/[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 1133 by chpe, Thu Oct 18 18:35:30 2012 UTC
# Line 53  supporting internal functions that are n Line 53  supporting internal functions that are n
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    
56  /* When PCRE_DEBUG is defined, we need the pcre(16)_printint() function, which  /* When PCRE_DEBUG is defined, we need the pcre(16|32)_printint() function, which
57  is also used by pcretest. PCRE_DEBUG is not defined when building a production  is also used by pcretest. PCRE_DEBUG is not defined when building a production
58  library. We do not need to select pcre16_printint.c specially, because the  library. We do not need to select pcre16_printint.c specially, because the
59  COMPILE_PCREx macro will already be appropriately set. */  COMPILE_PCREx macro will already be appropriately set. */
# Line 84  static int Line 84  static int
84      const pcre_uint32 *, unsigned int);      const pcre_uint32 *, unsigned int);
85    
86  static BOOL  static BOOL
87    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL,    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int,
88      int, int, int *, int *, branch_chain *, compile_data *, int *);      pcre_uint32 *, pcre_int32 *, pcre_uint32 *, pcre_int32 *, branch_chain *,
89        compile_data *, int *);
90    
91    
92    
# Line 121  overrun before it actually does run off Line 122  overrun before it actually does run off
122    
123  /* Private flags added to firstchar and reqchar. */  /* Private flags added to firstchar and reqchar. */
124    
125  #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */  #define REQ_CASELESS    (1 << 0)        /* Indicates caselessness */
126  #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */  #define REQ_VARY        (1 << 1)        /* Reqchar followed non-literal item */
127    /* 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 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    ;    ;
512    
513  /* Table to identify digits and hex digits. This is used when compiling  /* Table to identify digits and hex digits. This is used when compiling
# Line 643  static const pcre_uint8 ebcdic_chartab[] Line 648  static const pcre_uint8 ebcdic_chartab[]
648    
649    
650    
   
651  /*************************************************  /*************************************************
652  *            Find an error text                  *  *            Find an error text                  *
653  *************************************************/  *************************************************/
# Line 747  return (*p == CHAR_RIGHT_CURLY_BRACKET); Line 751  return (*p == CHAR_RIGHT_CURLY_BRACKET);
751  *************************************************/  *************************************************/
752    
753  /* This function is called when a \ has been encountered. It either returns a  /* This function is called when a \ has been encountered. It either returns a
754  positive value for a simple escape such as \n, or a negative value which  positive value for a simple escape such as \n, or 0 for a data character
755  encodes one of the more complicated things such as \d. A backreference to group  which will be placed in chptr. A backreference to group n is returned as
756  n is returned as -(ESC_REF + n); ESC_REF is the highest ESC_xxx macro. When  negative n. When UTF-8 is enabled, a positive value greater than 255 may
757  UTF-8 is enabled, a positive value greater than 255 may be returned. On entry,  be returned in chptr.
758  ptr is pointing at the \. On exit, it is on the final character of the escape  On entry,ptr is pointing at the \. On exit, it is on the final character of the
759  sequence.  escape sequence.
760    
761  Arguments:  Arguments:
762    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
763      chptr          points to the data character
764    errorcodeptr   points to the errorcode variable    errorcodeptr   points to the errorcode variable
765    bracount       number of previous extracting brackets    bracount       number of previous extracting brackets
766    options        the options bits    options        the options bits
767    isclass        TRUE if inside a character class    isclass        TRUE if inside a character class
768    
769  Returns:         zero or positive => a data character  Returns:         zero => a data character
770                   negative => a special escape sequence                   positive => a special escape sequence
771                     negative => a back reference
772                   on error, errorcodeptr is set                   on error, errorcodeptr is set
773  */  */
774    
775  static int  static int
776  check_escape(const pcre_uchar **ptrptr, int *errorcodeptr, int bracount,  check_escape(const pcre_uchar **ptrptr, pcre_uint32 *chptr, int *errorcodeptr,
777    int options, BOOL isclass)    int bracount, int options, BOOL isclass)
778  {  {
779  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
780  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
781  const pcre_uchar *ptr = *ptrptr + 1;  const pcre_uchar *ptr = *ptrptr + 1;
782  pcre_int32 c;  pcre_uint32 c;
783    int escape = 0;
784  int i;  int i;
785    
786  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */
# Line 790  Otherwise further processing may be requ Line 797  Otherwise further processing may be requ
797  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
798  /* Not alphanumeric */  /* Not alphanumeric */
799  else if (c < CHAR_0 || c > CHAR_z) {}  else if (c < CHAR_0 || c > CHAR_z) {}
800  else if ((i = escapes[c - CHAR_0]) != 0) c = i;  else if ((i = escapes[c - CHAR_0]) != 0) { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
801    
802  #else           /* EBCDIC coding */  #else           /* EBCDIC coding */
803  /* Not alphanumeric */  /* Not alphanumeric */
804  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}
805  else if ((i = escapes[c - 0x48]) != 0)  c = i;  else if ((i = escapes[c - 0x48]) != 0)  { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
806  #endif  #endif
807    
808  /* Escapes that need further processing, or are illegal. */  /* Escapes that need further processing, or are illegal. */
# Line 803  else if ((i = escapes[c - 0x48]) != 0) Line 810  else if ((i = escapes[c - 0x48]) != 0)
810  else  else
811    {    {
812    const pcre_uchar *oldptr;    const pcre_uchar *oldptr;
813    BOOL braced, negated;    BOOL braced, negated, overflow;
814      int s;
815    
816    switch (c)    switch (c)
817      {      {
# Line 828  else Line 836  else
836          c = 0;          c = 0;
837          for (i = 0; i < 4; ++i)          for (i = 0; i < 4; ++i)
838            {            {
839            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
840  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
841            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
842            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 838  else Line 846  else
846  #endif  #endif
847            }            }
848    
849  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
850          if (c > (utf ? 0x10ffff : 0xff))          if (c > (utf ? 0x10ffff : 0xff))
851  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
852          if (c > (utf ? 0x10ffff : 0xffff))          if (c > (utf ? 0x10ffff : 0xffff))
853  #endif  #elif defined COMPILE_PCRE32
854            if (utf && c > 0x10ffff)
855  #endif  #endif
856            {            {
857            *errorcodeptr = ERR76;            *errorcodeptr = ERR76;
# Line 875  else Line 883  else
883      (3) For Oniguruma compatibility we also support \g followed by a name or a      (3) For Oniguruma compatibility we also support \g followed by a name or a
884      number either in angle brackets or in single quotes. However, these are      number either in angle brackets or in single quotes. However, these are
885      (possibly recursive) subroutine calls, _not_ backreferences. Just return      (possibly recursive) subroutine calls, _not_ backreferences. Just return
886      the -ESC_g code (cf \k). */      the ESC_g code (cf \k). */
887    
888      case CHAR_g:      case CHAR_g:
889      if (isclass) break;      if (isclass) break;
890      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)
891        {        {
892        c = -ESC_g;        escape = ESC_g;
893        break;        break;
894        }        }
895    
# Line 894  else Line 902  else
902          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;
903        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)
904          {          {
905          c = -ESC_k;          escape = ESC_k;
906          break;          break;
907          }          }
908        braced = TRUE;        braced = TRUE;
# Line 910  else Line 918  else
918      else negated = FALSE;      else negated = FALSE;
919    
920      /* The integer range is limited by the machine's int representation. */      /* The integer range is limited by the machine's int representation. */
921      c = 0;      s = 0;
922        overflow = FALSE;
923      while (IS_DIGIT(ptr[1]))      while (IS_DIGIT(ptr[1]))
924        {        {
925        if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */        if (s > INT_MAX / 10 - 1) /* Integer overflow */
926          {          {
927          c = -1;          overflow = TRUE;
928          break;          break;
929          }          }
930        c = c * 10 + *(++ptr) - CHAR_0;        s = s * 10 + (int)(*(++ptr) - CHAR_0);
931        }        }
932      if (((unsigned int)c) > INT_MAX) /* Integer overflow */      if (overflow) /* Integer overflow */
933        {        {
934        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
935          ptr++;          ptr++;
# Line 934  else Line 943  else
943        break;        break;
944        }        }
945    
946      if (c == 0)      if (s == 0)
947        {        {
948        *errorcodeptr = ERR58;        *errorcodeptr = ERR58;
949        break;        break;
# Line 942  else Line 951  else
951    
952      if (negated)      if (negated)
953        {        {
954        if (c > bracount)        if (s > bracount)
955          {          {
956          *errorcodeptr = ERR15;          *errorcodeptr = ERR15;
957          break;          break;
958          }          }
959        c = bracount - (c - 1);        s = bracount - (s - 1);
960        }        }
961    
962      c = -(ESC_REF + c);      escape = -s;
963      break;      break;
964    
965      /* The handling of escape sequences consisting of a string of digits      /* The handling of escape sequences consisting of a string of digits
# Line 972  else Line 981  else
981        {        {
982        oldptr = ptr;        oldptr = ptr;
983        /* The integer range is limited by the machine's int representation. */        /* The integer range is limited by the machine's int representation. */
984        c -= CHAR_0;        s = (int)(c -CHAR_0);
985          overflow = FALSE;
986        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
987          {          {
988          if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */          if (s > INT_MAX / 10 - 1) /* Integer overflow */
989            {            {
990            c = -1;            overflow = TRUE;
991            break;            break;
992            }            }
993          c = c * 10 + *(++ptr) - CHAR_0;          s = s * 10 + (int)(*(++ptr) - CHAR_0);
994          }          }
995        if (((unsigned int)c) > INT_MAX) /* Integer overflow */        if (overflow) /* Integer overflow */
996          {          {
997          while (IS_DIGIT(ptr[1]))          while (IS_DIGIT(ptr[1]))
998            ptr++;            ptr++;
999          *errorcodeptr = ERR61;          *errorcodeptr = ERR61;
1000          break;          break;
1001          }          }
1002        if (c < 10 || c <= bracount)        if (s < 10 || s <= bracount)
1003          {          {
1004          c = -(ESC_REF + c);          escape = -s;
1005          break;          break;
1006          }          }
1007        ptr = oldptr;      /* Put the pointer back and fall through */        ptr = oldptr;      /* Put the pointer back and fall through */
# Line 1038  else Line 1048  else
1048          c = 0;          c = 0;
1049          for (i = 0; i < 2; ++i)          for (i = 0; i < 2; ++i)
1050            {            {
1051            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
1052  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1053            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1054            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 1056  else Line 1066  else
1066        const pcre_uchar *pt = ptr + 2;        const pcre_uchar *pt = ptr + 2;
1067    
1068        c = 0;        c = 0;
1069          overflow = FALSE;
1070        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)
1071          {          {
1072          register int cc = *pt++;          register pcre_uint32 cc = *pt++;
1073          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */
1074    
1075    #ifdef COMPILE_PCRE32
1076            if (c >= 0x10000000l) { overflow = TRUE; break; }
1077    #endif
1078    
1079  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1080          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1081          c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 1069  else Line 1084  else
1084          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
1085  #endif  #endif
1086    
1087  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1088          if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }          if (c > (utf ? 0x10ffff : 0xff)) { overflow = TRUE; break; }
1089  #else  #elif defined COMPILE_PCRE16
1090  #ifdef COMPILE_PCRE16          if (c > (utf ? 0x10ffff : 0xffff)) { overflow = TRUE; break; }
1091          if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }  #elif defined COMPILE_PCRE32
1092  #endif          if (utf && c > 0x10ffff) { overflow = TRUE; break; }
1093  #endif  #endif
1094          }          }
1095    
1096        if (c < 0)        if (overflow)
1097          {          {
1098          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;
1099          *errorcodeptr = ERR34;          *errorcodeptr = ERR34;
# Line 1100  else Line 1115  else
1115      c = 0;      c = 0;
1116      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)
1117        {        {
1118        int cc;                                  /* Some compilers don't like */        pcre_uint32 cc;                          /* Some compilers don't like */
1119        cc = *(++ptr);                           /* ++ in initializers */        cc = *(++ptr);                           /* ++ in initializers */
1120  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1121        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */
# Line 1159  else Line 1174  else
1174  newline". PCRE does not support \N{name}. However, it does support  newline". PCRE does not support \N{name}. However, it does support
1175  quantification such as \N{2,3}. */  quantification such as \N{2,3}. */
1176    
1177  if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&  if (escape == ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&
1178       !is_counted_repeat(ptr+2))       !is_counted_repeat(ptr+2))
1179    *errorcodeptr = ERR37;    *errorcodeptr = ERR37;
1180    
1181  /* If PCRE_UCP is set, we change the values for \d etc. */  /* If PCRE_UCP is set, we change the values for \d etc. */
1182    
1183  if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w)  if ((options & PCRE_UCP) != 0 && escape >= ESC_D && escape <= ESC_w)
1184    c -= (ESC_DU - ESC_D);    escape += (ESC_DU - ESC_D);
1185    
1186  /* Set the pointer to the final character before returning. */  /* Set the pointer to the final character before returning. */
1187    
1188  *ptrptr = ptr;  *ptrptr = ptr;
1189  return c;  *chptr = c;
1190    return escape;
1191  }  }
1192    
   
   
1193  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1194  /*************************************************  /*************************************************
1195  *               Handle \P and \p                 *  *               Handle \P and \p                 *
