else      else
3275  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3276    return (c != TABLE_GET((unsigned int)next, cd->fcc, next));  /* Non-UTF-8 mode */      return (c != TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3277    
3278    case OP_NOT:      case OP_NOT:
3279  #ifdef SUPPORT_UTF      return c == next;
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
 #endif  
   return c == next;  
3280    
3281    case OP_NOTI:      case OP_NOTI:
3282  #ifdef SUPPORT_UTF      if (c == next) return TRUE;
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
 #endif  
   if (c == next) return TRUE;  
3283  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3284    if (utf)      if (utf)
3285      {        {
3286      unsigned int othercase;        pcre_uint32 othercase;
3287      if (next < 128) othercase = cd->fcc[next]; else        if (next < 128) othercase = cd->fcc[next]; else
3288  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3289      othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE(next);
3290  #else  #else
3291      othercase = NOTACHAR;        othercase = NOTACHAR;
3292  #endif  #endif
3293      return (unsigned int)c == othercase;        return c == othercase;
3294      }        }
3295    else      else
3296  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3297    return (c == TABLE_GET((unsigned int)next, cd->fcc, next));  /* Non-UTF-8 mode */      return (c == TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3298    
3299    /* 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.
3300    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. */
3301    
3302    case OP_DIGIT:      case OP_DIGIT:
3303    return next > 255 || (cd->ctypes[next] & ctype_digit) == 0;      return next > 255 || (cd->ctypes[next] & ctype_digit) == 0;
3304    
3305    case OP_NOT_DIGIT:      case OP_NOT_DIGIT:
3306    return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;
3307    
3308    case OP_WHITESPACE:      case OP_WHITESPACE:
3309    return next > 255 || (cd->ctypes[next] & ctype_space) == 0;      return next > 255 || (cd->ctypes[next] & ctype_space) == 0;
3310    
3311    case OP_NOT_WHITESPACE:      case OP_NOT_WHITESPACE:
3312    return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;
3313    
3314    case OP_WORDCHAR:      case OP_WORDCHAR:
3315    return next > 255 || (cd->ctypes[next] & ctype_word) == 0;      return next > 255 || (cd->ctypes[next] & ctype_word) == 0;
3316    
3317    case OP_NOT_WORDCHAR:      case OP_NOT_WORDCHAR:
3318    return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;
3319    
3320    case OP_HSPACE:      case OP_HSPACE:
3321    case OP_NOT_HSPACE:      case OP_NOT_HSPACE:
3322    switch(next)      switch(next)
3323      {        {
3324      HSPACE_CASES:        HSPACE_CASES:
3325      return op_code == OP_NOT_HSPACE;        return op_code == OP_NOT_HSPACE;
   
     default:  
     return op_code != OP_NOT_HSPACE;  
     }  
3326    
3327    case OP_ANYNL:        default:
3328    case OP_VSPACE:        return op_code != OP_NOT_HSPACE;
3329    case OP_NOT_VSPACE:        }
3330    switch(next)  
3331      {      case OP_ANYNL:
3332      VSPACE_CASES:      case OP_VSPACE:
3333      return op_code == OP_NOT_VSPACE;      case OP_NOT_VSPACE:
3334        switch(next)
3335      default:        {
3336      return op_code != OP_NOT_VSPACE;        VSPACE_CASES:
3337      }        return op_code == OP_NOT_VSPACE;
3338    
3339          default:
3340          return op_code != OP_NOT_VSPACE;
3341          }
3342    
3343  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3344    case OP_PROP:      case OP_PROP:
3345    return check_char_prop(next, previous[0], previous[1], FALSE);      return check_char_prop(next, previous[0], previous[1], FALSE);
3346    
3347    case OP_NOTPROP:      case OP_NOTPROP:
3348    return check_char_prop(next, previous[0], previous[1], TRUE);      return check_char_prop(next, previous[0], previous[1], TRUE);
3349  #endif  #endif
3350    
3351    default:      default:
3352    return FALSE;      return FALSE;
3353        }
3354    }    }
3355    
   
3356  /* 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
3357  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
3358  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 3230  switch(op_code) Line 3363  switch(op_code)
3363    {    {
3364    case OP_CHAR:    case OP_CHAR:
3365    case OP_CHARI:    case OP_CHARI:
3366  #ifdef SUPPORT_UTF    switch(escape)
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
 #endif  
   switch(-next)  
3367      {      {
3368      case ESC_d:      case ESC_d:
3369      return c > 255 || (cd->ctypes[c] & ctype_digit) == 0;      return c > 255 || (cd->ctypes[c] & ctype_digit) == 0;
# Line 3259  switch(op_code) Line 3387  switch(op_code)
3387      case ESC_H:      case ESC_H:
3388      switch(c)      switch(c)
3389        {        {
3390        HSPACE_CASES:        HSPACE_CASES:
3391        return -next != ESC_h;        return escape != ESC_h;
3392    
3393        default:        default:
3394        return -next == ESC_h;        return escape == ESC_h;
3395        }        }
3396    
3397      case ESC_v:      case ESC_v:
3398      case ESC_V:      case ESC_V:
3399      switch(c)      switch(c)
3400        {        {
3401        VSPACE_CASES:        VSPACE_CASES:
3402        return -next != ESC_v;        return escape != ESC_v;
3403    
3404        default:        default:
3405        return -next == ESC_v;        return escape == ESC_v;
3406        }        }
3407    
3408      /* 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
3409      their substitutions and process them. The result will always be either      their substitutions and process them. The result will always be either
3410      -ESC_p or -ESC_P. Then fall through to process those values. */      ESC_p or ESC_P. Then fall through to process those values. */
3411    
3412  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3413      case ESC_du:      case ESC_du:
# Line 3290  switch(op_code) Line 3418  switch(op_code)
3418      case ESC_SU:      case ESC_SU:
3419        {        {
3420        int temperrorcode = 0;        int temperrorcode = 0;
3421        ptr = substitutes[-next - ESC_DU];        ptr = substitutes[escape - ESC_DU];
3422        next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);        escape = check_escape(&ptr, &next, &temperrorcode, 0, options, FALSE);
3423        if (temperrorcode != 0) return FALSE;        if (temperrorcode != 0) return FALSE;
3424        ptr++;    /* For compatibility */        ptr++;    /* For compatibility */
3425        }        }
# Line 3300  switch(op_code) Line 3428  switch(op_code)
3428      case ESC_p:      case ESC_p:
3429      case ESC_P:      case ESC_P:
3430        {        {
3431        int ptype, pdata, errorcodeptr;        unsigned int ptype = 0, pdata = 0;
3432          int errorcodeptr;
3433        BOOL negated;        BOOL negated;
3434    
3435        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
3436        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
3437        if (ptype < 0) return FALSE;          return FALSE;
3438        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
3439    
3440        /* 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 3318  switch(op_code) Line 3447  switch(op_code)
3447    
3448        /* Do the property check. */        /* Do the property check. */
3449    
3450        return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);        return check_char_prop(c, ptype, pdata, (escape == ESC_P) != negated);
3451        }        }
3452  #endif  #endif
3453    
# Line 3333  switch(op_code) Line 3462  switch(op_code)
3462    these op-codes are never generated.) */    these op-codes are never generated.) */
3463    
3464    case OP_DIGIT:    case OP_DIGIT:
3465    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||    return escape == ESC_D || escape == ESC_s || escape == ESC_W ||
3466           next == -ESC_h || next == -ESC_v || next == -ESC_R;           escape == ESC_h || escape == ESC_v || escape == ESC_R;
3467    
3468    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
3469    return next == -ESC_d;    return escape == ESC_d;
3470    
3471    case OP_WHITESPACE:    case OP_WHITESPACE:
3472    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return escape == ESC_S || escape == ESC_d || escape == ESC_w;
3473    
3474    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
3475    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;
3476    
3477    case OP_HSPACE:    case OP_HSPACE:
3478    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||    return escape == ESC_S || escape == ESC_H || escape == ESC_d ||
3479           next == -ESC_w || next == -ESC_v || next == -ESC_R;           escape == ESC_w || escape == ESC_v || escape == ESC_R;
3480    
3481    case OP_NOT_HSPACE:    case OP_NOT_HSPACE:
3482    return next == -ESC_h;    return escape == ESC_h;
3483    
3484    /* Can't have \S in here because VT matches \S (Perl anomaly) */    /* Can't have \S in here because VT matches \S (Perl anomaly) */
3485    case OP_ANYNL:    case OP_ANYNL:
3486    case OP_VSPACE:    case OP_VSPACE:
3487    return next == -ESC_V || next == -ESC_d || next == -ESC_w;    return escape == ESC_V || escape == ESC_d || escape == ESC_w;
3488    
3489    case OP_NOT_VSPACE:    case OP_NOT_VSPACE:
3490    return next == -ESC_v || next == -ESC_R;    return escape == ESC_v || escape == ESC_R;
3491    
3492    case OP_WORDCHAR:    case OP_WORDCHAR:
3493    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||    return escape == ESC_W || escape == ESC_s || escape == ESC_h ||
3494           next == -ESC_v || next == -ESC_R;           escape == ESC_v || escape == ESC_R;
3495    
3496    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
3497    return next == -ESC_w || next == -ESC_d;    return escape == ESC_w || escape == ESC_d;
3498    
3499    default:    default:
3500    return FALSE;    return FALSE;
# Line 3381  switch(op_code) Line 3510  switch(op_code)
3510  *************************************************/  *************************************************/
3511    
3512  /* This function packages up the logic of adding a character or range of  /* This function packages up the logic of adding a character or range of
3513  characters to a class. The character values in the arguments will be within the  characters to a class. The character values in the arguments will be within the
3514  valid values for the current mode (8-bit, 16-bit, UTF, etc). This function is  valid values for the current mode (8-bit, 16-bit, UTF, etc). This function is
3515  mutually recursive with the function immediately below.  mutually recursive with the function immediately below.