# Line 1189  escape sequence. Line 1203  escape sequence.
1203  Argument:  Argument:
1204    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
1205    negptr         points to a boolean that is set TRUE for negation else FALSE    negptr         points to a boolean that is set TRUE for negation else FALSE
1206    dptr           points to an int that is set to the detailed property value    ptypeptr       points to an unsigned int that is set to the type value
1207      pdataptr       points to an unsigned int that is set to the detailed property value
1208    errorcodeptr   points to the error code variable    errorcodeptr   points to the error code variable
1209    
1210  Returns:         type value from ucp_type_table, or -1 for an invalid type  Returns:         TRUE if the type value was found, or FALSE for an invalid type
1211  */  */
1212    
1213  static int  static BOOL
1214  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr)  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, unsigned int *ptypeptr,
1215      unsigned int *pdataptr, int *errorcodeptr)
1216  {  {
1217  int c, i, bot, top;  pcre_uchar c;
1218    int i, bot, top;
1219  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
1220  pcre_uchar name[32];  pcre_uchar name[32];
1221    
# Line 1245  top = PRIV(utt_size); Line 1262  top = PRIV(utt_size);
1262    
1263  while (bot < top)  while (bot < top)
1264    {    {
1265      int r;
1266    i = (bot + top) >> 1;    i = (bot + top) >> 1;
1267    c = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);    r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);
1268    if (c == 0)    if (r == 0)
1269      {      {
1270      *dptr = PRIV(utt)[i].value;      *ptypeptr = PRIV(utt)[i].type;
1271      return PRIV(utt)[i].type;      *pdataptr = PRIV(utt)[i].value;
1272        return TRUE;
1273      }      }
1274    if (c > 0) bot = i + 1; else top = i;    if (r > 0) bot = i + 1; else top = i;
1275    }    }
1276    
1277  *errorcodeptr = ERR47;  *errorcodeptr = ERR47;
1278  *ptrptr = ptr;  *ptrptr = ptr;
1279  return -1;  return FALSE;
1280    
1281  ERROR_RETURN:  ERROR_RETURN:
1282  *errorcodeptr = ERR46;  *errorcodeptr = ERR46;
1283  *ptrptr = ptr;  *ptrptr = ptr;
1284  return -1;  return FALSE;
1285  }  }
1286  #endif  #endif
1287    
# Line 1297  int max = -1; Line 1316  int max = -1;
1316  /* Read the minimum value and do a paranoid check: a negative value indicates  /* Read the minimum value and do a paranoid check: a negative value indicates
1317  an integer overflow. */  an integer overflow. */
1318    
1319  while (IS_DIGIT(*p)) min = min * 10 + *p++ - CHAR_0;  while (IS_DIGIT(*p)) min = min * 10 + (int)(*p++ - CHAR_0);
1320  if (min < 0 || min > 65535)  if (min < 0 || min > 65535)
1321    {    {
1322    *errorcodeptr = ERR5;    *errorcodeptr = ERR5;
# Line 1312  if (*p == CHAR_RIGHT_CURLY_BRACKET) max Line 1331  if (*p == CHAR_RIGHT_CURLY_BRACKET) max
1331    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
1332      {      {
1333      max = 0;      max = 0;
1334      while(IS_DIGIT(*p)) max = max * 10 + *p++ - CHAR_0;      while(IS_DIGIT(*p)) max = max * 10 + (int)(*p++ - CHAR_0);
1335      if (max < 0 || max > 65535)      if (max < 0 || max > 65535)
1336        {        {
1337        *errorcodeptr = ERR5;        *errorcodeptr = ERR5;
# Line 1367  Arguments: Line 1386  Arguments:
1386    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1387    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1388    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1389    utf          TRUE if we are in UTF-8 / UTF-16 mode    utf          TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
1390    count        pointer to the current capturing subpattern number (updated)    count        pointer to the current capturing subpattern number (updated)
1391    
1392  Returns:       the number of the named subpattern, or -1 if not found  Returns:       the number of the named subpattern, or -1 if not found
# Line 1443  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1462  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1462      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&
1463          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)
1464        {        {
1465        int term;        pcre_uchar term;
1466        const pcre_uchar *thisname;        const pcre_uchar *thisname;
1467        *count += 1;        *count += 1;
1468        if (name == NULL && *count == lorn) return *count;        if (name == NULL && *count == lorn) return *count;
# Line 1451  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1470  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1470        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;
1471        thisname = ptr;        thisname = ptr;
1472        while (*ptr != term) ptr++;        while (*ptr != term) ptr++;
1473        if (name != NULL && lorn == ptr - thisname &&        if (name != NULL && lorn == (int)(ptr - thisname) &&
1474            STRNCMP_UC_UC(name, thisname, lorn) == 0)            STRNCMP_UC_UC(name, thisname, (unsigned int)lorn) == 0)
1475          return *count;          return *count;
1476        term++;        term++;
1477        }        }
# Line 1601  Arguments: Line 1620  Arguments:
1620    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1621    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1622    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1623    utf          TRUE if we are in UTF-8 / UTF-16 mode    utf          TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
1624    
1625  Returns:       the number of the found subpattern, or -1 if not found  Returns:       the number of the found subpattern, or -1 if not found
1626  */  */
# Line 1704  and doing the check at the end; a flag s Line 1723  and doing the check at the end; a flag s
1723    
1724  Arguments:  Arguments:
1725    code     points to the start of the pattern (the bracket)    code     points to the start of the pattern (the bracket)
1726    utf      TRUE in UTF-8 / UTF-16 mode    utf      TRUE in UTF-8 / UTF-16 / UTF-32 mode
1727    atend    TRUE if called when the pattern is complete    atend    TRUE if called when the pattern is complete
1728    cd       the "compile data" structure    cd       the "compile data" structure
1729    
# Line 1730  for (;;) Line 1749  for (;;)
1749    {    {
1750    int d;    int d;
1751    pcre_uchar *ce, *cs;    pcre_uchar *ce, *cs;
1752    register int op = *cc;    register pcre_uchar op = *cc;
1753    
1754    switch (op)    switch (op)
1755      {      {
# Line 1850  for (;;) Line 1869  for (;;)
1869      case OP_EXACTI:      case OP_EXACTI:
1870      case OP_NOTEXACT:      case OP_NOTEXACT:
1871      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1872      branchlength += GET2(cc,1);      branchlength += (int)GET2(cc,1);
1873      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1874  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
1875      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
# Line 1859  for (;;) Line 1878  for (;;)
1878    
1879      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1880      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1881      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP) cc += 2;      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP)
1882          cc += 2;
1883      cc += 1 + IMM2_SIZE + 1;      cc += 1 + IMM2_SIZE + 1;
1884      break;      break;
1885    
# Line 1894  for (;;) Line 1914  for (;;)
1914    
1915      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1916    
1917  #if defined SUPPORT_UTF || defined COMPILE_PCRE16  #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32
1918      case OP_XCLASS:      case OP_XCLASS:
1919      cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];      cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];
1920      /* Fall through */      /* Fall through */
# Line 1917  for (;;) Line 1937  for (;;)
1937        case OP_CRRANGE:        case OP_CRRANGE:
1938        case OP_CRMINRANGE:        case OP_CRMINRANGE:
1939        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;
1940        branchlength += GET2(cc,1);        branchlength += (int)GET2(cc,1);
1941        cc += 1 + 2 * IMM2_SIZE;        cc += 1 + 2 * IMM2_SIZE;
1942        break;        break;
1943    
# Line 2033  length. Line 2053  length.
2053    
2054  Arguments:  Arguments:
2055    code        points to start of expression    code        points to start of expression
2056    utf         TRUE in UTF-8 / UTF-16 mode    utf         TRUE in UTF-8 / UTF-16 / UTF-32 mode
2057    number      the required bracket number or negative to find a lookbehind    number      the required bracket number or negative to find a lookbehind
2058    
2059  Returns:      pointer to the opcode for the bracket, or NULL if not found  Returns:      pointer to the opcode for the bracket, or NULL if not found
# Line 2044  PRIV(find_bracket)(const pcre_uchar *cod Line 2064  PRIV(find_bracket)(const pcre_uchar *cod
2064  {  {
2065  for (;;)  for (;;)
2066    {    {
2067    register int c = *code;    register pcre_uchar c = *code;
2068    
2069    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2070    
# Line 2067  for (;;) Line 2087  for (;;)
2087    else if (c == OP_CBRA || c == OP_SCBRA ||    else if (c == OP_CBRA || c == OP_SCBRA ||
2088             c == OP_CBRAPOS || c == OP_SCBRAPOS)             c == OP_CBRAPOS || c == OP_SCBRAPOS)
2089      {      {
2090      int n = GET2(code, 1+LINK_SIZE);      int n = (int)GET2(code, 1+LINK_SIZE);
2091      if (n == number) return (pcre_uchar *)code;      if (n == number) return (pcre_uchar *)code;
2092      code += PRIV(OP_lengths)[c];      code += PRIV(OP_lengths)[c];
2093      }      }
# Line 2097  for (;;) Line 2117  for (;;)
2117        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2118        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2119        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
2120        if (code[1 + IMM2_SIZE] == OP_PROP        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2121          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;          code += 2;
2122        break;        break;
2123    
2124        case OP_MARK:        case OP_MARK:
# Line 2120  for (;;) Line 2140  for (;;)
2140    a multi-byte character. The length in the table is a minimum, so we have to    a multi-byte character. The length in the table is a minimum, so we have to
2141    arrange to skip the extra bytes. */    arrange to skip the extra bytes. */
2142    
2143  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2144      if (utf) switch(c)      if (utf) switch(c)
2145        {        {
2146        case OP_CHAR:        case OP_CHAR:
# Line 2172  instance of OP_RECURSE. Line 2192  instance of OP_RECURSE.
2192    
2193  Arguments:  Arguments:
2194    code        points to start of expression    code        points to start of expression
2195    utf         TRUE in UTF-8 / UTF-16 mode    utf         TRUE in UTF-8 / UTF-16 / UTF-32 mode
2196    
2197  Returns:      pointer to the opcode for OP_RECURSE, or NULL if not found  Returns:      pointer to the opcode for OP_RECURSE, or NULL if not found
2198  */  */
# Line 2182  find_recurse(const pcre_uchar *code, BOO Line 2202  find_recurse(const pcre_uchar *code, BOO
2202  {  {
2203  for (;;)  for (;;)
2204    {    {
2205    register int c = *code;    register pcre_uchar c = *code;
2206    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2207    if (c == OP_RECURSE) return code;    if (c == OP_RECURSE) return code;
2208    
# Line 2217  for (;;) Line 2237  for (;;)
2237        case OP_TYPEUPTO:        case OP_TYPEUPTO:
2238        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2239        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2240        if (code[1 + IMM2_SIZE] == OP_PROP        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2241          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;          code += 2;
2242        break;        break;
2243    
2244        case OP_MARK:        case OP_MARK:
# Line 2240  for (;;) Line 2260  for (;;)
2260      by a multi-byte character. The length in the table is a minimum, so we have      by a multi-byte character. The length in the table is a minimum, so we have
2261      to arrange to skip the extra bytes. */      to arrange to skip the extra bytes. */
2262    
2263  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2264      if (utf) switch(c)      if (utf) switch(c)
2265        {        {
2266        case OP_CHAR:        case OP_CHAR:
# Line 2326  bracket whose current branch will alread Line 2346  bracket whose current branch will alread
2346  Arguments:  Arguments:
2347    code        points to start of search    code        points to start of search
2348    endcode     points to where to stop    endcode     points to where to stop
2349    utf         TRUE if in UTF-8 / UTF-16 mode    utf         TRUE if in UTF-8 / UTF-16 / UTF-32 mode
2350    cd          contains pointers to tables etc.    cd          contains pointers to tables etc.
2351    
2352  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2336  static BOOL Line 2356  static BOOL
2356  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,
2357    BOOL utf, compile_data *cd)    BOOL utf, compile_data *cd)
2358  {  {
2359  register int c;  register pcre_uchar c;
2360  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
2361       code < endcode;       code < endcode;
2362       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))
# Line 2370  for (code = first_significant_code(code Line 2390  for (code = first_significant_code(code
2390      /* Test for forward reference */      /* Test for forward reference */
2391    
2392      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)
2393        if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE;        if ((int)GET(scode, 0) == (int)(code + 1 - cd->start_code)) return TRUE;
2394    
2395      /* Not a forward reference, test for completed backward reference */      /* Not a forward reference, test for completed backward reference */
2396    
# Line 2543  for (code = first_significant_code(code Line 2563  for (code = first_significant_code(code
2563      case OP_TYPEUPTO:      case OP_TYPEUPTO:
2564      case OP_TYPEMINUPTO:      case OP_TYPEMINUPTO:
2565      case OP_TYPEPOSUPTO:      case OP_TYPEPOSUPTO:
2566      if (code[1 + IMM2_SIZE] == OP_PROP      if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2567        || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;        code += 2;
2568      break;      break;
2569    
2570      /* End of branch */      /* End of branch */
# Line 2559  for (code = first_significant_code(code Line 2579  for (code = first_significant_code(code
2579      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
2580      MINUPTO, and POSUPTO may be followed by a multibyte character */      MINUPTO, and POSUPTO may be followed by a multibyte character */
2581    
2582  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2583      case OP_STAR:      case OP_STAR:
2584      case OP_STARI:      case OP_STARI:
2585      case OP_MINSTAR:      case OP_MINSTAR:
# Line 2625  Arguments: Line 2645  Arguments:
2645    code        points to start of the recursion    code        points to start of the recursion
2646    endcode     points to where to stop (current RECURSE item)    endcode     points to where to stop (current RECURSE item)
2647    bcptr       points to the chain of current (unclosed) branch starts    bcptr       points to the chain of current (unclosed) branch starts
2648    utf         TRUE if in UTF-8 / UTF-16 mode    utf         TRUE if in UTF-8 / UTF-16 / UTF-32 mode
2649    cd          pointers to tables etc    cd          pointers to tables etc
2650    
2651  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2691  Returns:   TRUE or FALSE Line 2711  Returns:   TRUE or FALSE
2711  static BOOL  static BOOL
2712  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)
2713  {  {
2714  int terminator;          /* Don't combine these lines; the Solaris cc */  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */
2715  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
2716  for (++ptr; *ptr != 0; ptr++)  for (++ptr; *ptr != 0; ptr++)
2717    {    {
# Line 2740  register int yield = 0; Line 2760  register int yield = 0;
2760  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
2761    {    {
2762    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
2763      STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;      STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield;
2764    pn += posix_name_lengths[yield] + 1;    pn += posix_name_lengths[yield] + 1;
2765    yield++;    yield++;
2766    }    }
# Line 2772  value in the reference (which is a group Line 2792  value in the reference (which is a group
2792  Arguments:  Arguments:
2793    group      points to the start of the group    group      points to the start of the group
2794    adjust     the amount by which the group is to be moved    adjust     the amount by which the group is to be moved
2795    utf        TRUE in UTF-8 / UTF-16 mode    utf        TRUE in UTF-8 / UTF-16 / UTF-32 mode
2796    cd         contains pointers to tables etc.    cd         contains pointers to tables etc.
2797    save_hwm   the hwm forward reference pointer at the start of the group    save_hwm   the hwm forward reference pointer at the start of the group
2798    
# Line 2795  while ((ptr = (pcre_uchar *)find_recurse Line 2815  while ((ptr = (pcre_uchar *)find_recurse
2815    
2816    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
2817      {      {
2818      offset = GET(hc, 0);      offset = (int)GET(hc, 0);
2819      if (cd->start_code + offset == ptr + 1)      if (cd->start_code + offset == ptr + 1)
2820        {        {
2821        PUT(hc, 0, offset + adjust);        PUT(hc, 0, offset + adjust);
# Line 2808  while ((ptr = (pcre_uchar *)find_recurse Line 2828  while ((ptr = (pcre_uchar *)find_recurse
2828    
2829    if (hc >= cd->hwm)    if (hc >= cd->hwm)
2830      {      {
2831      offset = GET(ptr, 1);      offset = (int)GET(ptr, 1);
2832      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
2833      }      }
2834    
# Line 2894  Yield:        -1 when no more Line 2914  Yield:        -1 when no more
2914  */  */
2915    
2916  static int  static int
2917  get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr,  get_othercase_range(pcre_uint32 *cptr, pcre_uint32 d, pcre_uint32 *ocptr,
2918    unsigned int *odptr)    pcre_uint32 *odptr)
2919  {  {
2920  unsigned int c, othercase, next;  pcre_uint32 c, othercase, next;
2921  int co;  unsigned int co;
2922    
2923  /* 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
2924  cases, return its case offset value. */  cases, return its case offset value. */
# Line 2909  for (c = *cptr; c <= d; c++) Line 2929  for (c = *cptr; c <= d; c++)
2929      {      {
2930      *ocptr = c++;   /* Character that has the set */      *ocptr = c++;   /* Character that has the set */
2931      *cptr = c;      /* Rest of input range */      *cptr = c;      /* Rest of input range */
2932      return co;      return (int)co;
2933      }      }
2934    if ((othercase = UCD_OTHERCASE(c)) != c) break;    if ((othercase = UCD_OTHERCASE(c)) != c) break;
2935    }    }
# Line 2949  Returns:       TRUE if auto-possessifyin Line 2969  Returns:       TRUE if auto-possessifyin
2969  */  */
2970    
2971  static BOOL  static BOOL
2972  check_char_prop(int c, int ptype, int pdata, BOOL negated)  check_char_prop(pcre_uint32 c, unsigned int ptype, unsigned int pdata, BOOL negated)
2973  {  {
2974    #ifdef SUPPORT_UCP
2975    const pcre_uint32 *p;
2976    #endif
2977    
2978  const ucd_record *prop = GET_UCD(c);  const ucd_record *prop = GET_UCD(c);
2979    
2980  switch(ptype)  switch(ptype)
2981    {    {
2982    case PT_LAMP:    case PT_LAMP:
# Line 2989  switch(ptype) Line 3014  switch(ptype)
3014    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
3015            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
3016            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
3017    
3018    #ifdef SUPPORT_UCP
3019      case PT_CLIST:
3020      p = PRIV(ucd_caseless_sets) + prop->caseset;
3021      for (;;)
3022        {
3023        if ((unsigned int)c < *p) return !negated;
3024        if ((unsigned int)c == *p++) return negated;
3025        }
3026      break;  /* Control never reaches here */
3027    #endif
3028    }    }
3029    
3030  return FALSE;  return FALSE;
3031  }  }
3032  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
# Line 3006  sense to automatically possessify the re Line 3043  sense to automatically possessify the re
3043    
3044  Arguments:  Arguments:
3045    previous      pointer to the repeated opcode    previous      pointer to the repeated opcode
3046    utf           TRUE in UTF-8 / UTF-16 mode    utf           TRUE in UTF-8 / UTF-16 / UTF-32 mode
3047    ptr           next character in pattern    ptr           next character in pattern
3048    options       options bits    options       options bits
3049    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
# Line 3018  static BOOL Line 3055  static BOOL
3055  check_auto_possessive(const pcre_uchar *previous, BOOL utf,  check_auto_possessive(const pcre_uchar *previous, BOOL utf,
3056    const pcre_uchar *ptr, int options, compile_data *cd)    const pcre_uchar *ptr, int options, compile_data *cd)
3057  {  {
3058  pcre_int32 c, next;  pcre_uint32 c = NOTACHAR;
3059  int op_code = *previous++;  pcre_uint32 next;
3060    int escape;
3061    pcre_uchar op_code = *previous++;
3062    
3063  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
3064    
# Line 3050  value is a character, a negative value i Line 3089  value is a character, a negative value i
3089  if (*ptr == CHAR_BACKSLASH)  if (*ptr == CHAR_BACKSLASH)
3090    {    {
3091    int temperrorcode = 0;    int temperrorcode = 0;
3092    next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE);    escape = check_escape(&ptr, &next, &temperrorcode, cd->bracount, options, FALSE);
3093    if (temperrorcode != 0) return FALSE;    if (temperrorcode != 0) return FALSE;
3094    ptr++;    /* Point after the escape sequence */    ptr++;    /* Point after the escape sequence */
3095    }    }
3096  else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)  else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)
3097    {    {
3098      escape = 0;
3099  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3100    if (utf) { GETCHARINC(next, ptr); } else    if (utf) { GETCHARINC(next, ptr); } else
3101  #endif  #endif
# Line 3092  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3132  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3132    STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)    STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
3133      return FALSE;      return FALSE;
3134    
3135  /* Now compare the next item with the previous opcode. First, handle cases when  /* If the previous item is a character, get its value. */
 the next item is a character. */  
3136    
3137  if (next >= 0) switch(op_code)  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3138        op_code == OP_NOT || op_code == OP_NOTI)
3139    {    {
   case OP_CHAR:  
3140  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3141    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
3142  #else  #else
3143    c = *previous;    c = *previous;
3144  #endif  #endif
3145    return c != next;    }
3146    
3147    /* For CHARI (caseless character) we must check the other case. If we have  /* Now compare the next item with the previous opcode. First, handle cases when
3148    Unicode property support, we can use it to test the other case of  the next item is a character. */
   high-valued characters. */  
3149    
3150    case OP_CHARI:  if (escape == 0)
3151  #ifdef SUPPORT_UTF    {
3152    GETCHARTEST(c, previous);    /* For a caseless UTF match, the next character may have more than one other
3153  #else    case, which maps to the special PT_CLIST property. Check this first. */
3154    c = *previous;  
 #endif  
   if (c == next) return FALSE;  
 #ifdef SUPPORT_UTF  
   if (utf)  
     {  
     unsigned int othercase;  
     if (next < 128) othercase = cd->fcc[next]; else  
3155  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3156      othercase = UCD_OTHERCASE((unsigned int)next);    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3157  #else      {
3158      othercase = NOTACHAR;      unsigned int ocs = UCD_CASESET(next);
3159  #endif      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);
     return (unsigned int)c != othercase;  
3160      }      }
   else  
 #endif  /* SUPPORT_UTF */  
   return (c != TABLE_GET((unsigned int)next, cd->fcc, next));  /* Non-UTF-8 mode */  
   
   case OP_NOT:  
 #ifdef SUPPORT_UTF  
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
3161  #endif  #endif
   return c == next;  
3162    
3163    case OP_NOTI:    switch(op_code)
3164        {
3165        case OP_CHAR:
3166        return c != next;
3167    
3168        /* For CHARI (caseless character) we must check the other case. If we have
3169        Unicode property support, we can use it to test the other case of
3170        high-valued characters. We know that next can have only one other case,
3171        because multi-other-case characters are dealt with above. */
3172    
3173        case OP_CHARI:
3174        if (c == next) return FALSE;
3175  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3176    GETCHARTEST(c, previous);      if (utf)
3177          {
3178          pcre_uint32 othercase;
3179          if (next < 128) othercase = cd->fcc[next]; else
3180    #ifdef SUPPORT_UCP
3181          othercase = UCD_OTHERCASE(next);
3182  #else  #else
3183    c = *previous;        othercase = NOTACHAR;
3184  #endif  #endif
3185    if (c == next) return TRUE;        return c != othercase;
3186          }
3187        else
3188    #endif  /* SUPPORT_UTF */
3189        return (c != TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3190    
3191        case OP_NOT:
3192        return c == next;
3193    
3194        case OP_NOTI:
3195        if (c == next) return TRUE;
3196  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3197    if (utf)      if (utf)
3198      {        {
3199      unsigned int othercase;        pcre_uint32 othercase;
3200      if (next < 128) othercase = cd->fcc[next]; else        if (next < 128) othercase = cd->fcc[next]; else
3201  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3202      othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE(next);
3203  #else  #else
3204      othercase = NOTACHAR;        othercase = NOTACHAR;
3205  #endif  #endif
3206      return (unsigned int)c == othercase;        return c == othercase;
3207      }        }
3208    else      else
3209  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3210    return (c == TABLE_GET((unsigned int)next, cd->fcc, next));  /* Non-UTF-8 mode */      return (c == TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3211    
3212    /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.      /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
3213    When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */      When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
3214    
3215    case OP_DIGIT:      case OP_DIGIT:
3216    return next > 255 || (cd->ctypes[next] & ctype_digit) == 0;      return next > 255 || (cd->ctypes[next] & ctype_digit) == 0;
3217    
3218    case OP_NOT_DIGIT:      case OP_NOT_DIGIT:
3219    return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;
3220    
3221    case OP_WHITESPACE:      case OP_WHITESPACE:
3222    return next > 255 || (cd->ctypes[next] & ctype_space) == 0;      return next > 255 || (cd->ctypes[next] & ctype_space) == 0;
3223    
3224    case OP_NOT_WHITESPACE:      case OP_NOT_WHITESPACE:
3225    return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;
3226    
3227    case OP_WORDCHAR:      case OP_WORDCHAR:
3228    return next > 255 || (cd->ctypes[next] & ctype_word) == 0;      return next > 255 || (cd->ctypes[next] & ctype_word) == 0;
3229    
3230    case OP_NOT_WORDCHAR:      case OP_NOT_WORDCHAR:
3231    return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;
3232    
3233    case OP_HSPACE:      case OP_HSPACE:
3234    case OP_NOT_HSPACE:      case OP_NOT_HSPACE:
3235    switch(next)      switch(next)
3236      {        {
3237      HSPACE_CASES:        HSPACE_CASES:
3238      return op_code == OP_NOT_HSPACE;        return op_code == OP_NOT_HSPACE;
   
     default:  
     return op_code != OP_NOT_HSPACE;  
     }  
3239    
3240    case OP_ANYNL:        default:
3241    case OP_VSPACE:        return op_code != OP_NOT_HSPACE;
3242    case OP_NOT_VSPACE:        }
3243    switch(next)  
3244      {      case OP_ANYNL:
3245      VSPACE_CASES:      case OP_VSPACE:
3246      return op_code == OP_NOT_VSPACE;      case OP_NOT_VSPACE:
3247        switch(next)
3248      default:        {
3249      return op_code != OP_NOT_VSPACE;        VSPACE_CASES:
3250      }        return op_code == OP_NOT_VSPACE;
3251    
3252          default:
3253          return op_code != OP_NOT_VSPACE;
3254          }
3255    
3256  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3257    case OP_PROP:      case OP_PROP:
3258    return check_char_prop(next, previous[0], previous[1], FALSE);      return check_char_prop(next, previous[0], previous[1], FALSE);
3259    
3260    case OP_NOTPROP:      case OP_NOTPROP:
3261    return check_char_prop(next, previous[0], previous[1], TRUE);      return check_char_prop(next, previous[0], previous[1], TRUE);
3262  #endif  #endif
3263    
3264    default:      default:
3265    return FALSE;      return FALSE;
3266        }
3267    }    }
3268    
   
3269  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
3270  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
3271  generated only when PCRE_UCP is *not* set, that is, when only ASCII  generated only when PCRE_UCP is *not* set, that is, when only ASCII
# Line 3230  switch(op_code) Line 3276  switch(op_code)
3276    {    {
3277    case OP_CHAR:    case OP_CHAR:
3278    case OP_CHARI:    case OP_CHARI:
3279  #ifdef SUPPORT_UTF    switch(escape)
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
 #endif  
   switch(-next)  
3280      {      {
3281      case ESC_d:      case ESC_d:
3282      return c > 255 || (cd->ctypes[c] & ctype_digit) == 0;      return c > 255 || (cd->ctypes[c] & ctype_digit) == 0;
# Line 3260  switch(op_code) Line 3301  switch(op_code)
3301      switch(c)      switch(c)
3302        {        {
3303        HSPACE_CASES:        HSPACE_CASES:
3304        return -next != ESC_h;        return escape != ESC_h;
3305    
3306        default:        default:
3307        return -next == ESC_h;        return escape == ESC_h;
3308        }        }
3309    
3310      case ESC_v:      case ESC_v:
# Line 3271  switch(op_code) Line 3312  switch(op_code)
3312      switch(c)      switch(c)
3313        {        {
3314        VSPACE_CASES:        VSPACE_CASES:
3315        return -next != ESC_v;        return escape != ESC_v;
3316    
3317        default:        default:
3318        return -next == ESC_v;        return escape == ESC_v;
3319        }        }
3320    
3321      /* When PCRE_UCP is set, these values get generated for \d etc. Find      /* When PCRE_UCP is set, these values get generated for \d etc. Find
3322      their substitutions and process them. The result will always be either      their substitutions and process them. The result will always be either
3323      -ESC_p or -ESC_P. Then fall through to process those values. */      ESC_p or ESC_P. Then fall through to process those values. */
3324    
3325  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3326      case ESC_du:      case ESC_du:
# Line 3290  switch(op_code) Line 3331  switch(op_code)
3331      case ESC_SU:      case ESC_SU:
3332        {        {
3333        int temperrorcode = 0;        int temperrorcode = 0;
3334        ptr = substitutes[-next - ESC_DU];        ptr = substitutes[escape - ESC_DU];
3335        next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);        escape = check_escape(&ptr, &next, &temperrorcode, 0, options, FALSE);
3336        if (temperrorcode != 0) return FALSE;        if (temperrorcode != 0) return FALSE;
3337        ptr++;    /* For compatibility */        ptr++;    /* For compatibility */
3338        }        }
# Line 3300  switch(op_code) Line 3341  switch(op_code)
3341      case ESC_p:      case ESC_p:
3342      case ESC_P:      case ESC_P:
3343        {        {
3344        int ptype, pdata, errorcodeptr;        unsigned int ptype = 0, pdata = 0;
3345          int errorcodeptr;
3346        BOOL negated;        BOOL negated;
3347    
3348        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
3349        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
3350        if (ptype < 0) return FALSE;          return FALSE;
3351        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
3352    
3353        /* If the property item is optional, we have to give up. (When generated        /* If the property item is optional, we have to give up. (When generated
# Line 3318  switch(op_code) Line 3360  switch(op_code)
3360    
3361        /* Do the property check. */        /* Do the property check. */
3362    
3363        return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);        return check_char_prop(c, ptype, pdata, (escape == ESC_P) != negated);
3364        }        }
3365  #endif  #endif
3366    
# Line 3333  switch(op_code) Line 3375  switch(op_code)
3375    these op-codes are never generated.) */    these op-codes are never generated.) */
3376    
3377    case OP_DIGIT:    case OP_DIGIT:
3378    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||    return escape == ESC_D || escape == ESC_s || escape == ESC_W ||
3379           next == -ESC_h || next == -ESC_v || next == -ESC_R;           escape == ESC_h || escape == ESC_v || escape == ESC_R;
3380    
3381    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
3382    return next == -ESC_d;    return escape == ESC_d;
3383    
3384    case OP_WHITESPACE:    case OP_WHITESPACE:
3385    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return escape == ESC_S || escape == ESC_d || escape == ESC_w;
3386    
3387    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
3388    return next == -ESC_s || next == -ESC_h || next == -ESC_v || next == -ESC_R;    return escape == ESC_s || escape == ESC_h || escape == ESC_v || escape == ESC_R;
3389    
3390    case OP_HSPACE:    case OP_HSPACE:
3391    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||    return escape == ESC_S || escape == ESC_H || escape == ESC_d ||
3392           next == -ESC_w || next == -ESC_v || next == -ESC_R;           escape == ESC_w || escape == ESC_v || escape == ESC_R;
3393    
3394    case OP_NOT_HSPACE:    case OP_NOT_HSPACE:
3395    return next == -ESC_h;    return escape == ESC_h;
3396    
3397    /* Can't have \S in here because VT matches \S (Perl anomaly) */    /* Can't have \S in here because VT matches \S (Perl anomaly) */
3398    case OP_ANYNL:    case OP_ANYNL:
3399    case OP_VSPACE:    case OP_VSPACE:
3400    return next == -ESC_V || next == -ESC_d || next == -ESC_w;    return escape == ESC_V || escape == ESC_d || escape == ESC_w;
3401    
3402    case OP_NOT_VSPACE:    case OP_NOT_VSPACE:
3403    return next == -ESC_v || next == -ESC_R;    return escape == ESC_v || escape == ESC_R;
3404    
3405    case OP_WORDCHAR:    case OP_WORDCHAR:
3406    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||    return escape == ESC_W || escape == ESC_s || escape == ESC_h ||
3407           next == -ESC_v || next == -ESC_R;           escape == ESC_v || escape == ESC_R;
3408    
3409    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
3410    return next == -ESC_w || next == -ESC_d;    return escape == ESC_w || escape == ESC_d;
3411    
3412    default:    default:
3413    return FALSE;    return FALSE;
# Line 3399  Returns:        the number of < 256 char Line 3441  Returns:        the number of < 256 char
3441    
3442  static int  static int
3443  add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,  add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,
3444    compile_data *cd, unsigned int start, unsigned int end)    compile_data *cd, pcre_uint32 start, pcre_uint32 end)
3445  {  {
3446  unsigned int c;  pcre_uint32 c;
3447  int n8 = 0;  int n8 = 0;
3448    
3449  /* If caseless matching is required, scan the range and process alternate  /* If caseless matching is required, scan the range and process alternate
# Line 3415  if ((options & PCRE_CASELESS) != 0) Line 3457  if ((options & PCRE_CASELESS) != 0)
3457    if ((options & PCRE_UTF8) != 0)    if ((options & PCRE_UTF8) != 0)
3458      {      {
3459      int rc;      int rc;
3460      unsigned int oc, od;      pcre_uint32 oc, od;
3461    
3462      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */
3463      c = start;      c = start;
# Line 3456  if ((options & PCRE_CASELESS) != 0) Line 3498  if ((options & PCRE_CASELESS) != 0)
3498  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
3499  in all cases. */  in all cases. */
3500    
3501  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
3502  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3503    if ((options & PCRE_UTF8) == 0)    if ((options & PCRE_UTF8) == 0)
3504  #endif  #endif
3505    if (end > 0xff) end = 0xff;    if (end > 0xff) end = 0xff;
 #endif  
3506    
3507  #ifdef COMPILE_PCRE16  #elif defined COMPILE_PCRE16
3508  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3509    if ((options & PCRE_UTF16) == 0)    if ((options & PCRE_UTF16) == 0)
3510  #endif  #endif
3511    if (end > 0xffff) end = 0xffff;    if (end > 0xffff) end = 0xffff;
3512  #endif  
3513    #endif /* COMPILE_PCRE[8|16] */
3514    
3515  /* 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
3516  data. */  data. */
# Line 3598  static int Line 3640  static int
3640  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,
3641    int options, compile_data *cd, const pcre_uint32 *p)    int options, compile_data *cd, const pcre_uint32 *p)
3642  {  {
3643    BOOL utf = (options & PCRE_UTF8) != 0;
3644  int n8 = 0;  int n8 = 0;
3645  if (p[0] > 0)  if (p[0] > 0)
3646    n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);    n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);
3647  while (p[0] < NOTACHAR)  while (p[0] < NOTACHAR)
3648    {    {
3649    while (p[1] == p[0] + 1) p++;    while (p[1] == p[0] + 1) p++;
3650    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,
3651      (p[1] == NOTACHAR)? 0x10ffff : p[1] - 1);      (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);
3652    p++;    p++;
3653    }    }
3654  return n8;  return n8;
# Line 3628  Arguments: Line 3671  Arguments:
3671    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3672    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3673    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3674    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr    place to put the first required character
3675    reqcharptr     set to the last literal character required, else < 0    firstcharflagsptr place to put the first character flags, or a negative number
3676      reqcharptr     place to put the last required character
3677      reqcharflagsptr place to put the last required character flags, or a negative number
3678    bcptr          points to current branch chain    bcptr          points to current branch chain
3679    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3680    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3642  Returns:         TRUE on success Line 3687  Returns:         TRUE on success
3687    
3688  static BOOL  static BOOL
3689  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3690    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,    const pcre_uchar **ptrptr, int *errorcodeptr,
3691    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,    pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
3692      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
3693      branch_chain *bcptr, int cond_depth,
3694    compile_data *cd, int *lengthptr)    compile_data *cd, int *lengthptr)
3695  {  {
3696  int repeat_type, op_type;  int repeat_type, op_type;
3697  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3698  int bravalue = 0;  int bravalue = 0;
3699  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3700  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
3701  pcre_int32 zeroreqchar, zerofirstchar;  pcre_int32 firstcharflags, reqcharflags;
3702    pcre_uint32 zeroreqchar, zerofirstchar;
3703    pcre_int32 zeroreqcharflags, zerofirstcharflags;
3704  pcre_int32 req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3705  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3706  int after_manual_callout = 0;  int after_manual_callout = 0;
3707  int length_prevgroup = 0;  int length_prevgroup = 0;
3708  register int c;  register pcre_uint32 c;
3709    int escape;
3710  register pcre_uchar *code = *codeptr;  register pcre_uchar *code = *codeptr;
3711  pcre_uchar *last_code = code;  pcre_uchar *last_code = code;
3712  pcre_uchar *orig_code = code;  pcre_uchar *orig_code = code;
# Line 3676  must not do this for other options (e.g. Line 3726  must not do this for other options (e.g.
3726  dynamically as we process the pattern. */  dynamically as we process the pattern. */
3727    
3728  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3729  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3730  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3731    #ifndef COMPILE_PCRE32
3732  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3733    #endif
3734  #else  #else
3735  BOOL utf = FALSE;  BOOL utf = FALSE;
3736  #endif  #endif
3737    
3738  /* Helper variables for OP_XCLASS opcode (for characters > 255). */  /* Helper variables for OP_XCLASS opcode (for characters > 255). We define
3739    class_uchardata always so that it can be passed to add_to_class() always,
3740    though it will not be used in non-UTF 8-bit cases. This avoids having to supply
3741    alternative calls for the different cases. */
3742    
3743    pcre_uchar *class_uchardata;
3744  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3745  BOOL xclass;  BOOL xclass;
 pcre_uchar *class_uchardata;  
3746  pcre_uchar *class_uchardata_base;  pcre_uchar *class_uchardata_base;
3747  #endif  #endif
3748    
# Line 3710  to take the zero repeat into account. Th Line 3765  to take the zero repeat into account. Th
3765  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3766  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3767    
3768  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
3769    firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;
3770    
3771  /* The variable req_caseopt contains either the REQ_CASELESS value  /* The variable req_caseopt contains either the REQ_CASELESS value
3772  or zero, according to the current setting of the caseless flag. The  or zero, according to the current setting of the caseless flag. The
# Line 3736  for (;; ptr++) Line 3792  for (;; ptr++)
3792    int recno;    int recno;
3793    int refsign;    int refsign;
3794    int skipbytes;    int skipbytes;
3795    int subreqchar;    pcre_uint32 subreqchar, subfirstchar;
3796    int subfirstchar;    pcre_int32 subreqcharflags, subfirstcharflags;
3797    int terminator;    int terminator;
3798    int mclength;    unsigned int mclength;
3799    int tempbracount;    int tempbracount;
3800      pcre_uint32 ec;
3801    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3802    
3803    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 3903  for (;; ptr++) Line 3960  for (;; ptr++)
3960      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3961      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3962      *firstcharptr = firstchar;      *firstcharptr = firstchar;
3963        *firstcharflagsptr = firstcharflags;
3964      *reqcharptr = reqchar;      *reqcharptr = reqchar;
3965        *reqcharflagsptr = reqcharflags;
3966      *codeptr = code;      *codeptr = code;
3967      *ptrptr = ptr;      *ptrptr = ptr;
3968      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3927  for (;; ptr++) Line 3986  for (;; ptr++)
3986      previous = NULL;      previous = NULL;
3987      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3988        {        {
3989        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
3990        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3991        }        }
3992      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3942  for (;; ptr++) Line 4001  for (;; ptr++)
4001      repeats. The value of reqchar doesn't change either. */      repeats. The value of reqchar doesn't change either. */
4002    
4003      case CHAR_DOT:      case CHAR_DOT:
4004      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4005      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4006        zerofirstcharflags = firstcharflags;
4007      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4008        zeroreqcharflags = reqcharflags;
4009      previous = code;      previous = code;
4010      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
4011      break;      break;
# Line 4018  for (;; ptr++) Line 4079  for (;; ptr++)
4079          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
4080        {        {
4081        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
4082        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4083        zerofirstchar = firstchar;        zerofirstchar = firstchar;
4084          zerofirstcharflags = firstcharflags;
4085        break;        break;
4086        }        }
4087    
# Line 4133  for (;; ptr++) Line 4195  for (;; ptr++)
4195          alpha. This relies on the fact that the class table starts with          alpha. This relies on the fact that the class table starts with
4196          alpha, lower, upper as the first 3 entries. */          alpha, lower, upper as the first 3 entries. */
4197    
4198          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
4199            posix_class = 0;            posix_class = 0;
4200    
4201          /* When PCRE_UCP is set, some of the POSIX classes are converted to          /* When PCRE_UCP is set, some of the POSIX classes are converted to
# Line 4210  for (;; ptr++) Line 4272  for (;; ptr++)
4272    
4273        if (c == CHAR_BACKSLASH)        if (c == CHAR_BACKSLASH)
4274          {          {
4275          c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);          escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, TRUE);
4276    
4277          if (*errorcodeptr != 0) goto FAILED;          if (*errorcodeptr != 0) goto FAILED;
4278    
4279          if (-c == ESC_b) c = CHAR_BS;    /* \b is backspace in a class */          if (escape == 0)
4280          else if (-c == ESC_N)            /* \N is not supported in a class */            c = ec;
4281            else if (escape == ESC_b) c = CHAR_BS; /* \b is backspace in a class */
4282            else if (escape == ESC_N)            /* \N is not supported in a class */
4283            {            {
4284            *errorcodeptr = ERR71;            *errorcodeptr = ERR71;
4285            goto FAILED;            goto FAILED;
4286            }            }
4287          else if (-c == ESC_Q)            /* Handle start of quoted string */          else if (escape == ESC_Q)            /* Handle start of quoted string */
4288            {            {
4289            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
4290              {              {
# Line 4228  for (;; ptr++) Line 4293  for (;; ptr++)
4293            else inescq = TRUE;            else inescq = TRUE;
4294            continue;            continue;
4295            }            }
4296          else if (-c == ESC_E) continue;  /* Ignore orphan \E */          else if (escape == ESC_E) continue;  /* Ignore orphan \E */
4297    
4298          if (c < 0)          else
4299            {            {
4300            register const pcre_uint8 *cbits = cd->cbits;            register const pcre_uint8 *cbits = cd->cbits;
4301            /* Every class contains at least two < 256 characters. */            /* Every class contains at least two < 256 characters. */
# Line 4238  for (;; ptr++) Line 4303  for (;; ptr++)
4303            /* Every class contains at least two characters. */            /* Every class contains at least two characters. */
4304            class_one_char += 2;            class_one_char += 2;
4305    
4306            switch (-c)            switch (escape)
4307              {              {
4308  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4309              case ESC_du:     /* These are the values given for \d etc */              case ESC_du:     /* These are the values given for \d etc */
# Line 4248  for (;; ptr++) Line 4313  for (;; ptr++)
4313              case ESC_su:     /* of the default ASCII testing. */              case ESC_su:     /* of the default ASCII testing. */
4314              case ESC_SU:              case ESC_SU:
4315              nestptr = ptr;              nestptr = ptr;
4316              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */              ptr = substitutes[escape - ESC_DU] - 1;  /* Just before substitute */
4317              class_has_8bitchar--;                /* Undo! */              class_has_8bitchar--;                /* Undo! */
4318              continue;              continue;
4319  #endif  #endif
# Line 4314  for (;; ptr++) Line 4379  for (;; ptr++)
4379              case ESC_P:              case ESC_P:
4380                {                {
4381                BOOL negated;                BOOL negated;
4382                int pdata;                unsigned int ptype = 0, pdata = 0;
4383                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
4384                if (ptype < 0) goto FAILED;                  goto FAILED;
4385                *class_uchardata++ = ((-c == ESC_p) != negated)?                *class_uchardata++ = ((escape == ESC_p) != negated)?