3516    
3517  Arguments:  Arguments:
3518    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3519    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3520    options       the options word    options       the options word
3521    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3522    start         start of range character    start         start of range character
3523    end           end of range character    end           end of range character
3524    
3525  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3526                  the pointer to extra data is updated                  the pointer to extra data is updated
3527  */  */
3528    
3529  static int  static int
3530  add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,  add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,
3531    compile_data *cd, unsigned int start, unsigned int end)    compile_data *cd, pcre_uint32 start, pcre_uint32 end)
3532  {  {
3533  unsigned int c;  pcre_uint32 c;
3534  int n8 = 0;  int n8 = 0;
3535    
3536  /* If caseless matching is required, scan the range and process alternate  /* If caseless matching is required, scan the range and process alternate
3537  cases. In Unicode, there are 8-bit characters that have alternate cases that  cases. In Unicode, there are 8-bit characters that have alternate cases that
3538  are greater than 255 and vice-versa. Sometimes we can just extend the original  are greater than 255 and vice-versa. Sometimes we can just extend the original
3539  range. */  range. */
3540    
3541  if ((options & PCRE_CASELESS) != 0)  if ((options & PCRE_CASELESS) != 0)
3542    {    {
3543  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3544    if ((options & PCRE_UTF8) != 0)    if ((options & PCRE_UTF8) != 0)
3545      {      {
3546      int rc;      int rc;
3547      unsigned int oc, od;      pcre_uint32 oc, od;
3548    
3549      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */
3550      c = start;      c = start;
3551    
3552      while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)      while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)
3553        {        {
3554        /* Handle a single character that has more than one other case. */        /* Handle a single character that has more than one other case. */
3555    
3556        if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,        if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,
3557          PRIV(ucd_caseless_sets) + rc, oc);          PRIV(ucd_caseless_sets) + rc, oc);
3558    
3559        /* Do nothing if the other case range is within the original range. */        /* Do nothing if the other case range is within the original range. */
3560    
3561        else if (oc >= start && od <= end) continue;        else if (oc >= start && od <= end) continue;
3562    
3563        /* Extend the original range if there is overlap, noting that if oc < c, we        /* Extend the original range if there is overlap, noting that if oc < c, we
3564        can't have od > end because a subrange is always shorter than the basic        can't have od > end because a subrange is always shorter than the basic
3565        range. Otherwise, use a recursive call to add the additional range. */        range. Otherwise, use a recursive call to add the additional range. */
3566    
3567        else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */        else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */
3568        else if (od > end && oc <= end + 1) end = od;       /* Extend upwards */        else if (od > end && oc <= end + 1) end = od;       /* Extend upwards */
3569        else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);        else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);
# Line 3444  if ((options & PCRE_CASELESS) != 0) Line 3573  if ((options & PCRE_CASELESS) != 0)
3573  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3574    
3575    /* Not UTF-mode, or no UCP */    /* Not UTF-mode, or no UCP */
3576    
3577    for (c = start; c <= end && c < 256; c++)    for (c = start; c <= end && c < 256; c++)
3578      {      {
3579      SETBIT(classbits, cd->fcc[c]);      SETBIT(classbits, cd->fcc[c]);
3580      n8++;      n8++;
3581      }      }
3582    }    }
3583    
3584  /* Now handle the original range. Adjust the final value according to the bit  /* Now handle the original range. Adjust the final value according to the bit
3585  length - this means that the same lists of (e.g.) horizontal spaces can be used  length - this means that the same lists of (e.g.) horizontal spaces can be used
3586  in all cases. */  in all cases. */
3587    
3588  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
3589  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3590    if ((options & PCRE_UTF8) == 0)    if ((options & PCRE_UTF8) == 0)
3591  #endif  #endif
3592    if (end > 0xff) end = 0xff;    if (end > 0xff) end = 0xff;
 #endif  
3593    
3594  #ifdef COMPILE_PCRE16  #elif defined COMPILE_PCRE16
3595  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3596    if ((options & PCRE_UTF16) == 0)    if ((options & PCRE_UTF16) == 0)
3597  #endif  #endif
3598    if (end > 0xffff) end = 0xffff;    if (end > 0xffff) end = 0xffff;
3599  #endif  
3600    #endif /* COMPILE_PCRE[8|16] */
3601    
3602  /* If all characters are less than 256, use the bit map. Otherwise use extra  /* If all characters are less than 256, use the bit map. Otherwise use extra
3603  data. */  data. */
# Line 3477  if (end < 0x100) Line 3606  if (end < 0x100)
3606    {    {
3607    for (c = start; c <= end; c++)    for (c = start; c <= end; c++)
3608      {      {
3609      n8++;      n8++;
3610      SETBIT(classbits, c);      SETBIT(classbits, c);
3611      }      }
3612    }    }
3613    
3614  else  else
3615    {    {
3616    pcre_uchar *uchardata = *uchardptr;    pcre_uchar *uchardata = *uchardptr;
3617    
3618  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3619    if ((options & PCRE_UTF8) != 0)  /* All UTFs use the same flag bit */    if ((options & PCRE_UTF8) != 0)  /* All UTFs use the same flag bit */
3620      {      {
3621      if (start < end)      if (start < end)
3622        {        {
3623        *uchardata++ = XCL_RANGE;        *uchardata++ = XCL_RANGE;
3624        uchardata += PRIV(ord2utf)(start, uchardata);        uchardata += PRIV(ord2utf)(start, uchardata);
3625        uchardata += PRIV(ord2utf)(end, uchardata);        uchardata += PRIV(ord2utf)(end, uchardata);
3626        }        }
3627      else if (start == end)      else if (start == end)
3628        {        {
3629        *uchardata++ = XCL_SINGLE;        *uchardata++ = XCL_SINGLE;
3630        uchardata += PRIV(ord2utf)(start, uchardata);        uchardata += PRIV(ord2utf)(start, uchardata);
3631        }        }
3632      }      }
3633    else    else
3634  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3635    
3636    /* Without UTF support, character values are constrained by the bit length,    /* Without UTF support, character values are constrained by the bit length,
3637    and can only be > 256 for 16-bit and 32-bit libraries. */    and can only be > 256 for 16-bit and 32-bit libraries. */
3638    
3639  #ifdef COMPILE_PCRE8  #ifdef COMPILE_PCRE8
3640      {}      {}
3641  #else  #else
3642    if (start < end)    if (start < end)
3643      {      {
3644      *uchardata++ = XCL_RANGE;      *uchardata++ = XCL_RANGE;
# Line 3520  else Line 3649  else
3649      {      {
3650      *uchardata++ = XCL_SINGLE;      *uchardata++ = XCL_SINGLE;
3651      *uchardata++ = start;      *uchardata++ = start;
3652      }      }
3653  #endif  #endif
3654    
3655    *uchardptr = uchardata;   /* Updata extra data pointer */    *uchardptr = uchardata;   /* Updata extra data pointer */
3656    }    }
3657    
3658  return n8;    /* Number of 8-bit characters */  return n8;    /* Number of 8-bit characters */
3659  }  }
3660    
3661    
3662    
3663    
3664  /*************************************************  /*************************************************
3665  *        Add a list of characters to a class     *  *        Add a list of characters to a class     *
3666  *************************************************/  *************************************************/
3667    
3668  /* This function is used for adding a list of case-equivalent characters to a  /* This function is used for adding a list of case-equivalent characters to a
3669  class, and also for adding a list of horizontal or vertical whitespace. If the  class, and also for adding a list of horizontal or vertical whitespace. If the
3670  list is in order (which it should be), ranges of characters are detected and  list is in order (which it should be), ranges of characters are detected and
3671  handled appropriately. This function is mutually recursive with the function  handled appropriately. This function is mutually recursive with the function
# Line 3546  Arguments: Line 3675  Arguments:
3675    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3676    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3677    options       the options word    options       the options word
3678    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3679    p             points to row of 32-bit values, terminated by NOTACHAR    p             points to row of 32-bit values, terminated by NOTACHAR
3680    except        character to omit; this is used when adding lists of    except        character to omit; this is used when adding lists of
3681                    case-equivalent characters to avoid including the one we                    case-equivalent characters to avoid including the one we
3682                    already know about                    already know about
3683    
3684  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3685                  the pointer to extra data is updated                  the pointer to extra data is updated
3686  */  */
# Line 3565  while (p[0] < NOTACHAR) Line 3694  while (p[0] < NOTACHAR)
3694    {    {
3695    int n = 0;    int n = 0;
3696    if (p[0] != except)    if (p[0] != except)
3697      {      {
3698      while(p[n+1] == p[0] + n + 1) n++;      while(p[n+1] == p[0] + n + 1) n++;
3699      n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);      n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);
3700      }      }
3701    p += n + 1;    p += n + 1;
3702    }    }
3703  return n8;  return n8;
3704  }  }
3705    
3706    
3707    
# Line 3587  Arguments: Line 3716  Arguments:
3716    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3717    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3718    options       the options word    options       the options word
3719    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3720    p             points to row of 32-bit values, terminated by NOTACHAR    p             points to row of 32-bit values, terminated by NOTACHAR
3721    
3722  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3723                  the pointer to extra data is updated                  the pointer to extra data is updated
3724  */  */
3725    
3726  static int  static int
3727  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,
3728    int options, compile_data *cd, const pcre_uint32 *p)    int options, compile_data *cd, const pcre_uint32 *p)
3729  {  {
3730    BOOL utf = (options & PCRE_UTF8) != 0;
3731  int n8 = 0;  int n8 = 0;
3732  if (p[0] > 0)  if (p[0] > 0)
3733    n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);    n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);
3734  while (p[0] < NOTACHAR)  while (p[0] < NOTACHAR)
3735    {    {
3736    while (p[1] == p[0] + 1) p++;    while (p[1] == p[0] + 1) p++;
3737    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,
3738      (p[1] == NOTACHAR)? 0x10ffff : p[1] - 1);      (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);
3739    p++;    p++;
3740    }    }
3741  return n8;  return n8;
3742  }  }
3743    
3744    
3745    
# Line 3624  to find out the amount of memory needed, Line 3754  to find out the amount of memory needed,
3754  phase. The value of lengthptr distinguishes the two phases.  phase. The value of lengthptr distinguishes the two phases.
3755    
3756  Arguments:  Arguments:
3757    optionsptr     pointer to the option bits    optionsptr        pointer to the option bits
3758    codeptr        points to the pointer to the current code point    codeptr           points to the pointer to the current code point
3759    ptrptr         points to the current pattern pointer    ptrptr            points to the current pattern pointer
3760    errorcodeptr   points to error code variable    errorcodeptr      points to error code variable
3761    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr      place to put the first required character
3762    reqcharptr     set to the last literal character required, else < 0    firstcharflagsptr place to put the first character flags, or a negative number
3763    bcptr          points to current branch chain    reqcharptr        place to put the last required character
3764    cond_depth     conditional nesting depth    reqcharflagsptr   place to put the last required character flags, or a negative number
3765    cd             contains pointers to tables etc.    bcptr             points to current branch chain
3766    lengthptr      NULL during the real compile phase    cond_depth        conditional nesting depth
3767                   points to length accumulator during pre-compile phase    cd                contains pointers to tables etc.
3768      lengthptr         NULL during the real compile phase
3769                        points to length accumulator during pre-compile phase
3770    
3771  Returns:         TRUE on success  Returns:            TRUE on success
3772                   FALSE, with *errorcodeptr set non-zero on error                      FALSE, with *errorcodeptr set non-zero on error
3773  */  */
3774    
3775  static BOOL  static BOOL
3776  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3777    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,    const pcre_uchar **ptrptr, int *errorcodeptr,
3778    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,    pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
3779      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
3780      branch_chain *bcptr, int cond_depth,
3781    compile_data *cd, int *lengthptr)    compile_data *cd, int *lengthptr)
3782  {  {
3783  int repeat_type, op_type;  int repeat_type, op_type;
3784  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3785  int bravalue = 0;  int bravalue = 0;
3786  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3787  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
3788  pcre_int32 zeroreqchar, zerofirstchar;  pcre_int32 firstcharflags, reqcharflags;
3789    pcre_uint32 zeroreqchar, zerofirstchar;
3790    pcre_int32 zeroreqcharflags, zerofirstcharflags;
3791  pcre_int32 req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3792  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3793  int after_manual_callout = 0;  int after_manual_callout = 0;
3794  int length_prevgroup = 0;  int length_prevgroup = 0;
3795  register int c;  register pcre_uint32 c;
3796    int escape;
3797  register pcre_uchar *code = *codeptr;  register pcre_uchar *code = *codeptr;
3798  pcre_uchar *last_code = code;  pcre_uchar *last_code = code;
3799  pcre_uchar *orig_code = code;  pcre_uchar *orig_code = code;