4386                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4387                *class_uchardata++ = ptype;                *class_uchardata++ = ptype;
4388                *class_uchardata++ = pdata;                *class_uchardata++ = pdata;
# Line 4345  for (;; ptr++) Line 4410  for (;; ptr++)
4410          /* Fall through if the escape just defined a single character (c >= 0).          /* Fall through if the escape just defined a single character (c >= 0).
4411          This may be greater than 256. */          This may be greater than 256. */
4412    
4413            escape = 0;
4414    
4415          }   /* End of backslash handling */          }   /* End of backslash handling */
4416    
4417        /* A 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 4435  for (;; ptr++)
4435    
4436        if (!inescq && ptr[1] == CHAR_MINUS)        if (!inescq && ptr[1] == CHAR_MINUS)
4437          {          {
4438          int d;          pcre_uint32 d;
4439          ptr += 2;          ptr += 2;
4440          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;
4441    
# Line 4410  for (;; ptr++) Line 4477  for (;; ptr++)
4477    
4478          if (!inescq && d == CHAR_BACKSLASH)          if (!inescq && d == CHAR_BACKSLASH)
4479            {            {
4480            d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);            int descape;
4481              descape = check_escape(&ptr, &d, errorcodeptr, cd->bracount, options, TRUE);
4482            if (*errorcodeptr != 0) goto FAILED;            if (*errorcodeptr != 0) goto FAILED;
4483    
4484            /* \b is backspace; any other special means the '-' was literal. */            /* \b is backspace; any other special means the '-' was literal. */
4485    
4486            if (d < 0)            if (descape != 0)
4487              {              {
4488              if (d == -ESC_b) d = CHAR_BS; else              if (descape == ESC_b) d = CHAR_BS; else
4489                {                {
4490                ptr = oldptr;                ptr = oldptr;
4491                goto CLASS_SINGLE_CHARACTER;  /* A few lines below */                goto CLASS_SINGLE_CHARACTER;  /* A few lines below */
# Line 4473  for (;; ptr++) Line 4541  for (;; ptr++)
4541          {          {
4542          ptr++;          ptr++;
4543          zeroreqchar = reqchar;          zeroreqchar = reqchar;
4544            zeroreqcharflags = reqcharflags;
4545    
4546          if (negate_class)          if (negate_class)
4547            {            {
4548            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;  #ifdef SUPPORT_UCP
4549              int d;
4550    #endif
4551              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4552            zerofirstchar = firstchar;            zerofirstchar = firstchar;
4553            *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;            zerofirstcharflags = firstcharflags;
4554  #ifdef SUPPORT_UTF  
4555            if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)            /* For caseless UTF-8 mode when UCP support is available, check
4556              code += PRIV(ord2utf)(c, code);            whether this character has more than one other case. If so, generate
4557            else            a special OP_NOTPROP item instead of OP_NOTI. */
4558    
4559    #ifdef SUPPORT_UCP
4560              if (utf && (options & PCRE_CASELESS) != 0 &&
4561                  (d = UCD_CASESET(c)) != 0)
4562                {
4563                *code++ = OP_NOTPROP;
4564                *code++ = PT_CLIST;
4565                *code++ = d;
4566                }
4567              else
4568  #endif  #endif
4569              *code++ = c;            /* Char has only one other case, or UCP not available */
4570            goto NOT_CHAR;  
4571                {
4572                *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
4573    #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4574                if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
4575                  code += PRIV(ord2utf)(c, code);
4576                else
4577    #endif
4578                  *code++ = c;
4579                }
4580    
4581              /* We are finished with this character class */
4582    
4583              goto END_CLASS;
4584            }            }
4585    
4586          /* For a single, positive character, get the value into mcbuffer, and          /* For a single, positive character, get the value into mcbuffer, and
4587          then we can handle this with the normal one-character code. */          then we can handle this with the normal one-character code. */
4588    
4589  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4590          if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)          if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
4591            mclength = PRIV(ord2utf)(c, mcbuffer);            mclength = PRIV(ord2utf)(c, mcbuffer);
4592          else          else
# Line 4543  for (;; ptr++) Line 4638  for (;; ptr++)
4638      setting, whatever the repeat count. Any reqchar setting must remain      setting, whatever the repeat count. Any reqchar setting must remain
4639      unchanged after any kind of repeat. */      unchanged after any kind of repeat. */
4640    
4641      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4642      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4643        zerofirstcharflags = firstcharflags;
4644      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4645        zeroreqcharflags = reqcharflags;
4646    
4647      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4648      extended class, with its own opcode, unless there was a negated special      extended class, with its own opcode, unless there was a negated special
# Line 4601  for (;; ptr++) Line 4698  for (;; ptr++)
4698        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
4699        }        }
4700      code += 32 / sizeof(pcre_uchar);      code += 32 / sizeof(pcre_uchar);
4701      NOT_CHAR:  
4702        END_CLASS:
4703      break;      break;
4704    
4705    
# Line 4639  for (;; ptr++) Line 4737  for (;; ptr++)
4737      if (repeat_min == 0)      if (repeat_min == 0)
4738        {        {
4739        firstchar = zerofirstchar;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4740          firstcharflags = zerofirstcharflags;
4741        reqchar = zeroreqchar;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4742          reqcharflags = zeroreqcharflags;
4743        }        }
4744    
4745      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4725  for (;; ptr++) Line 4825  for (;; ptr++)
4825        hold the length of the character in bytes, plus UTF_LENGTH to flag that        hold the length of the character in bytes, plus UTF_LENGTH to flag that
4826        it's a length rather than a small character. */        it's a length rather than a small character. */
4827    
4828  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4829        if (utf && NOT_FIRSTCHAR(code[-1]))        if (utf && NOT_FIRSTCHAR(code[-1]))
4830          {          {
4831          pcre_uchar *lastchar = code - 1;          pcre_uchar *lastchar = code - 1;
# Line 4742  for (;; ptr++) Line 4842  for (;; ptr++)
4842          {          {
4843          c = code[-1];          c = code[-1];
4844          if (*previous <= OP_CHARI && repeat_min > 1)          if (*previous <= OP_CHARI && repeat_min > 1)
4845            reqchar = c | req_caseopt | cd->req_varyopt;            {
4846              reqchar = c;
4847              reqcharflags = req_caseopt | cd->req_varyopt;
4848              }
4849          }          }
4850    
4851        /* If the repetition is unlimited, it pays to see if the next thing on        /* If the repetition is unlimited, it pays to see if the next thing on
# Line 4861  for (;; ptr++) Line 4964  for (;; ptr++)
4964    
4965          if (repeat_max < 0)          if (repeat_max < 0)
4966            {            {
4967  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4968            if (utf && (c & UTF_LENGTH) != 0)            if (utf && (c & UTF_LENGTH) != 0)
4969              {              {
4970              memcpy(code, utf_chars, IN_UCHARS(c & 7));              memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 4886  for (;; ptr++) Line 4989  for (;; ptr++)
4989    
4990          else if (repeat_max != repeat_min)          else if (repeat_max != repeat_min)
4991            {            {
4992  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4993            if (utf && (c & UTF_LENGTH) != 0)            if (utf && (c & UTF_LENGTH) != 0)
4994              {              {
4995              memcpy(code, utf_chars, IN_UCHARS(c & 7));              memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 4916  for (;; ptr++) Line 5019  for (;; ptr++)
5019    
5020        /* The character or character type itself comes last in all cases. */        /* The character or character type itself comes last in all cases. */
5021    
5022  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
5023        if (utf && (c & UTF_LENGTH) != 0)        if (utf && (c & UTF_LENGTH) != 0)
5024          {          {
5025          memcpy(code, utf_chars, IN_UCHARS(c & 7));          memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 5124  for (;; ptr++) Line 5227  for (;; ptr++)
5227    
5228            else            else
5229              {              {
5230              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;              if (groupsetfirstchar && reqcharflags < 0)
5231                  {
5232                  reqchar = firstchar;
5233                  reqcharflags = firstcharflags;
5234                  }
5235    
5236              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5237                {                {
# Line 5501  for (;; ptr++) Line 5608  for (;; ptr++)
5608          arg = ++ptr;          arg = ++ptr;
5609          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5610          arglen = (int)(ptr - arg);          arglen = (int)(ptr - arg);
5611          if (arglen > (int)MAX_MARK)          if ((unsigned int)arglen > MAX_MARK)
5612            {            {
5613            *errorcodeptr = ERR75;            *errorcodeptr = ERR75;
5614            goto FAILED;            goto FAILED;
# Line 5544  for (;; ptr++) Line 5651  for (;; ptr++)
5651                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5652    
5653              /* Do not set firstchar after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5654              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
5655              }              }
5656    
5657            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5710  for (;; ptr++) Line 5817  for (;; ptr++)
5817          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)
5818            {            {
5819            if (recno >= 0)            if (recno >= 0)
5820              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;              recno = (IS_DIGIT(*ptr))? recno * 10 + (int)(*ptr - CHAR_0) : -1;
5821            ptr++;            ptr++;
5822            }            }
5823          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5824    
5825          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) ||
5826              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
5827            {            {
5828            ptr--;      /* Error offset */            ptr--;      /* Error offset */
# Line 5955  for (;; ptr++) Line 6062  for (;; ptr++)
6062    