# Line 3676  must not do this for other options (e.g. Line 3813  must not do this for other options (e.g.
3813  dynamically as we process the pattern. */  dynamically as we process the pattern. */
3814    
3815  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3816  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3817  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3818    #ifndef COMPILE_PCRE32
3819  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3820    #endif
3821  #else  #else
3822  BOOL utf = FALSE;  BOOL utf = FALSE;
3823  #endif  #endif
3824    
3825  /* Helper variables for OP_XCLASS opcode (for characters > 255). */  /* Helper variables for OP_XCLASS opcode (for characters > 255). We define
3826    class_uchardata always so that it can be passed to add_to_class() always,
3827    though it will not be used in non-UTF 8-bit cases. This avoids having to supply
3828    alternative calls for the different cases. */
3829    
3830    pcre_uchar *class_uchardata;
3831  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3832  BOOL xclass;  BOOL xclass;
 pcre_uchar *class_uchardata;  
3833  pcre_uchar *class_uchardata_base;  pcre_uchar *class_uchardata_base;
3834  #endif  #endif
3835    
# Line 3710  to take the zero repeat into account. Th Line 3852  to take the zero repeat into account. Th
3852  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3853  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3854    
3855  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
3856    firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;
3857    
3858  /* The variable req_caseopt contains either the REQ_CASELESS value  /* The variable req_caseopt contains either the REQ_CASELESS value
3859  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 3736  for (;; ptr++) Line 3879  for (;; ptr++)
3879    int recno;    int recno;
3880    int refsign;    int refsign;
3881    int skipbytes;    int skipbytes;
3882    int subreqchar;    pcre_uint32 subreqchar, subfirstchar;
3883    int subfirstchar;    pcre_int32 subreqcharflags, subfirstcharflags;
3884    int terminator;    int terminator;
3885    int mclength;    unsigned int mclength;
3886    int tempbracount;    unsigned int tempbracount;
3887      pcre_uint32 ec;
3888    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3889    
3890    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 3750  for (;; ptr++) Line 3894  for (;; ptr++)
3894    /* If we are at the end of a nested substitution, revert to the outer level    /* If we are at the end of a nested substitution, revert to the outer level
3895    string. Nesting only happens one level deep. */    string. Nesting only happens one level deep. */
3896    
3897    if (c == 0 && nestptr != NULL)    if (c == CHAR_NULL && nestptr != NULL)
3898      {      {
3899      ptr = nestptr;      ptr = nestptr;
3900      nestptr = NULL;      nestptr = NULL;
# Line 3825  for (;; ptr++) Line 3969  for (;; ptr++)
3969    
3970    /* If in \Q...\E, check for the end; if not, we have a literal */    /* If in \Q...\E, check for the end; if not, we have a literal */
3971    
3972    if (inescq && c != 0)    if (inescq && c != CHAR_NULL)
3973      {      {
3974      if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)      if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)
3975        {        {
# Line 3873  for (;; ptr++) Line 4017  for (;; ptr++)
4017      if (c == CHAR_NUMBER_SIGN)      if (c == CHAR_NUMBER_SIGN)
4018        {        {
4019        ptr++;        ptr++;
4020        while (*ptr != 0)        while (*ptr != CHAR_NULL)
4021          {          {
4022          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
4023          ptr++;          ptr++;
# Line 3881  for (;; ptr++) Line 4025  for (;; ptr++)
4025          if (utf) FORWARDCHAR(ptr);          if (utf) FORWARDCHAR(ptr);
4026  #endif  #endif
4027          }          }
4028        if (*ptr != 0) continue;        if (*ptr != CHAR_NULL) continue;
4029    
4030        /* Else fall through to handle end of string */        /* Else fall through to handle end of string */
4031        c = 0;        c = 0;
# Line 3903  for (;; ptr++) Line 4047  for (;; ptr++)
4047      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
4048      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
4049      *firstcharptr = firstchar;      *firstcharptr = firstchar;
4050        *firstcharflagsptr = firstcharflags;
4051      *reqcharptr = reqchar;      *reqcharptr = reqchar;
4052        *reqcharflagsptr = reqcharflags;
4053      *codeptr = code;      *codeptr = code;
4054      *ptrptr = ptr;      *ptrptr = ptr;
4055      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3927  for (;; ptr++) Line 4073  for (;; ptr++)
4073      previous = NULL;      previous = NULL;
4074      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
4075        {        {
4076        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4077        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
4078        }        }
4079      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3942  for (;; ptr++) Line 4088  for (;; ptr++)
4088      repeats. The value of reqchar doesn't change either. */      repeats. The value of reqchar doesn't change either. */
4089    
4090      case CHAR_DOT:      case CHAR_DOT:
4091      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4092      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4093        zerofirstcharflags = firstcharflags;
4094      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4095        zeroreqcharflags = reqcharflags;
4096      previous = code;      previous = code;
4097      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
4098      break;      break;
# Line 4018  for (;; ptr++) Line 4166  for (;; ptr++)
4166          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
4167        {        {
4168        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
4169        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4170        zerofirstchar = firstchar;        zerofirstchar = firstchar;
4171          zerofirstcharflags = firstcharflags;
4172        break;        break;
4173        }        }
4174    
# Line 4054  for (;; ptr++) Line 4203  for (;; ptr++)
4203      means that an initial ] is taken as a data character. At the start of the      means that an initial ] is taken as a data character. At the start of the
4204      loop, c contains the first byte of the character. */      loop, c contains the first byte of the character. */
4205    
4206      if (c != 0) do      if (c != CHAR_NULL) do
4207        {        {
4208        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
4209    
# Line 4069  for (;; ptr++) Line 4218  for (;; ptr++)
4218        /* In the pre-compile phase, accumulate the length of any extra        /* In the pre-compile phase, accumulate the length of any extra
4219        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
4220        contain a zillion > 255 characters no longer overwrite the work space        contain a zillion > 255 characters no longer overwrite the work space
4221        (which is on the stack). We have to remember that there was XCLASS data,        (which is on the stack). We have to remember that there was XCLASS data,
4222        however. */        however. */
4223    
4224        if (lengthptr != NULL && class_uchardata > class_uchardata_base)        if (lengthptr != NULL && class_uchardata > class_uchardata_base)
# Line 4210  for (;; ptr++) Line 4359  for (;; ptr++)
4359    
4360        if (c == CHAR_BACKSLASH)        if (c == CHAR_BACKSLASH)
4361          {          {
4362          c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);          escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options,
4363              TRUE);
4364          if (*errorcodeptr != 0) goto FAILED;          if (*errorcodeptr != 0) goto FAILED;
4365            if (escape == 0) c = ec;
4366          if (-c == ESC_b) c = CHAR_BS;    /* \b is backspace in a class */          else if (escape == ESC_b) c = CHAR_BS; /* \b is backspace in a class */
4367          else if (-c == ESC_N)            /* \N is not supported in a class */          else if (escape == ESC_N)          /* \N is not supported in a class */
4368            {            {
4369            *errorcodeptr = ERR71;            *errorcodeptr = ERR71;
4370            goto FAILED;            goto FAILED;
4371            }            }
4372          else if (-c == ESC_Q)            /* Handle start of quoted string */          else if (escape == ESC_Q)            /* Handle start of quoted string */
4373            {            {
4374            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
4375              {              {
# Line 4228  for (;; ptr++) Line 4378  for (;; ptr++)
4378            else inescq = TRUE;            else inescq = TRUE;
4379            continue;            continue;
4380            }            }
4381          else if (-c == ESC_E) continue;  /* Ignore orphan \E */          else if (escape == ESC_E) continue;  /* Ignore orphan \E */
4382    
4383          if (c < 0)          else
4384            {            {
4385            register const pcre_uint8 *cbits = cd->cbits;            register const pcre_uint8 *cbits = cd->cbits;
4386            /* Every class contains at least two < 256 characters. */            /* Every class contains at least two < 256 characters. */
# Line 4238  for (;; ptr++) Line 4388  for (;; ptr++)
4388            /* Every class contains at least two characters. */            /* Every class contains at least two characters. */
4389            class_one_char += 2;            class_one_char += 2;
4390    
4391            switch (-c)            switch (escape)
4392              {              {
4393  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4394              case ESC_du:     /* These are the values given for \d etc */              case ESC_du:     /* These are the values given for \d etc */
# Line 4248  for (;; ptr++) Line 4398  for (;; ptr++)
4398              case ESC_su:     /* of the default ASCII testing. */              case ESC_su:     /* of the default ASCII testing. */
4399              case ESC_SU:              case ESC_SU:
4400              nestptr = ptr;              nestptr = ptr;
4401              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */              ptr = substitutes[escape - ESC_DU] - 1;  /* Just before substitute */
4402              class_has_8bitchar--;                /* Undo! */              class_has_8bitchar--;                /* Undo! */
4403              continue;              continue;
4404  #endif  #endif
# Line 4286  for (;; ptr++) Line 4436  for (;; ptr++)
4436              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];
4437              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */
4438              continue;              continue;
4439    
4440              /* The rest apply in both UCP and non-UCP cases. */              /* The rest apply in both UCP and non-UCP cases. */
4441    
4442              case ESC_h:              case ESC_h:
4443              (void)add_list_to_class(classbits, &class_uchardata, options, cd,              (void)add_list_to_class(classbits, &class_uchardata, options, cd,
4444                PRIV(hspace_list), NOTACHAR);                PRIV(hspace_list), NOTACHAR);
4445              continue;              continue;
4446    
4447              case ESC_H:              case ESC_H:
4448              (void)add_not_list_to_class(classbits, &class_uchardata, options,              (void)add_not_list_to_class(classbits, &class_uchardata, options,
4449                cd, PRIV(hspace_list));                cd, PRIV(hspace_list));
4450              continue;              continue;
4451    
4452              case ESC_v:              case ESC_v:
4453              (void)add_list_to_class(classbits, &class_uchardata, options, cd,              (void)add_list_to_class(classbits, &class_uchardata, options, cd,
4454                PRIV(vspace_list), NOTACHAR);                PRIV(vspace_list), NOTACHAR);
4455              continue;              continue;
4456    
4457              case ESC_V:              case ESC_V:
4458              (void)add_not_list_to_class(classbits, &class_uchardata, options,              (void)add_not_list_to_class(classbits, &class_uchardata, options,
4459                cd, PRIV(vspace_list));                cd, PRIV(vspace_list));
4460              continue;              continue;
4461    
4462  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 4314  for (;; ptr++) Line 4464  for (;; ptr++)
4464              case ESC_P:              case ESC_P:
4465                {                {
4466                BOOL negated;                BOOL negated;
4467                int pdata;                unsigned int ptype = 0, pdata = 0;
4468                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
4469                if (ptype < 0) goto FAILED;                  goto FAILED;
4470                *class_uchardata++ = ((-c == ESC_p) != negated)?                *class_uchardata++ = ((escape == ESC_p) != negated)?
4471                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4472                *class_uchardata++ = ptype;                *class_uchardata++ = ptype;
4473                *class_uchardata++ = pdata;                *class_uchardata++ = pdata;
# Line 4344  for (;; ptr++) Line 4494  for (;; ptr++)
4494    
4495          /* Fall through if the escape just defined a single character (c >= 0).          /* Fall through if the escape just defined a single character (c >= 0).