6063            if (lengthptr != NULL)            if (lengthptr != NULL)
6064              {              {
6065              if (*ptr != terminator)              if (*ptr != (pcre_uchar)terminator)
6066                {                {
6067                *errorcodeptr = ERR42;                *errorcodeptr = ERR42;
6068                goto FAILED;                goto FAILED;
# Line 6097  for (;; ptr++) Line 6204  for (;; ptr++)
6204              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
6205              goto FAILED;              goto FAILED;
6206              }              }
6207            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6208              {              {
6209              *errorcodeptr = ERR42;              *errorcodeptr = ERR42;
6210              goto FAILED;              goto FAILED;
# Line 6203  for (;; ptr++) Line 6310  for (;; ptr++)
6310            while(IS_DIGIT(*ptr))            while(IS_DIGIT(*ptr))
6311              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6312    
6313            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6314              {              {
6315              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
6316              goto FAILED;              goto FAILED;
# Line 6307  for (;; ptr++) Line 6414  for (;; ptr++)
6414    
6415          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6416    
6417          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6418          continue;          continue;
6419    
6420    
# Line 6441  for (;; ptr++) Line 6548  for (;; ptr++)
6548           cond_depth +           cond_depth +
6549             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6550           &subfirstchar,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6551             &subfirstcharflags,
6552           &subreqchar,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6553             &subreqcharflags,
6554           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6555           cd,                              /* Tables block */           cd,                              /* Tables block */
6556           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6502  for (;; ptr++) Line 6611  for (;; ptr++)
6611            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6612            goto FAILED;            goto FAILED;
6613            }            }
6614          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;          if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
6615          }          }
6616        }        }
6617    
# Line 6551  for (;; ptr++) Line 6660  for (;; ptr++)
6660      back off. */      back off. */
6661    
6662      zeroreqchar = reqchar;      zeroreqchar = reqchar;
6663        zeroreqcharflags = reqcharflags;
6664      zerofirstchar = firstchar;      zerofirstchar = firstchar;
6665        zerofirstcharflags = firstcharflags;
6666      groupsetfirstchar = FALSE;      groupsetfirstchar = FALSE;
6667    
6668      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
# Line 6562  for (;; ptr++) Line 6673  for (;; ptr++)
6673        no firstchar, set "none" for the whole branch. In both cases, a zero        no firstchar, set "none" for the whole branch. In both cases, a zero
6674        repeat forces firstchar to "none". */        repeat forces firstchar to "none". */
6675    
6676        if (firstchar == REQ_UNSET)        if (firstcharflags == REQ_UNSET)
6677          {          {
6678          if (subfirstchar >= 0)          if (subfirstcharflags >= 0)
6679            {            {
6680            firstchar = subfirstchar;            firstchar = subfirstchar;
6681              firstcharflags = subfirstcharflags;
6682            groupsetfirstchar = TRUE;            groupsetfirstchar = TRUE;
6683            }            }
6684          else firstchar = REQ_NONE;          else firstcharflags = REQ_NONE;
6685          zerofirstchar = REQ_NONE;          zerofirstcharflags = REQ_NONE;
6686          }          }
6687    
6688        /* If firstchar was previously set, convert the subpattern's firstchar        /* If firstchar was previously set, convert the subpattern's firstchar
6689        into reqchar if there wasn't one, using the vary flag that was in        into reqchar if there wasn't one, using the vary flag that was in
6690        existence beforehand. */        existence beforehand. */
6691    
6692        else if (subfirstchar >= 0 && subreqchar < 0)        else if (subfirstcharflags >= 0 && subreqcharflags < 0)
6693          subreqchar = subfirstchar | tempreqvary;          {
6694            subreqchar = subfirstchar;
6695            subreqcharflags = subfirstcharflags | tempreqvary;
6696            }
6697    
6698        /* If the subpattern set a required byte (or set a first byte that isn't        /* If the subpattern set a required byte (or set a first byte that isn't
6699        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6700    
6701        if (subreqchar >= 0) reqchar = subreqchar;        if (subreqcharflags >= 0)
6702            {
6703            reqchar = subreqchar;
6704            reqcharflags = subreqcharflags;
6705            }
6706        }        }
6707    
6708      /* For a forward assertion, we take the reqchar, if set. This can be      /* For a forward assertion, we take the reqchar, if set. This can be
# Line 6594  for (;; ptr++) Line 6713  for (;; ptr++)
6713      of a firstchar. This is overcome by a scan at the end if there's no      of a firstchar. This is overcome by a scan at the end if there's no
6714      firstchar, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6715    
6716      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;      else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
6717          {
6718          reqchar = subreqchar;
6719          reqcharflags = subreqcharflags;
6720          }
6721      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6722    
6723    
# Line 6602  for (;; ptr++) Line 6725  for (;; ptr++)
6725      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
6726      are arranged to be the negation of the corresponding OP_values in the      are arranged to be the negation of the corresponding OP_values in the
6727      default case when PCRE_UCP is not set. For the back references, the values      default case when PCRE_UCP is not set. For the back references, the values
6728      are ESC_REF plus the reference number. Only back references and those types      are negative the reference number. Only back references and those types
6729      that consume a character may be repeated. We can test for values between      that consume a character may be repeated. We can test for values between
6730      ESC_b and ESC_Z for the latter; this may have to change if any new ones are      ESC_b and ESC_Z for the latter; this may have to change if any new ones are
6731      ever created. */      ever created. */
6732    
6733      case CHAR_BACKSLASH:      case CHAR_BACKSLASH:
6734      tempptr = ptr;      tempptr = ptr;
6735      c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE);      escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, FALSE);
6736    
6737      if (*errorcodeptr != 0) goto FAILED;      if (*errorcodeptr != 0) goto FAILED;
6738    
6739      if (c < 0)      if (escape == 0)
6740          c = ec;
6741        else
6742        {        {
6743        if (-c == ESC_Q)            /* Handle start of quoted string */        if (escape == ESC_Q)            /* Handle start of quoted string */
6744          {          {
6745          if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)          if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
6746            ptr += 2;               /* avoid empty string */            ptr += 2;               /* avoid empty string */
# Line 6622  for (;; ptr++) Line 6748  for (;; ptr++)
6748          continue;          continue;
6749          }          }
6750    
6751        if (-c == ESC_E) continue;  /* Perl ignores an orphan \E */        if (escape == ESC_E) continue;  /* Perl ignores an orphan \E */
6752    
6753        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6754        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6755    
6756        if (firstchar == REQ_UNSET && -c > ESC_b && -c < ESC_Z)        if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6757          firstchar = REQ_NONE;          firstcharflags = REQ_NONE;
6758    
6759        /* Set values to reset to if this is followed by a zero repeat. */        /* Set values to reset to if this is followed by a zero repeat. */
6760    
6761        zerofirstchar = firstchar;        zerofirstchar = firstchar;
6762          zerofirstcharflags = firstcharflags;
6763        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6764          zeroreqcharflags = reqcharflags;
6765    
6766        /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'        /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'
6767        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
6768        -ESC_g is returned only for these cases. So we don't need to check for <        ESC_g is returned only for these cases. So we don't need to check for <
6769        or ' if the value is -ESC_g. For the Perl syntax \g{n} the value is        or ' if the value is ESC_g. For the Perl syntax \g{n} the value is
6770        -ESC_REF+n, and for the Perl syntax \g{name} the result is -ESC_k (as        -n, and for the Perl syntax \g{name} the result is ESC_k (as
6771        that is a synonym for a named back reference). */        that is a synonym for a named back reference). */
6772    
6773        if (-c == ESC_g)        if (escape == ESC_g)
6774          {          {
6775          const pcre_uchar *p;          const pcre_uchar *p;
6776          save_hwm = cd->hwm;   /* Normally this is set when '(' is read */          save_hwm = cd->hwm;   /* Normally this is set when '(' is read */
# Line 6662  for (;; ptr++) Line 6790  for (;; ptr++)
6790          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
6791            {            {
6792            BOOL is_a_number = TRUE;            BOOL is_a_number = TRUE;
6793            for (p = ptr + 1; *p != 0 && *p != terminator; p++)            for (p = ptr + 1; *p != 0 && *p != (pcre_uchar)terminator; p++)
6794              {              {
6795              if (!MAX_255(*p)) { is_a_number = FALSE; break; }              if (!MAX_255(*p)) { is_a_number = FALSE; break; }
6796              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
6797              if ((cd->ctypes[*p] & ctype_word) == 0) break;              if ((cd->ctypes[*p] & ctype_word) == 0) break;
6798              }              }
6799            if (*p != terminator)            if (*p != (pcre_uchar)terminator)
6800              {              {
6801              *errorcodeptr = ERR57;              *errorcodeptr = ERR57;
6802              break;              break;
# Line 6686  for (;; ptr++) Line 6814  for (;; ptr++)
6814    
6815          p = ptr + 2;          p = ptr + 2;
6816          while (IS_DIGIT(*p)) p++;          while (IS_DIGIT(*p)) p++;
6817          if (*p != terminator)          if (*p != (pcre_uchar)terminator)
6818            {            {
6819            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
6820            break;            break;
# Line 6698  for (;; ptr++) Line 6826  for (;; ptr++)
6826        /* \k<name> or \k'name' is a back reference by name (Perl syntax).        /* \k<name> or \k'name' is a back reference by name (Perl syntax).