4496          This may be greater than 256. */          This may be greater than 256. */
4497    
4498            escape = 0;
4499    
4500          }   /* End of backslash handling */          }   /* End of backslash handling */
4501    
4502        /* A character may be followed by '-' to form a range. However, Perl does        /* A character may be followed by '-' to form a range. However, Perl does
# Line 4368  for (;; ptr++) Line 4520  for (;; ptr++)
4520    
4521        if (!inescq && ptr[1] == CHAR_MINUS)        if (!inescq && ptr[1] == CHAR_MINUS)
4522          {          {
4523          int d;          pcre_uint32 d;
4524          ptr += 2;          ptr += 2;
4525          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;
4526    
# Line 4383  for (;; ptr++) Line 4535  for (;; ptr++)
4535            inescq = TRUE;            inescq = TRUE;
4536            break;            break;
4537            }            }
4538    
4539          /* Minus (hyphen) at the end of a class is treated as a literal, so put          /* Minus (hyphen) at the end of a class is treated as a literal, so put
4540          back the pointer and jump to handle the character that preceded it. */          back the pointer and jump to handle the character that preceded it. */
4541    
4542          if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))          if (*ptr == CHAR_NULL || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))
4543            {            {
4544            ptr = oldptr;            ptr = oldptr;
4545            goto CLASS_SINGLE_CHARACTER;            goto CLASS_SINGLE_CHARACTER;
4546            }            }
4547    
4548          /* Otherwise, we have a potential range; pick up the next character */          /* Otherwise, we have a potential range; pick up the next character */
4549    
4550  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
# Line 4410  for (;; ptr++) Line 4562  for (;; ptr++)
4562    
4563          if (!inescq && d == CHAR_BACKSLASH)          if (!inescq && d == CHAR_BACKSLASH)
4564            {            {
4565            d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);            int descape;
4566              descape = check_escape(&ptr, &d, errorcodeptr, cd->bracount, options, TRUE);
4567            if (*errorcodeptr != 0) goto FAILED;            if (*errorcodeptr != 0) goto FAILED;
4568    
4569            /* \b is backspace; any other special means the '-' was literal. */            /* \b is backspace; any other special means the '-' was literal. */
4570    
4571            if (d < 0)            if (descape != 0)
4572              {              {
4573              if (d == -ESC_b) d = CHAR_BS; else              if (descape == ESC_b) d = CHAR_BS; else
4574                {                {
4575                ptr = oldptr;                ptr = oldptr;
4576                goto CLASS_SINGLE_CHARACTER;  /* A few lines below */                goto CLASS_SINGLE_CHARACTER;  /* A few lines below */
# Line 4438  for (;; ptr++) Line 4591  for (;; ptr++)
4591          /* We have found a character range, so single character optimizations          /* We have found a character range, so single character optimizations
4592          cannot be done anymore. Any value greater than 1 indicates that there          cannot be done anymore. Any value greater than 1 indicates that there
4593          is more than one character. */          is more than one character. */
4594    
4595          class_one_char = 2;          class_one_char = 2;
4596    
4597          /* Remember an explicit \r or \n, and add the range to the class. */          /* Remember an explicit \r or \n, and add the range to the class. */
4598    
4599          if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;          if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;
4600    
4601          class_has_8bitchar +=          class_has_8bitchar +=
4602            add_to_class(classbits, &class_uchardata, options, cd, c, d);            add_to_class(classbits, &class_uchardata, options, cd, c, d);
4603    
4604          continue;   /* Go get the next char in the class */          continue;   /* Go get the next char in the class */
4605          }          }
4606    
4607        /* Handle a single character - we can get here for a normal non-escape        /* Handle a single character - we can get here for a normal non-escape
4608        char, or after \ that introduces a single character or for an apparent        char, or after \ that introduces a single character or for an apparent
4609        range that isn't. Only the value 1 matters for class_one_char, so don't        range that isn't. Only the value 1 matters for class_one_char, so don't
4610        increase it if it is already 2 or more ... just in case there's a class        increase it if it is already 2 or more ... just in case there's a class
4611        with a zillion characters in it. */        with a zillion characters in it. */
4612    
4613        CLASS_SINGLE_CHARACTER:        CLASS_SINGLE_CHARACTER:
# Line 4473  for (;; ptr++) Line 4626  for (;; ptr++)
4626          {          {
4627          ptr++;          ptr++;
4628          zeroreqchar = reqchar;          zeroreqchar = reqchar;
4629            zeroreqcharflags = reqcharflags;
4630    
4631          if (negate_class)          if (negate_class)
4632            {            {
4633            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;  #ifdef SUPPORT_UCP
4634              int d;
4635    #endif
4636              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4637            zerofirstchar = firstchar;            zerofirstchar = firstchar;
4638            *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;            zerofirstcharflags = firstcharflags;
4639  #ifdef SUPPORT_UTF  
4640            if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)            /* For caseless UTF-8 mode when UCP support is available, check
4641              code += PRIV(ord2utf)(c, code);            whether this character has more than one other case. If so, generate
4642              a special OP_NOTPROP item instead of OP_NOTI. */
4643    
4644    #ifdef SUPPORT_UCP
4645              if (utf && (options & PCRE_CASELESS) != 0 &&
4646                  (d = UCD_CASESET(c)) != 0)
4647                {
4648                *code++ = OP_NOTPROP;
4649                *code++ = PT_CLIST;
4650                *code++ = d;
4651                }
4652            else            else
4653  #endif  #endif
4654              *code++ = c;            /* Char has only one other case, or UCP not available */
4655            goto NOT_CHAR;  
4656                {
4657                *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
4658    #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4659                if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
4660                  code += PRIV(ord2utf)(c, code);
4661                else
4662    #endif
4663                  *code++ = c;
4664                }
4665    
4666              /* We are finished with this character class */
4667    
4668              goto END_CLASS;
4669            }            }
4670    
4671          /* For a single, positive character, get the value into mcbuffer, and          /* For a single, positive character, get the value into mcbuffer, and
4672          then we can handle this with the normal one-character code. */          then we can handle this with the normal one-character code. */
4673    
4674  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4675          if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)          if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
4676            mclength = PRIV(ord2utf)(c, mcbuffer);            mclength = PRIV(ord2utf)(c, mcbuffer);
4677          else          else
# Line 4502  for (;; ptr++) Line 4682  for (;; ptr++)
4682            }            }
4683          goto ONE_CHAR;          goto ONE_CHAR;
4684          }       /* End of 1-char optimization */          }       /* End of 1-char optimization */
4685    
4686        /* There is more than one character in the class, or an XCLASS item        /* There is more than one character in the class, or an XCLASS item
4687        has been generated. Add this character to the class. */        has been generated. Add this character to the class. */
4688    
4689        class_has_8bitchar +=        class_has_8bitchar +=
4690          add_to_class(classbits, &class_uchardata, options, cd, c, c);          add_to_class(classbits, &class_uchardata, options, cd, c, c);
4691        }        }
4692    
# Line 4514  for (;; ptr++) Line 4694  for (;; ptr++)
4694      If we are at the end of an internal nested string, revert to the outer      If we are at the end of an internal nested string, revert to the outer
4695      string. */      string. */
4696    
4697      while (((c = *(++ptr)) != 0 ||      while (((c = *(++ptr)) != CHAR_NULL ||
4698             (nestptr != NULL &&             (nestptr != NULL &&
4699               (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != 0)) &&               (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != CHAR_NULL)) &&
4700             (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));             (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));
4701    
4702      /* Check for missing terminating ']' */      /* Check for missing terminating ']' */
4703    
4704      if (c == 0)      if (c == CHAR_NULL)
4705        {        {
4706        *errorcodeptr = ERR6;        *errorcodeptr = ERR6;
4707        goto FAILED;        goto FAILED;
4708        }        }
4709    
4710      /* We will need an XCLASS if data has been placed in class_uchardata. In      /* We will need an XCLASS if data has been placed in class_uchardata. In
4711      the second phase this is a sufficient test. However, in the pre-compile      the second phase this is a sufficient test. However, in the pre-compile
4712      phase, class_uchardata gets emptied to prevent workspace overflow, so it      phase, class_uchardata gets emptied to prevent workspace overflow, so it
4713      only if the very last character in the class needs XCLASS will it contain      only if the very last character in the class needs XCLASS will it contain
4714      anything at this point. For this reason, xclass gets set TRUE above when      anything at this point. For this reason, xclass gets set TRUE above when
4715      uchar_classdata is emptied, and that's why this code is the way it is here      uchar_classdata is emptied, and that's why this code is the way it is here
4716      instead of just doing a test on class_uchardata below. */      instead of just doing a test on class_uchardata below. */
4717    
4718  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
4719      if (class_uchardata > class_uchardata_base) xclass = TRUE;      if (class_uchardata > class_uchardata_base) xclass = TRUE;
4720  #endif  #endif
# Line 4543  for (;; ptr++) Line 4723  for (;; ptr++)
4723      setting, whatever the repeat count. Any reqchar setting must remain      setting, whatever the repeat count. Any reqchar setting must remain
4724      unchanged after any kind of repeat. */      unchanged after any kind of repeat. */
4725    
4726      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4727      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4728        zerofirstcharflags = firstcharflags;
4729      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4730        zeroreqcharflags = reqcharflags;
4731    
4732      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4733      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 4601  for (;; ptr++) Line 4783  for (;; ptr++)
4783        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
4784        }        }
4785      code += 32 / sizeof(pcre_uchar);      code += 32 / sizeof(pcre_uchar);
4786      NOT_CHAR:  
4787        END_CLASS:
4788      break;      break;
4789    
4790    
# Line 4639  for (;; ptr++) Line 4822  for (;; ptr++)
4822      if (repeat_min == 0)      if (repeat_min == 0)
4823        {        {
4824        firstchar = zerofirstchar;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4825          firstcharflags = zerofirstcharflags;
4826        reqchar = zeroreqchar;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4827          reqcharflags = zeroreqcharflags;
4828        }        }
4829    
4830      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4725  for (;; ptr++) Line 4910  for (;; ptr++)
4910        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
4911        it's a length rather than a small character. */        it's a length rather than a small character. */
4912    
4913  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4914        if (utf && NOT_FIRSTCHAR(code[-1]))        if (utf && NOT_FIRSTCHAR(code[-1]))
4915          {          {
4916          pcre_uchar *lastchar = code - 1;          pcre_uchar *lastchar = code - 1;
# Line 4742  for (;; ptr++) Line 4927  for (;; ptr++)
4927          {          {
4928          c = code[-1];          c = code[-1];
4929          if (*previous <= OP_CHARI && repeat_min > 1)          if (*previous <= OP_CHARI && repeat_min > 1)
4930            reqchar = c | req_caseopt | cd->req_varyopt;            {
4931              reqchar = c;
4932              reqcharflags = req_caseopt | cd->req_varyopt;
4933              }
4934          }          }
4935    
4936        /* 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 4799  for (;; ptr++) Line 4987  for (;; ptr++)
4987    
4988        if (repeat_max == 0) goto END_REPEAT;        if (repeat_max == 0) goto END_REPEAT;
4989    
       /*--------------------------------------------------------------------*/  
       /* This code is obsolete from release 8.00; the restriction was finally  
       removed: */  
   
       /* All real repeats make it impossible to handle partial matching (maybe  
       one day we will be able to remove this restriction). */  
   
       /* if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; */  
       /*--------------------------------------------------------------------*/  
   
4990        /* Combine the op_type with the repeat_type */        /* Combine the op_type with the repeat_type */
4991    
4992        repeat_type += op_type;        repeat_type += op_type;
# Line 4861  for (;; ptr++) Line 5039  for (;; ptr++)
5039    
5040          if (repeat_max < 0)          if (repeat_max < 0)
5041            {            {
5042  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
5043            if (utf && (c & UTF_LENGTH) != 0)            if (utf && (c & UTF_LENGTH) != 0)
5044              {              {
5045              memcpy(code, utf_chars, IN_UCHARS(c & 7));              memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 4886  for (;; ptr++) Line 5064  for (;; ptr++)
5064    
5065          else if (repeat_max != repeat_min)          else if (repeat_max != repeat_min)
5066            {            {
5067  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
5068            if (utf && (c & UTF_LENGTH) != 0)            if (utf && (c & UTF_LENGTH) != 0)
5069              {              {
5070              memcpy(code, utf_chars, IN_UCHARS(c & 7));              memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 4916  for (;; ptr++) Line 5094  for (;; ptr++)
5094    
5095        /* The character or character type itself comes last in all cases. */        /* The character or character type itself comes last in all cases. */
5096    
5097  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
5098        if (utf && (c & UTF_LENGTH) != 0)        if (utf && (c & UTF_LENGTH) != 0)
5099          {          {
5100          memcpy(code, utf_chars, IN_UCHARS(c & 7));          memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 4955  for (;; ptr++) Line 5133  for (;; ptr++)
5133          goto END_REPEAT;          goto END_REPEAT;
5134          }          }
5135    
       /*--------------------------------------------------------------------*/  
       /* This code is obsolete from release 8.00; the restriction was finally  
       removed: */  
   
       /* All real repeats make it impossible to handle partial matching (maybe  
       one day we will be able to remove this restriction). */  
   
       /* if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; */  
       /*--------------------------------------------------------------------*/  
   
5136        if (repeat_min == 0 && repeat_max == -1)        if (repeat_min == 0 && repeat_max == -1)
5137          *code++ = OP_CRSTAR + repeat_type;          *code++ = OP_CRSTAR + repeat_type;
5138        else if (repeat_min == 1 && repeat_max == -1)        else if (repeat_min == 1 && repeat_max == -1)
# Line 5124  for (;; ptr++) Line 5292  for (;; ptr++)
5292    
5293            else            else
5294              {              {
5295              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;              if (groupsetfirstchar && reqcharflags < 0)
5296                  {
5297                  reqchar = firstchar;
5298                  reqcharflags = firstcharflags;
5299                  }
5300    
5301              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5302                {                {
# Line 5303  for (;; ptr++) Line 5475  for (;; ptr++)
5475              pcre_uchar *scode = bracode;              pcre_uchar *scode = bracode;
5476              do              do
5477                {                {
5478                if (could_be_empty_branch(scode, ketcode, utf, cd))                if (could_be_empty_branch(scode, ketcode, utf, cd, NULL))
5479                  {                  {
5480                  *bracode += OP_SBRA - OP_BRA;                  *bracode += OP_SBRA - OP_BRA;
5481                  break;                  break;
# Line 5499  for (;; ptr++) Line 5671  for (;; ptr++)
5671        if (*ptr == CHAR_COLON)        if (*ptr == CHAR_COLON)
5672          {          {
5673          arg = ++ptr;          arg = ++ptr;
5674          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5675          arglen = (int)(ptr - arg);          arglen = (int)(ptr - arg);
5676          if (arglen > (int)MAX_MARK)          if ((unsigned int)arglen > MAX_MARK)
5677            {            {
5678            *errorcodeptr = ERR75;            *errorcodeptr = ERR75;
5679            goto FAILED;            goto FAILED;
# Line 5544  for (;; ptr++) Line 5716  for (;; ptr++)
5716                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5717    
5718              /* Do not set firstchar after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5719              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
5720              }              }
5721    
5722            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5613  for (;; ptr++) Line 5785  for (;; ptr++)
5785          {          {
5786          case CHAR_NUMBER_SIGN:                 /* Comment; skip to ket */          case CHAR_NUMBER_SIGN:                 /* Comment; skip to ket */
5787          ptr++;          ptr++;
5788          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5789          if (*ptr == 0)          if (*ptr == CHAR_NULL)
5790            {            {
5791            *errorcodeptr = ERR18;            *errorcodeptr = ERR18;
5792            goto FAILED;            goto FAILED;
# Line 5637  for (;; ptr++) Line 5809  for (;; ptr++)
5809          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
5810          case CHAR_LEFT_PARENTHESIS:          case CHAR_LEFT_PARENTHESIS:
5811          bravalue = OP_COND;       /* Conditional group */          bravalue = OP_COND;       /* Conditional group */
5812            tempptr = ptr;
5813    
5814          /* A condition can be an assertion, a number (referring to a numbered          /* A condition can be an assertion, a number (referring to a numbered
5815          group), a name (referring to a named group), or 'R', referring to          group), a name (referring to a named group), or 'R', referring to
# Line 5649  for (;; ptr++) Line 5822  for (;; ptr++)
5822          be the recursive thing or the name 'R' (and similarly for 'R' followed          be the recursive thing or the name 'R' (and similarly for 'R' followed
5823          by digits), and (b) a number could be a name that consists of digits.          by digits), and (b) a number could be a name that consists of digits.
5824          In both cases, we look for a name first; if not found, we try the other          In both cases, we look for a name first; if not found, we try the other
5825          cases. */          cases.
5826    
5827            For compatibility with auto-callouts, we allow a callout to be
5828            specified before a condition that is an assertion. First, check for the
5829            syntax of a callout; if found, adjust the temporary pointer that is
5830            used to check for an assertion condition. That's all that is needed! */
5831    
5832            if (ptr[1] == CHAR_QUESTION_MARK && ptr[2] == CHAR_C)
5833              {
5834              for (i = 3;; i++) if (!IS_DIGIT(ptr[i])) break;
5835              if (ptr[i] == CHAR_RIGHT_PARENTHESIS)
5836                tempptr += i + 1;
5837              }
5838    
5839          /* For conditions that are assertions, check the syntax, and then exit          /* For conditions that are assertions, check the syntax, and then exit
5840          the switch. This will take control down to where bracketed groups,          the switch. This will take control down to where bracketed groups,
5841          including assertions, are processed. */          including assertions, are processed. */
5842    
5843          if (ptr[1] == CHAR_QUESTION_MARK && (ptr[2] == CHAR_EQUALS_SIGN ||          if (tempptr[1] == CHAR_QUESTION_MARK &&
5844              ptr[2] == CHAR_EXCLAMATION_MARK || ptr[2] == CHAR_LESS_THAN_SIGN))                (tempptr[2] == CHAR_EQUALS_SIGN ||
5845                   tempptr[2] == CHAR_EXCLAMATION_MARK ||
5846                   tempptr[2] == CHAR_LESS_THAN_SIGN))
5847            break;            break;
5848    
5849          /* Most other conditions use OP_CREF (a couple change to OP_RREF          /* Most other conditions use OP_CREF (a couple change to OP_RREF
# Line 5690  for (;; ptr++) Line 5877  for (;; ptr++)
5877            }            }
5878          else          else
5879            {            {
5880            terminator = 0;            terminator = CHAR_NULL;
5881            if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);            if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);
5882            }            }
5883    
# Line 5710  for (;; ptr++) Line 5897  for (;; ptr++)
5897          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)
5898            {            {
5899            if (recno >= 0)            if (recno >= 0)
5900              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;              recno = (IS_DIGIT(*ptr))? recno * 10 + (int)(*ptr - CHAR_0) : -1;
5901            ptr++;            ptr++;
5902            }            }
5903          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5904    
5905          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) ||
5906              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
5907            {            {
5908            ptr--;      /* Error offset */            ptr--;      /* Error offset */
# Line 5780  for (;; ptr++) Line 5967  for (;; ptr++)
5967            code[1+LINK_SIZE]++;            code[1+LINK_SIZE]++;
5968            }            }
5969    
5970          /* If terminator == 0 it means that the name followed directly after          /* If terminator == CHAR_NULL it means that the name followed directly
5971          the opening parenthesis [e.g. (?(abc)...] and in this case there are          after the opening parenthesis [e.g. (?(abc)...] and in this case there
5972          some further alternatives to try. For the cases where terminator != 0          are some further alternatives to try. For the cases where terminator !=
5973          [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have          0 [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have
5974          now checked all the possibilities, so give an error. */          now checked all the possibilities, so give an error. */
5975    
5976          else if (terminator != 0)          else if (terminator != CHAR_NULL)
5977            {            {
5978            *errorcodeptr = ERR15;            *errorcodeptr = ERR15;
5979            goto FAILED;            goto FAILED;
# Line 5955  for (;; ptr++) Line 6142  for (;; ptr++)
6142    
6143            if (lengthptr != NULL)            if (lengthptr != NULL)
6144              {              {
6145              if (*ptr != terminator)              if (*ptr != (pcre_uchar)terminator)
6146                {                {
6147                *errorcodeptr = ERR42;                *errorcodeptr = ERR42;
6148                goto FAILED;                goto FAILED;
# Line 6097  for (;; ptr++) Line 6284  for (;; ptr++)
6284              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
6285              goto FAILED;              goto FAILED;
6286              }              }
6287            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6288              {              {
6289              *errorcodeptr = ERR42;              *errorcodeptr = ERR42;
6290              goto FAILED;              goto FAILED;
# Line 6203  for (;; ptr++) Line 6390  for (;; ptr++)
6390            while(IS_DIGIT(*ptr))            while(IS_DIGIT(*ptr))
6391              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6392    
6393            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6394              {              {
6395              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
6396              goto FAILED;              goto FAILED;
# Line 6307  for (;; ptr++) Line 6494  for (;; ptr++)
6494    
6495          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6496    
6497          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6498          continue;          continue;
6499    
6500    
# Line 6441  for (;; ptr++) Line 6628  for (;; ptr++)
6628           cond_depth +           cond_depth +
6629             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6630           &subfirstchar,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6631             &subfirstcharflags,
6632           &subreqchar,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6633             &subreqcharflags,
6634           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6635           cd,                              /* Tables block */           cd,                              /* Tables block */
6636           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6502  for (;; ptr++) Line 6691  for (;; ptr++)
6691            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6692            goto FAILED;            goto FAILED;
6693            }            }
6694          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;          if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
6695          }          }
6696        }        }
6697    
# Line 6551  for (;; ptr++) Line 6740  for (;; ptr++)
6740      back off. */      back off. */
6741    
6742      zeroreqchar = reqchar;      zeroreqchar = reqchar;
6743        zeroreqcharflags = reqcharflags;
6744      zerofirstchar = firstchar;      zerofirstchar = firstchar;
6745        zerofirstcharflags = firstcharflags;
6746      groupsetfirstchar = FALSE;      groupsetfirstchar = FALSE;
6747    
6748      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
# Line 6562  for (;; ptr++) Line 6753  for (;; ptr++)
6753        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
6754        repeat forces firstchar to "none". */        repeat forces firstchar to "none". */
6755    
6756        if (firstchar == REQ_UNSET)        if (firstcharflags == REQ_UNSET)
6757          {          {
6758          if (subfirstchar >= 0)          if (subfirstcharflags >= 0)
6759            {            {
6760            firstchar = subfirstchar;            firstchar = subfirstchar;
6761              firstcharflags = subfirstcharflags;
6762            groupsetfirstchar = TRUE;            groupsetfirstchar = TRUE;
6763            }            }
6764          else firstchar = REQ_NONE;          else firstcharflags = REQ_NONE;
6765          zerofirstchar = REQ_NONE;          zerofirstcharflags = REQ_NONE;
6766          }          }
6767    
6768        /* If firstchar was previously set, convert the subpattern's firstchar        /* If firstchar was previously set, convert the subpattern's firstchar
6769        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
6770        existence beforehand. */        existence beforehand. */
6771    
6772        else if (subfirstchar >= 0 && subreqchar < 0)        else if (subfirstcharflags >= 0 && subreqcharflags < 0)
6773          subreqchar = subfirstchar | tempreqvary;          {
6774            subreqchar = subfirstchar;
6775            subreqcharflags = subfirstcharflags | tempreqvary;
6776            }
6777    
6778        /* 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
6779        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6780    
6781        if (subreqchar >= 0) reqchar = subreqchar;        if (subreqcharflags >= 0)
6782            {
6783            reqchar = subreqchar;
6784            reqcharflags = subreqcharflags;
6785            }
6786        }        }
6787    
6788      /* 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 6594  for (;; ptr++) Line 6793  for (;; ptr++)
6793      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
6794      firstchar, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6795    
6796      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;      else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
6797          {
6798          reqchar = subreqchar;
6799          reqcharflags = subreqcharflags;
6800          }
6801      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6802    
6803    
# Line 6602  for (;; ptr++) Line 6805  for (;; ptr++)
6805      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
6806      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
6807      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
6808      are ESC_REF plus the reference number. Only back references and those types      are negative the reference number. Only back references and those types
6809      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
6810      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
6811      ever created. */      ever created. */
6812    
6813      case CHAR_BACKSLASH:      case CHAR_BACKSLASH:
6814      tempptr = ptr;      tempptr = ptr;
6815      c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE);      escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, FALSE);
6816      if (*errorcodeptr != 0) goto FAILED;      if (*errorcodeptr != 0) goto FAILED;
6817    
6818      if (c < 0)      if (escape == 0)                  /* The escape coded a single character */
6819          c = ec;
6820        else
6821        {        {
6822        if (-c == ESC_Q)            /* Handle start of quoted string */        if (escape == ESC_Q)            /* Handle start of quoted string */
6823          {          {
6824          if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)          if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
6825            ptr += 2;               /* avoid empty string */            ptr += 2;               /* avoid empty string */
# Line 6622  for (;; ptr++) Line 6827  for (;; ptr++)
6827          continue;          continue;
6828          }          }
6829    
6830        if (-c == ESC_E) continue;  /* Perl ignores an orphan \E */        if (escape == ESC_E) continue;  /* Perl ignores an orphan \E */
6831    
6832        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6833        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6834    
6835        if (firstchar == REQ_UNSET && -c > ESC_b && -c < ESC_Z)        if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6836          firstchar = REQ_NONE;          firstcharflags = REQ_NONE;
6837    
6838        /* 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. */
6839    
6840        zerofirstchar = firstchar;        zerofirstchar = firstchar;
6841          zerofirstcharflags = firstcharflags;
6842        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6843          zeroreqcharflags = reqcharflags;
6844    
6845        /* \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'
6846        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
6847        -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 <
6848        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
6849        -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
6850        that is a synonym for a named back reference). */        that is a synonym for a named back reference). */
6851    
6852        if (-c == ESC_g)        if (escape == ESC_g)
6853          {          {
6854          const pcre_uchar *p;          const pcre_uchar *p;
6855          save_hwm = cd->hwm;   /* Normally this is set when '(' is read */          save_hwm = cd->hwm;   /* Normally this is set when '(' is read */
# Line 6662  for (;; ptr++) Line 6869  for (;; ptr++)
6869          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
6870            {            {
6871            BOOL is_a_number = TRUE;            BOOL is_a_number = TRUE;
6872            for (p = ptr + 1; *p != 0 && *p != terminator; p++)            for (p = ptr + 1; *p != CHAR_NULL && *p != (pcre_uchar)terminator; p++)
6873              {              {
6874              if (!MAX_255(*p)) { is_a_number = FALSE; break; }              if (!MAX_255(*p)) { is_a_number = FALSE; break; }
6875              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
6876              if ((cd->ctypes[*p] & ctype_word) == 0) break;              if ((cd->ctypes[*p] & ctype_word) == 0) break;
6877              }              }
6878            if (*p != terminator)            if (*p != (pcre_uchar)terminator)
6879              {              {
6880              *errorcodeptr = ERR57;              *errorcodeptr = ERR57;
6881              break;              break;
# Line 6686  for (;; ptr++) Line 6893  for (;; ptr++)
6893    
6894          p = ptr + 2;          p = ptr + 2;
6895          while (IS_DIGIT(*p)) p++;          while (IS_DIGIT(*p)) p++;
6896          if (*p != terminator)          if (*p != (pcre_uchar)terminator)
6897            {            {
6898            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
6899            break;            break;
# Line 6698  for (;; ptr++) Line 6905  for (;; ptr++)
6905        /* \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).
6906        We also support \k{name} (.NET syntax).  */        We also support \k{name} (.NET syntax).  */
6907    
6908        if (-c == ESC_k)        if (escape == ESC_k)
6909          {          {
6910          if ((ptr[1] != CHAR_LESS_THAN_SIGN &&          if ((ptr[1] != CHAR_LESS_THAN_SIGN &&
6911            ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))            ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))
# Line 6717  for (;; ptr++) Line 6924  for (;; ptr++)
6924        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
6925        ':' later. */        ':' later. */
6926    
6927        if (-c >= ESC_REF)        if (escape < 0)
6928          {          {
6929          open_capitem *oc;          open_capitem *oc;
6930          recno = -c - ESC_REF;          recno = -escape;
6931    
6932          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6933          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6934          previous = code;          previous = code;
6935          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6936          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6747  for (;; ptr++) Line 6954  for (;; ptr++)
6954        /* So are Unicode property matches, if supported. */        /* So are Unicode property matches, if supported. */
6955    
6956  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6957        else if (-c == ESC_P || -c == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6958          {          {
6959          BOOL negated;          BOOL negated;
6960          int pdata;          unsigned int ptype = 0, pdata = 0;
6961          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
6962          if (ptype < 0) goto FAILED;            goto FAILED;
6963          previous = code;          previous = code;
6964          *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6965          *code++ = ptype;          *code++ = ptype;
6966          *code++ = pdata;          *code++ = pdata;
6967          }          }
# Line 6763  for (;; ptr++) Line 6970  for (;; ptr++)
6970        /* If Unicode properties are not supported, \X, \P, and \p are not        /* If Unicode properties are not supported, \X, \P, and \p are not
6971        allowed. */        allowed. */
6972    
6973        else if (-c == ESC_X || -c == ESC_P || -c == ESC_p)        else if (escape == ESC_X || escape == ESC_P || escape == ESC_p)
6974          {          {
6975          *errorcodeptr = ERR45;          *errorcodeptr = ERR45;
6976          goto FAILED;          goto FAILED;
# Line 6774  for (;; ptr++) Line 6981  for (;; ptr++)
6981        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
6982        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
6983        Unicode property tests. Note that \b and \B do a one-character        Unicode property tests. Note that \b and \B do a one-character
6984        lookbehind. */        lookbehind, and \A also behaves as if it does. */
6985    
6986        else        else
6987          {          {
6988          if ((-c == ESC_b || -c == ESC_B) && cd->max_lookbehind == 0)          if ((escape == ESC_b || escape == ESC_B || escape == ESC_A) &&
6989                 cd->max_lookbehind == 0)
6990            cd->max_lookbehind = 1;            cd->max_lookbehind = 1;
6991  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6992          if (-c >= ESC_DU && -c <= ESC_wu)          if (escape >= ESC_DU && escape <= ESC_wu)
6993            {            {
6994            nestptr = ptr + 1;                   /* Where to resume */            nestptr = ptr + 1;                   /* Where to resume */
6995            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */            ptr = substitutes[escape - ESC_DU] - 1;  /* Just before substitute */
6996            }            }
6997          else          else
6998  #endif  #endif
# Line 6792  for (;; ptr++) Line 7000  for (;; ptr++)
7000          so that it works in DFA mode and in lookbehinds. */          so that it works in DFA mode and in lookbehinds. */
7001    
7002            {            {
7003            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (escape > ESC_b && escape < ESC_Z)? code : NULL;
7004            *code++ = (!utf && c == -ESC_C)? OP_ALLANY : -c;            *code++ = (!utf && escape == ESC_C)? OP_ALLANY : escape;
7005            }            }
7006          }          }
7007        continue;        continue;
# Line 6803  for (;; ptr++) Line 7011  for (;; ptr++)
7011      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
7012      handle it as a data character. */      handle it as a data character. */
7013    
7014  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
7015      if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)      if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
7016        mclength = PRIV(ord2utf)(c, mcbuffer);        mclength = PRIV(ord2utf)(c, mcbuffer);
7017      else      else
# Line 6836  for (;; ptr++) Line 7044  for (;; ptr++)
7044    
7045      ONE_CHAR:      ONE_CHAR:
7046      previous = code;      previous = code;
7047    
7048        /* For caseless UTF-8 mode when UCP support is available, check whether
7049        this character has more than one other case. If so, generate a special
7050        OP_PROP item instead of OP_CHARI. */
7051    
7052    #ifdef SUPPORT_UCP
7053        if (utf && (options & PCRE_CASELESS) != 0)
7054          {
7055          GETCHAR(c, mcbuffer);
7056          if ((c = UCD_CASESET(c)) != 0)
7057            {
7058            *code++ = OP_PROP;
7059            *code++ = PT_CLIST;
7060            *code++ = c;
7061            if (firstcharflags == REQ_UNSET)
7062              firstcharflags = zerofirstcharflags = REQ_NONE;
7063            break;
7064            }
7065          }
7066    #endif
7067    
7068        /* Caseful matches, or not one of the multicase characters. */
7069    
7070      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;
7071      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
7072    
# Line 6849  for (;; ptr++) Line 7080  for (;; ptr++)
7080      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
7081      repeat. */      repeat. */
7082    
7083      if (firstchar == REQ_UNSET)      if (firstcharflags == REQ_UNSET)
7084        {        {
7085        zerofirstchar = REQ_NONE;        zerofirstcharflags = REQ_NONE;
7086        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7087          zeroreqcharflags = reqcharflags;
7088    
7089        /* 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
7090        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
# Line 6860  for (;; ptr++) Line 7092  for (;; ptr++)
7092        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7093          {          {
7094          firstchar = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
7095          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;          firstchar = mcbuffer[0];
7096            firstcharflags = req_caseopt;
7097    
7098            if (mclength != 1)
7099              {
7100              reqchar = code[-1];
7101              reqcharflags = cd->req_varyopt;
7102              }
7103          }          }
7104        else firstchar = reqchar = REQ_NONE;        else firstcharflags = reqcharflags = REQ_NONE;
7105        }        }
7106    
7107      /* 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 6871  for (;; ptr++) Line 7110  for (;; ptr++)
7110      else      else
7111        {        {
7112        zerofirstchar = firstchar;        zerofirstchar = firstchar;
7113          zerofirstcharflags = firstcharflags;
7114        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7115          zeroreqcharflags = reqcharflags;
7116        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7117          reqchar = code[-1] | req_caseopt | cd->req_varyopt;          {
7118            reqchar = code[-1];
7119            reqcharflags = req_caseopt | cd->req_varyopt;
7120            }
7121        }        }
7122    
7123      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6892  return FALSE; Line 7136  return FALSE;
7136    
7137    
7138    
   
7139  /*************************************************  /*************************************************
7140  *     Compile sequence of alternatives           *  *     Compile sequence of alternatives           *
7141  *************************************************/  *************************************************/
# Line 6905  out the amount of memory needed, as well Line 7148  out the amount of memory needed, as well
7148  value of lengthptr distinguishes the two phases.  value of lengthptr distinguishes the two phases.
7149    
7150  Arguments:  Arguments:
7151    options        option bits, including any changes for this subpattern    options           option bits, including any changes for this subpattern
7152    codeptr        -> the address of the current code pointer    codeptr           -> the address of the current code pointer
7153    ptrptr         -> the address of the current pattern pointer    ptrptr            -> the address of the current pattern pointer
7154    errorcodeptr   -> pointer to error code variable    errorcodeptr      -> pointer to error code variable
7155    lookbehind     TRUE if this is a lookbehind assertion    lookbehind        TRUE if this is a lookbehind assertion
7156    reset_bracount TRUE to reset the count for each branch    reset_bracount    TRUE to reset the count for each branch
7157    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes         skip this many bytes at start (for brackets and OP_COND)
7158    cond_depth     depth of nesting for conditional subpatterns    cond_depth        depth of nesting for conditional subpatterns
7159    firstcharptr   place to put the first required character, or a negative number    firstcharptr      place to put the first required character
7160    reqcharptr     place to put the last required character, or a negative number    firstcharflagsptr place to put the first character flags, or a negative number
7161    bcptr          pointer to the chain of currently open branches    reqcharptr        place to put the last required character
7162    cd             points to the data block with tables pointers etc.    reqcharflagsptr   place to put the last required character flags, or a negative number
7163    lengthptr      NULL during the real compile phase    bcptr             pointer to the chain of currently open branches
7164                   points to length accumulator during pre-compile phase    cd                points to the data block with tables pointers etc.
7165      lengthptr         NULL during the real compile phase
7166                        points to length accumulator during pre-compile phase
7167    
7168  Returns:         TRUE on success  Returns:            TRUE on success
7169  */  */
7170    
7171  static BOOL  static BOOL
7172  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
7173    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
7174    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,    int cond_depth,
7175      pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
7176      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
7177    branch_chain *bcptr, compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
7178  {  {
7179  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
# Line 6936  pcre_uchar *start_bracket = code; Line 7183  pcre_uchar *start_bracket = code;
7183  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
7184  open_capitem capitem;  open_capitem capitem;
7185  int capnumber = 0;  int capnumber = 0;
7186  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7187  pcre_int32 branchfirstchar, branchreqchar;  pcre_int32 firstcharflags, reqcharflags;
7188    pcre_uint32 branchfirstchar, branchreqchar;
7189    pcre_int32 branchfirstcharflags, branchreqcharflags;
7190  int length;  int length;
7191  int orig_bracount;  unsigned int orig_bracount;
7192  int max_bracount;  unsigned int max_bracount;
7193  branch_chain bc;  branch_chain bc;
7194    
7195  bc.outer = bcptr;  bc.outer = bcptr;
7196  bc.current_branch = code;  bc.current_branch = code;
7197    
7198  firstchar = reqchar = REQ_UNSET;  firstchar = reqchar = 0;
7199    firstcharflags = reqcharflags = REQ_UNSET;
7200    
7201  /* 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
7202  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 7006  for (;;) Line 7256  for (;;)
7256    into the length. */    into the length. */
7257    
7258    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
7259          &branchreqchar, &bc, cond_depth, cd,          &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
7260          (lengthptr == NULL)? NULL : &length))          cond_depth, cd, (lengthptr == NULL)? NULL : &length))
7261      {      {
7262      *ptrptr = ptr;      *ptrptr = ptr;
7263      return FALSE;      return FALSE;
# Line 7028  for (;;) Line 7278  for (;;)
7278      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
7279        {        {
7280        firstchar = branchfirstchar;        firstchar = branchfirstchar;
7281          firstcharflags = branchfirstcharflags;
7282        reqchar = branchreqchar;        reqchar = branchreqchar;
7283          reqcharflags = branchreqcharflags;
7284        }        }
7285    
7286      /* 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 7042  for (;;) Line 7294  for (;;)
7294        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
7295        previously no reqchar, it takes on the value of the old firstchar. */        previously no reqchar, it takes on the value of the old firstchar. */
7296    
7297        if (firstchar >= 0 && firstchar != branchfirstchar)        if (firstcharflags >= 0 &&
7298              (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar))
7299          {          {
7300          if (reqchar < 0) reqchar = firstchar;          if (reqcharflags < 0)
7301          firstchar = REQ_NONE;            {
7302              reqchar = firstchar;
7303              reqcharflags = firstcharflags;
7304              }
7305            firstcharflags = REQ_NONE;
7306          }          }
7307    
7308        /* 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
7309        branch becomes a reqchar if there isn't a branch reqchar. */        branch becomes a reqchar if there isn't a branch reqchar. */
7310    
7311        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)        if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
7312            branchreqchar = branchfirstchar;          {
7313            branchreqchar = branchfirstchar;
7314            branchreqcharflags = branchfirstcharflags;
7315            }
7316    
7317        /* Now ensure that the reqchars match */        /* Now ensure that the reqchars match */
7318    
7319        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))        if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) ||
7320          reqchar = REQ_NONE;            reqchar != branchreqchar)
7321        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */          reqcharflags = REQ_NONE;
7322          else
7323            {
7324            reqchar = branchreqchar;
7325            reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */
7326            }
7327        }        }
7328    
7329      /* 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 7154  for (;;) Line 7419  for (;;)
7419      *codeptr = code;      *codeptr = code;
7420      *ptrptr = ptr;      *ptrptr = ptr;
7421      *firstcharptr = firstchar;      *firstcharptr = firstchar;
7422        *firstcharflagsptr = firstcharflags;
7423      *reqcharptr = reqchar;      *reqcharptr = reqchar;
7424        *reqcharflagsptr = reqcharflags;
7425      if (lengthptr != NULL)      if (lengthptr != NULL)
7426        {        {
7427        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7389  do { Line 7656  do {
7656       {       {
7657       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7658       }       }
7659    
7660     /* Atomic brackets */     /* Atomic brackets */
7661    
7662     else if (op == OP_ONCE || op == OP_ONCE_NC)     else if (op == OP_ONCE || op == OP_ONCE_NC)
# Line 7435  return TRUE; Line 7702  return TRUE;
7702  discarded, because they can cause conflicts with actual literals that follow.  discarded, because they can cause conflicts with actual literals that follow.
7703  However, if we end up without a first char setting for an unanchored pattern,  However, if we end up without a first char setting for an unanchored pattern,
7704  it is worth scanning the regex to see if there is an initial asserted first  it is worth scanning the regex to see if there is an initial asserted first
7705  char. If all branches start with the same asserted char, or with a bracket all  char. If all branches start with the same asserted char, or with a
7706  of whose alternatives start with the same asserted char (recurse ad lib), then  non-conditional bracket all of whose alternatives start with the same asserted
7707  we return that char, otherwise -1.  char (recurse ad lib), then we return that char, with the flags set to zero or
7708    REQ_CASELESS; otherwise return zero with REQ_NONE in the flags.
7709    
7710  Arguments:  Arguments:
7711    code       points to start of expression (the bracket)    code       points to start of expression (the bracket)
7712      flags      points to the first char flags, or to REQ_NONE
7713    inassert   TRUE if in an assertion    inassert   TRUE if in an assertion
7714    
7715  Returns:     -1 or the fixed first char  Returns:     the fixed first char, or 0 with REQ_NONE in flags
7716  */  */
7717    
7718  static int  static pcre_uint32
7719  find_firstassertedchar(const pcre_uchar *code, BOOL inassert)  find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags,
7720      BOOL inassert)
7721  {  {
7722  register int c = -1;  register pcre_uint32 c = 0;
7723    int cflags = REQ_NONE;
7724    
7725    *flags = REQ_NONE;
7726  do {  do {
7727     int d;     pcre_uint32 d;
7728       int dflags;
7729     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7730               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7731     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7732       TRUE);       TRUE);
7733     register int op = *scode;     register pcre_uchar op = *scode;
7734    
7735     switch(op)     switch(op)
7736       {       {
7737       default:       default:
7738       return -1;       return 0;
7739    
7740       case OP_BRA:       case OP_BRA:
7741       case OP_BRAPOS:       case OP_BRAPOS:
# Line 7472  do { Line 7746  do {
7746       case OP_ASSERT:       case OP_ASSERT:
7747       case OP_ONCE:       case OP_ONCE:
7748       case OP_ONCE_NC:       case OP_ONCE_NC:
7749       case OP_COND:       d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
7750       if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)       if (dflags < 0)
7751         return -1;         return 0;
7752       if (c < 0) c = d; else if (c != d) return -1;       if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0;
7753       break;       break;
7754    
7755       case OP_EXACT:       case OP_EXACT:
# Line 7486  do { Line 7760  do {
7760       case OP_PLUS:       case OP_PLUS:
7761       case OP_MINPLUS:       case OP_MINPLUS:
7762       case OP_POSPLUS:       case OP_POSPLUS:
7763       if (!inassert) return -1;       if (!inassert) return 0;
7764       if (c < 0) c = scode[1];       if (cflags < 0) { c = scode[1]; cflags = 0; }
7765         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7766       break;       break;
7767    
7768       case OP_EXACTI:       case OP_EXACTI:
# Line 7499  do { Line 7773  do {
7773       case OP_PLUSI:       case OP_PLUSI:
7774       case OP_MINPLUSI:       case OP_MINPLUSI:
7775       case OP_POSPLUSI:       case OP_POSPLUSI:
7776       if (!inassert) return -1;       if (!inassert) return 0;
7777       if (c < 0) c = scode[1] | REQ_CASELESS;       if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
7778         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7779       break;       break;
7780       }       }
7781    
7782     code += GET(code, 1);     code += GET(code, 1);
7783     }     }
7784  while (*code == OP_ALT);  while (*code == OP_ALT);
7785    
7786    *flags = cflags;
7787  return c;  return c;
7788  }  }
7789    
# Line 7535  Returns:        pointer to compiled data Line 7811  Returns:        pointer to compiled data
7811                  with errorptr and erroroffset set                  with errorptr and erroroffset set
7812  */  */
7813    
7814  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7815  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7816  pcre_compile(const char *pattern, int options, const char **errorptr,  pcre_compile(const char *pattern, int options, const char **errorptr,
7817    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7818  #else  #elif defined COMPILE_PCRE16
7819  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
7820  pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,  pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,
7821    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7822    #elif defined COMPILE_PCRE32
7823    PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
7824    pcre32_compile(PCRE_SPTR32 pattern, int options, const char **errorptr,
7825      int *erroroffset, const unsigned char *tables)
7826  #endif  #endif
7827  {  {
7828  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7829  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7830  #else  #elif defined COMPILE_PCRE16
7831  return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7832    #elif defined COMPILE_PCRE32
7833    return pcre32_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7834  #endif  #endif
7835  }  }
7836    
7837    
7838  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7839  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7840  pcre_compile2(const char *pattern, int options, int *errorcodeptr,  pcre_compile2(const char *pattern, int options, int *errorcodeptr,
7841    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7842  #else  #elif defined COMPILE_PCRE16
7843  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
7844  pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,  pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,
7845    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7846    #elif defined COMPILE_PCRE32
7847    PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
7848    pcre32_compile2(PCRE_SPTR32 pattern, int options, int *errorcodeptr,
7849      const char **errorptr, int *erroroffset, const unsigned char *tables)
7850  #endif  #endif
7851  {  {
7852  REAL_PCRE *re;  REAL_PCRE *re;
7853  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7854  pcre_int32 firstchar, reqchar;  pcre_int32 firstcharflags, reqcharflags;
7855    pcre_uint32 firstchar, reqchar;
7856    pcre_uint32 limit_match = PCRE_UINT32_MAX;
7857    pcre_uint32 limit_recursion = PCRE_UINT32_MAX;
7858  int newline;  int newline;
7859  int errorcode = 0;  int errorcode = 0;
7860  int skipatstart = 0;  int skipatstart = 0;
7861  BOOL utf;  BOOL utf;
7862    BOOL never_utf = FALSE;
7863  size_t size;  size_t size;
7864  pcre_uchar *code;  pcre_uchar *code;
7865  const pcre_uchar *codestart;  const pcre_uchar *codestart;
# Line 7629  if ((options & ~PUBLIC_COMPILE_OPTIONS) Line 7919  if ((options & ~PUBLIC_COMPILE_OPTIONS)
7919    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
7920    }    }
7921    
7922    /* If PCRE_NEVER_UTF is set, remember it. */
7923    
7924    if ((options & PCRE_NEVER_UTF) != 0) never_utf = TRUE;
7925    
7926  /* Check for global one-time settings at the start of the pattern, and remember  /* Check for global one-time settings at the start of the pattern, and remember
7927  the offset for later. */  the offset for later. */
7928    
7929    cd->external_flags = 0;   /* Initialize here for LIMIT_MATCH/RECURSION */
7930    
7931  while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS &&  while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS &&
7932         ptr[skipatstart+1] == CHAR_ASTERISK)         ptr[skipatstart+1] == CHAR_ASTERISK)
7933    {    {
7934    int newnl = 0;    int newnl = 0;
7935    int newbsr = 0;    int newbsr = 0;
7936    
7937    /* For completeness and backward compatibility, (*UTFn) is supported in the
7938    relevant libraries, but (*UTF) is generic and always supported. Note that
7939    PCRE_UTF8 == PCRE_UTF16 == PCRE_UTF32. */
7940    
7941  #ifdef COMPILE_PCRE8  #ifdef COMPILE_PCRE8
7942    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 5) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF8_RIGHTPAR, 5) == 0)
7943      { skipatstart += 7; options |= PCRE_UTF8; continue; }      { skipatstart += 7; options |= PCRE_UTF8; continue; }
7944  #endif  #endif
7945  #ifdef COMPILE_PCRE16  #ifdef COMPILE_PCRE16
7946    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF16_RIGHTPAR, 6) == 0)
7947      { skipatstart += 8; options |= PCRE_UTF16; continue; }      { skipatstart += 8; options |= PCRE_UTF16; continue; }
7948  #endif  #endif
7949    #ifdef COMPILE_PCRE32
7950      if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF32_RIGHTPAR, 6) == 0)
7951        { skipatstart += 8; options |= PCRE_UTF32; continue; }
7952    #endif
7953    
7954      else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 4) == 0)
7955        { skipatstart += 6; options |= PCRE_UTF8; continue; }
7956    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)
7957      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
7958    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)
7959      { skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; }      { skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; }
7960    
7961      else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LIMIT_MATCH_EQ, 12) == 0)
7962        {
7963        pcre_uint32 c = 0;
7964        int p = skipatstart + 14;
7965        while (isdigit(ptr[p]))
7966          {
7967          if (c > PCRE_UINT32_MAX / 10 - 1) break;   /* Integer overflow */
7968          c = c*10 + ptr[p++] - CHAR_0;
7969          }
7970        if (ptr[p++] != CHAR_RIGHT_PARENTHESIS) break;
7971        if (c < limit_match)
7972          {
7973          limit_match = c;
7974          cd->external_flags |= PCRE_MLSET;
7975          }
7976        skipatstart = p;
7977        continue;
7978        }
7979    
7980      else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LIMIT_RECURSION_EQ, 16) == 0)
7981        {
7982        pcre_uint32 c = 0;
7983        int p = skipatstart + 18;
7984        while (isdigit(ptr[p]))
7985          {
7986          if (c > PCRE_UINT32_MAX / 10 - 1) break;   /* Integer overflow check */
7987          c = c*10 + ptr[p++] - CHAR_0;
7988          }
7989        if (ptr[p++] != CHAR_RIGHT_PARENTHESIS) break;
7990        if (c < limit_recursion)
7991          {
7992          limit_recursion = c;
7993          cd->external_flags |= PCRE_RLSET;
7994          }
7995        skipatstart = p;
7996        continue;
7997        }
7998    
7999    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CR_RIGHTPAR, 3) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CR_RIGHTPAR, 3) == 0)
8000      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }
8001    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LF_RIGHTPAR, 3)  == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LF_RIGHTPAR, 3)  == 0)
# Line 7674  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 8019  while (ptr[skipatstart] == CHAR_LEFT_PAR
8019    else break;    else break;
8020    }    }
8021    
8022  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF(16|32) have the same value as PCRE_UTF8. */
8023  utf = (options & PCRE_UTF8) != 0;  utf = (options & PCRE_UTF8) != 0;
8024    if (utf && never_utf)
8025      {
8026      errorcode = ERR78;
8027      goto PCRE_EARLY_ERROR_RETURN2;
8028      }
8029    
8030  /* 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
8031  return of an error code from PRIV(valid_utf)() is a new feature, introduced in  return of an error code from PRIV(valid_utf)() is a new feature, introduced in
# Line 7686  not used here. */ Line 8036  not used here. */
8036  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&
8037       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)
8038    {    {
8039  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8040    errorcode = ERR44;    errorcode = ERR44;
8041  #else  #elif defined COMPILE_PCRE16
8042    errorcode = ERR74;    errorcode = ERR74;
8043    #elif defined COMPILE_PCRE32
8044      errorcode = ERR77;
8045  #endif  #endif
8046    goto PCRE_EARLY_ERROR_RETURN2;    goto PCRE_EARLY_ERROR_RETURN2;
8047    }    }
# Line 7796  cd->req_varyopt = 0; Line 8148  cd->req_varyopt = 0;
8148  cd->assert_depth = 0;  cd->assert_depth = 0;
8149  cd->max_lookbehind = 0;  cd->max_lookbehind = 0;
8150  cd->external_options = options;  cd->external_options = options;
 cd->external_flags = 0;  
8151  cd->open_caps = NULL;  cd->open_caps = NULL;
8152    
8153  /* Now do the pre-compile. On error, errorcode will be set non-zero, so we  /* Now do the pre-compile. On error, errorcode will be set non-zero, so we
# Line 7809  ptr += skipatstart; Line 8160  ptr += skipatstart;
8160  code = cworkspace;  code = cworkspace;
8161  *code = OP_BRA;  *code = OP_BRA;
8162  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8163    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8164      cd, &length);
8165  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
8166    
8167  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 7845  re->magic_number = MAGIC_NUMBER; Line 8197  re->magic_number = MAGIC_NUMBER;
8197  re->size = (int)size;  re->size = (int)size;
8198  re->options = cd->external_options;  re->options = cd->external_options;
8199  re->flags = cd->external_flags;  re->flags = cd->external_flags;
8200    re->limit_match = limit_match;
8201    re->limit_recursion = limit_recursion;
8202  re->first_char = 0;  re->first_char = 0;
8203  re->req_char = 0;  re->req_char = 0;
8204  re->name_table_offset = sizeof(REAL_PCRE) / sizeof(pcre_uchar);  re->name_table_offset = sizeof(REAL_PCRE) / sizeof(pcre_uchar);
# Line 7853  re->name_count = cd->names_found; Line 8207  re->name_count = cd->names_found;
8207  re->ref_count = 0;  re->ref_count = 0;
8208  re->tables = (tables == PRIV(default_tables))? NULL : tables;  re->tables = (tables == PRIV(default_tables))? NULL : tables;
8209  re->nullpad = NULL;  re->nullpad = NULL;
8210    #ifdef COMPILE_PCRE32
8211    re->dummy = 0;
8212    #else
8213    re->dummy1 = re->dummy2 = re->dummy3 = 0;
8214    #endif
8215    
8216  /* 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
8217  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 7884  ptr = (const pcre_uchar *)pattern + skip Line 8243  ptr = (const pcre_uchar *)pattern + skip
8243  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
8244  *code = OP_BRA;  *code = OP_BRA;
8245  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
8246    &firstchar, &reqchar, NULL, cd, NULL);    &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
8247  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
8248  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
8249  re->max_lookbehind = cd->max_lookbehind;  re->max_lookbehind = cd->max_lookbehind;
8250  re->flags = cd->external_flags | PCRE_MODE;  re->flags = cd->external_flags | PCRE_MODE;
8251    
8252  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept)
8253      {
8254      reqchar = 0;              /* Must disable after (*ACCEPT) */
8255      reqcharflags = REQ_NONE;
8256      }
8257    
8258  /* 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. */
8259    
8260  if (errorcode == 0 && *ptr != 0) errorcode = ERR22;  if (errorcode == 0 && *ptr != CHAR_NULL) errorcode = ERR22;
8261    
8262  /* Fill in the terminating state and check for disastrous overflow, but  /* Fill in the terminating state and check for disastrous overflow, but
8263  if debugging, leave the test till after things are printed out. */  if debugging, leave the test till after things are printed out. */
# Line 7905  if debugging, leave the test till after Line 8268  if debugging, leave the test till after
8268  if (code - codestart > length) errorcode = ERR23;  if (code - codestart > length) errorcode = ERR23;
8269  #endif  #endif
8270    
8271    #ifdef SUPPORT_VALGRIND
8272    /* If the estimated length exceeds the really used length, mark the extra
8273    allocated memory as unaddressable, so that any out-of-bound reads can be
8274    detected. */
8275    VALGRIND_MAKE_MEM_NOACCESS(code, (length - (code - codestart)) * sizeof(pcre_uchar));
8276    #endif
8277    
8278  /* Fill in any forward references that are required. There may be repeated  /* Fill in any forward references that are required. There may be repeated
8279  references; optimize for them, as searching a large regex takes time. */  references; optimize for them, as searching a large regex takes time. */
8280    
# Line 7928  if (cd->hwm > cd->start_workspace) Line 8298  if (cd->hwm > cd->start_workspace)
8298      }      }
8299    }    }
8300    
8301  /* If the workspace had to be expanded, free the new memory. */  /* If the workspace had to be expanded, free the new memory. Set the pointer to
8302    NULL to indicate that forward references have been filled in. */
8303    
8304  if (cd->workspace_size > COMPILE_WORK_SIZE)  if (cd->workspace_size > COMPILE_WORK_SIZE)
8305    (PUBL(free))((void *)cd->start_workspace);    (PUBL(free))((void *)cd->start_workspace);
8306    cd->start_workspace = NULL;
8307    
8308  /* Give an error if there's back reference to a non-existent capturing  /* Give an error if there's back reference to a non-existent capturing
8309  subpattern. */  subpattern. */
# Line 7940  if (errorcode == 0 && re->top_backref > Line 8312  if (errorcode == 0 && re->top_backref >
8312    
8313  /* If there were any lookbehind assertions that contained OP_RECURSE  /* If there were any lookbehind assertions that contained OP_RECURSE
8314  (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,
8315  because they may contain forward references. Actual recursions can't be fixed  because they may contain forward references. Actual recursions cannot be fixed
8316  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
8317  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
8318  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 8011  if ((re->options & PCRE_ANCHORED) == 0) Line 8383  if ((re->options & PCRE_ANCHORED) == 0)
8383    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
8384    else    else
8385      {      {
8386      if (firstchar < 0)      if (firstcharflags < 0)
8387        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
8388      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstcharflags >= 0)   /* Remove caseless flag for non-caseable chars */
8389        {        {
8390  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8391        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8392  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
8393        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8394    #elif defined COMPILE_PCRE32
8395          re->first_char = firstchar;
8396  #endif  #endif
8397  #endif        if ((firstcharflags & REQ_CASELESS) != 0)
       if ((firstchar & REQ_CASELESS) != 0)  
8398          {          {
8399  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8400          /* We ignore non-ASCII first chars in 8 bit mode. */          /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8054  if ((re->options & PCRE_ANCHORED) == 0) Line 8426  if ((re->options & PCRE_ANCHORED) == 0)
8426  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
8427  bytes. */  bytes. */
8428    
8429  if (reqchar >= 0 &&  if (reqcharflags >= 0 &&
8430       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0))
8431    {    {
8432  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8433    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8434  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
8435    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8436    #elif defined COMPILE_PCRE32
8437      re->req_char = reqchar;
8438  #endif  #endif
8439  #endif    if ((reqcharflags & REQ_CASELESS) != 0)
   if ((reqchar & REQ_CASELESS) != 0)  
8440      {      {
8441  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8442      /* We ignore non-ASCII first chars in 8 bit mode. */      /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8114  if ((re->flags & PCRE_REQCHSET) != 0) Line 8486  if ((re->flags & PCRE_REQCHSET) != 0)
8486      else printf("Req char = \\x%02x%s\n", ch, caseless);      else printf("Req char = \\x%02x%s\n", ch, caseless);
8487    }    }
8488    
8489  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8490  pcre_printint((pcre *)re, stdout, TRUE);  pcre_printint((pcre *)re, stdout, TRUE);
8491  #else  #elif defined COMPILE_PCRE16
8492  pcre16_printint((pcre *)re, stdout, TRUE);  pcre16_printint((pcre *)re, stdout, TRUE);
8493    #elif defined COMPILE_PCRE32
8494    pcre32_printint((pcre *)re, stdout, TRUE);
8495  #endif  #endif
8496    
8497  /* 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 8133  if (code - codestart > length) Line 8507  if (code - codestart > length)
8507    }    }
8508  #endif   /* PCRE_DEBUG */  #endif   /* PCRE_DEBUG */
8509    
8510  #ifdef COMPILE_PCRE8  /* Check for a pattern than can match an empty string, so that this information
8511    can be provided to applications. */
8512    
8513    do
8514      {
8515      if (could_be_empty_branch(codestart, code, utf, cd, NULL))
8516        {
8517        re->flags |= PCRE_MATCH_EMPTY;
8518        break;
8519        }
8520      codestart += GET(codestart, 1);
8521      }
8522    while (*codestart == OP_ALT);
8523    
8524    #if defined COMPILE_PCRE8
8525  return (pcre *)re;  return (pcre *)re;
8526  #else  #elif defined COMPILE_PCRE16
8527  return (pcre16 *)re;  return (pcre16 *)re;
8528    #elif defined COMPILE_PCRE32
8529    return (pcre32 *)re;
8530  #endif  #endif
8531  }  }
8532    

Legend:
Removed from v.1045  
changed lines
  Added in v.1352

  ViewVC Help
Powered by ViewVC 1.1.5