6827        We also support \k{name} (.NET syntax).  */        We also support \k{name} (.NET syntax).  */
6828    
6829        if (-c == ESC_k)        if (escape == ESC_k)
6830          {          {
6831          if ((ptr[1] != CHAR_LESS_THAN_SIGN &&          if ((ptr[1] != CHAR_LESS_THAN_SIGN &&
6832            ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))            ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))
# Line 6717  for (;; ptr++) Line 6845  for (;; ptr++)
6845        not set to cope with cases like (?=(\w+))\1: which would otherwise set        not set to cope with cases like (?=(\w+))\1: which would otherwise set
6846        ':' later. */        ':' later. */
6847    
6848        if (-c >= ESC_REF)        if (escape < 0)
6849          {          {
6850          open_capitem *oc;          open_capitem *oc;
6851          recno = -c - ESC_REF;          recno = -escape;
6852    
6853          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6854          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6855          previous = code;          previous = code;
6856          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6857          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6747  for (;; ptr++) Line 6875  for (;; ptr++)
6875        /* So are Unicode property matches, if supported. */        /* So are Unicode property matches, if supported. */
6876    
6877  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6878        else if (-c == ESC_P || -c == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6879          {          {
6880          BOOL negated;          BOOL negated;
6881          int pdata;          unsigned int ptype = 0, pdata = 0;
6882          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
6883          if (ptype < 0) goto FAILED;            goto FAILED;
6884          previous = code;          previous = code;
6885          *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6886          *code++ = ptype;          *code++ = ptype;
6887          *code++ = pdata;          *code++ = pdata;
6888          }          }
# Line 6763  for (;; ptr++) Line 6891  for (;; ptr++)
6891        /* If Unicode properties are not supported, \X, \P, and \p are not        /* If Unicode properties are not supported, \X, \P, and \p are not
6892        allowed. */        allowed. */
6893    
6894        else if (-c == ESC_X || -c == ESC_P || -c == ESC_p)        else if (escape == ESC_X || escape == ESC_P || escape == ESC_p)
6895          {          {
6896          *errorcodeptr = ERR45;          *errorcodeptr = ERR45;
6897          goto FAILED;          goto FAILED;
# Line 6778  for (;; ptr++) Line 6906  for (;; ptr++)
6906    
6907        else        else
6908          {          {
6909          if ((-c == ESC_b || -c == ESC_B) && cd->max_lookbehind == 0)          if ((escape == ESC_b || escape == ESC_B) && cd->max_lookbehind == 0)
6910            cd->max_lookbehind = 1;            cd->max_lookbehind = 1;
6911  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6912          if (-c >= ESC_DU && -c <= ESC_wu)          if (escape >= ESC_DU && escape <= ESC_wu)
6913            {            {
6914            nestptr = ptr + 1;                   /* Where to resume */            nestptr = ptr + 1;                   /* Where to resume */
6915            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */            ptr = substitutes[escape - ESC_DU] - 1;  /* Just before substitute */
6916            }            }
6917          else          else
6918  #endif  #endif
# Line 6792  for (;; ptr++) Line 6920  for (;; ptr++)
6920          so that it works in DFA mode and in lookbehinds. */          so that it works in DFA mode and in lookbehinds. */
6921    
6922            {            {
6923            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (escape > ESC_b && escape < ESC_Z)? code : NULL;
6924            *code++ = (!utf && c == -ESC_C)? OP_ALLANY : -c;            *code++ = (!utf && escape == ESC_C)? OP_ALLANY : escape;
6925            }            }
6926          }          }
6927        continue;        continue;
# Line 6803  for (;; ptr++) Line 6931  for (;; ptr++)
6931      a value > 127. We set its representation in the length/buffer, and then      a value > 127. We set its representation in the length/buffer, and then
6932      handle it as a data character. */      handle it as a data character. */
6933    
6934  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
6935      if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)      if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
6936        mclength = PRIV(ord2utf)(c, mcbuffer);        mclength = PRIV(ord2utf)(c, mcbuffer);
6937      else      else
# Line 6836  for (;; ptr++) Line 6964  for (;; ptr++)
6964    
6965      ONE_CHAR:      ONE_CHAR:
6966      previous = code;      previous = code;
6967    
6968        /* For caseless UTF-8 mode when UCP support is available, check whether
6969        this character has more than one other case. If so, generate a special
6970        OP_PROP item instead of OP_CHARI. */
6971    
6972    #ifdef SUPPORT_UCP
6973        if (utf && (options & PCRE_CASELESS) != 0)
6974          {
6975          GETCHAR(c, mcbuffer);
6976          if ((c = UCD_CASESET(c)) != 0)
6977            {
6978            *code++ = OP_PROP;
6979            *code++ = PT_CLIST;
6980            *code++ = c;
6981            if (firstcharflags == REQ_UNSET) firstcharflags = zerofirstcharflags = REQ_NONE;
6982            break;
6983            }
6984          }
6985    #endif
6986    
6987        /* Caseful matches, or not one of the multicase characters. */
6988    
6989      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;
6990      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
6991    
# Line 6849  for (;; ptr++) Line 6999  for (;; ptr++)
6999      Otherwise, leave the firstchar value alone, and don't change it on a zero      Otherwise, leave the firstchar value alone, and don't change it on a zero
7000      repeat. */      repeat. */
7001    
7002      if (firstchar == REQ_UNSET)      if (firstcharflags == REQ_UNSET)
7003        {        {
7004        zerofirstchar = REQ_NONE;        zerofirstcharflags = REQ_NONE;
7005        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7006          zeroreqcharflags = reqcharflags;
7007    
7008        /* If the character is more than one byte long, we can set firstchar        /* If the character is more than one byte long, we can set firstchar
7009        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
# Line 6860  for (;; ptr++) Line 7011  for (;; ptr++)
7011        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7012          {          {
7013          firstchar = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
7014          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;          firstchar = mcbuffer[0];
7015            firstcharflags = req_caseopt;
7016    
7017            if (mclength != 1)
7018              {
7019              reqchar = code[-1];
7020              reqcharflags = cd->req_varyopt;
7021              }
7022          }          }
7023        else firstchar = reqchar = REQ_NONE;        else firstcharflags = reqcharflags = REQ_NONE;
7024        }        }
7025    
7026      /* firstchar was previously set; we can set reqchar only if the length is      /* firstchar was previously set; we can set reqchar only if the length is
# Line 6871  for (;; ptr++) Line 7029  for (;; ptr++)
7029      else      else
7030        {        {
7031        zerofirstchar = firstchar;        zerofirstchar = firstchar;
7032          zerofirstcharflags = firstcharflags;
7033        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7034          zeroreqcharflags = reqcharflags;
7035        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7036          reqchar = code[-1] | req_caseopt | cd->req_varyopt;          {
7037            reqchar = code[-1];
7038            reqcharflags = req_caseopt | cd->req_varyopt;
7039            }
7040        }        }
7041    
7042      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6892  return FALSE; Line 7055  return FALSE;
7055    
7056    
7057    
   
7058  /*************************************************  /*************************************************
7059  *     Compile sequence of alternatives           *  *     Compile sequence of alternatives           *
7060  *************************************************/  *************************************************/
# Line 6913  Arguments: Line 7075  Arguments:
7075    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
7076    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
7077    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
7078    firstcharptr   place to put the first required character, or a negative number    firstcharptr    place to put the first required character
7079    reqcharptr     place to put the last required character, or a negative number    firstcharflagsptr place to put the first character flags, or a negative number
7080      reqcharptr     place to put the last required character
7081      reqcharflagsptr place to put the last required character flags, or a negative number
7082    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
7083    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
7084    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 6926  Returns:         TRUE on success Line 7090  Returns:         TRUE on success
7090  static BOOL  static BOOL
7091  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
7092    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
7093    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,    int cond_depth,
7094      pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
7095      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
7096    branch_chain *bcptr, compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
7097  {  {
7098  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
# Line 6936  pcre_uchar *start_bracket = code; Line 7102  pcre_uchar *start_bracket = code;
7102  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
7103  open_capitem capitem;  open_capitem capitem;
7104  int capnumber = 0;  int capnumber = 0;
7105  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7106  pcre_int32 branchfirstchar, branchreqchar;  pcre_int32 firstcharflags, reqcharflags;
7107    pcre_uint32 branchfirstchar, branchreqchar;
7108    pcre_int32 branchfirstcharflags, branchreqcharflags;
7109  int length;  int length;
7110  int orig_bracount;  int orig_bracount;
7111  int max_bracount;  int max_bracount;
# Line 6946  branch_chain bc; Line 7114  branch_chain bc;
7114  bc.outer = bcptr;  bc.outer = bcptr;
7115  bc.current_branch = code;  bc.current_branch = code;
7116    
7117  firstchar = reqchar = REQ_UNSET;  firstchar = reqchar = 0;
7118    firstcharflags = reqcharflags = REQ_UNSET;
7119    
7120  /* Accumulate the length for use in the pre-compile phase. Start with the  /* Accumulate the length for use in the pre-compile phase. Start with the
7121  length of the BRA and KET and any extra bytes that are required at the  length of the BRA and KET and any extra bytes that are required at the
# Line 7006  for (;;) Line 7175  for (;;)
7175    into the length. */    into the length. */
7176    
7177    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
7178          &branchreqchar, &bc, cond_depth, cd,          &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
7179          (lengthptr == NULL)? NULL : &length))          cond_depth, cd, (lengthptr == NULL)? NULL : &length))
7180      {      {
7181      *ptrptr = ptr;      *ptrptr = ptr;
7182      return FALSE;      return FALSE;
# Line 7028  for (;;) Line 7197  for (;;)
7197      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
7198        {        {
7199        firstchar = branchfirstchar;        firstchar = branchfirstchar;
7200          firstcharflags = branchfirstcharflags;
7201        reqchar = branchreqchar;        reqchar = branchreqchar;
7202          reqcharflags = branchreqcharflags;
7203        }        }
7204    
7205      /* If this is not the first branch, the first char and reqchar have to      /* If this is not the first branch, the first char and reqchar have to
# Line 7042  for (;;) Line 7213  for (;;)
7213        we have to abandon the firstchar for the regex, but if there was        we have to abandon the firstchar for the regex, but if there was
7214        previously no reqchar, it takes on the value of the old firstchar. */        previously no reqchar, it takes on the value of the old firstchar. */
7215    
7216        if (firstchar >= 0 && firstchar != branchfirstchar)        if (firstcharflags >= 0 &&
7217              (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar))
7218          {          {
7219          if (reqchar < 0) reqchar = firstchar;          if (reqcharflags < 0)
7220          firstchar = REQ_NONE;            {
7221              reqchar = firstchar;
7222              reqcharflags = firstcharflags;
7223              }
7224            firstcharflags = REQ_NONE;
7225          }          }
7226    
7227        /* If we (now or from before) have no firstchar, a firstchar from the        /* If we (now or from before) have no firstchar, a firstchar from the
7228        branch becomes a reqchar if there isn't a branch reqchar. */        branch becomes a reqchar if there isn't a branch reqchar. */
7229    
7230        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)        if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
7231            branchreqchar = branchfirstchar;          {
7232            branchreqchar = branchfirstchar;
7233            branchreqcharflags = branchfirstcharflags;
7234            }
7235    
7236        /* Now ensure that the reqchars match */        /* Now ensure that the reqchars match */
7237    
7238        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))        if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) ||
7239          reqchar = REQ_NONE;            reqchar != branchreqchar)
7240        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */          reqcharflags = REQ_NONE;
7241          else
7242            {
7243            reqchar = branchreqchar;
7244            reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */
7245            }
7246        }        }
7247    
7248      /* If lookbehind, check that this branch matches a fixed-length string, and      /* If lookbehind, check that this branch matches a fixed-length string, and
# Line 7154  for (;;) Line 7338  for (;;)
7338      *codeptr = code;      *codeptr = code;
7339      *ptrptr = ptr;      *ptrptr = ptr;
7340      *firstcharptr = firstchar;      *firstcharptr = firstchar;
7341        *firstcharflagsptr = firstcharflags;
7342      *reqcharptr = reqchar;      *reqcharptr = reqchar;
7343        *reqcharflagsptr = reqcharflags;
7344      if (lengthptr != NULL)      if (lengthptr != NULL)
7345        {        {
7346        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7441  we return that char, otherwise -1. Line 7627  we return that char, otherwise -1.
7627    
7628  Arguments:  Arguments:
7629    code       points to start of expression (the bracket)    code       points to start of expression (the bracket)
7630      flags       points to the first char flags, or to REQ_NONE
7631    inassert   TRUE if in an assertion    inassert   TRUE if in an assertion
7632    
7633  Returns:     -1 or the fixed first char  Returns:     the fixed first char, or 0 with REQ_NONE in flags
7634  */  */
7635    
7636  static int  static pcre_uint32
7637  find_firstassertedchar(const pcre_uchar *code, BOOL inassert)  find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags,
7638      BOOL inassert)
7639  {  {
7640  register int c = -1;  register pcre_uint32 c = 0;
7641    int cflags = REQ_NONE;
7642    
7643    *flags = REQ_NONE;
7644  do {  do {
7645     int d;     pcre_uint32 d;
7646       int dflags;
7647     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7648               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7649     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7650       TRUE);       TRUE);
7651     register int op = *scode;     register pcre_uchar op = *scode;
7652    
7653     switch(op)     switch(op)
7654       {       {
7655       default:       default:
7656       return -1;       return 0;
7657    
7658       case OP_BRA:       case OP_BRA:
7659       case OP_BRAPOS:       case OP_BRAPOS:
# Line 7473  do { Line 7665  do {
7665       case OP_ONCE:       case OP_ONCE:
7666       case OP_ONCE_NC:       case OP_ONCE_NC:
7667       case OP_COND:       case OP_COND:
7668       if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)       d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
7669         return -1;       if (dflags < 0)
7670       if (c < 0) c = d; else if (c != d) return -1;         return 0;
7671         if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0;
7672       break;       break;
7673    
7674       case OP_EXACT:       case OP_EXACT:
# Line 7486  do { Line 7679  do {
7679       case OP_PLUS:       case OP_PLUS:
7680       case OP_MINPLUS:       case OP_MINPLUS:
7681       case OP_POSPLUS:       case OP_POSPLUS:
7682       if (!inassert) return -1;       if (!inassert) return 0;
7683       if (c < 0) c = scode[1];       if (cflags < 0) { c = scode[1]; cflags = 0; }
7684         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7685       break;       break;
7686    
7687       case OP_EXACTI:       case OP_EXACTI:
# Line 7499  do { Line 7692  do {
7692       case OP_PLUSI:       case OP_PLUSI:
7693       case OP_MINPLUSI:       case OP_MINPLUSI:
7694       case OP_POSPLUSI:       case OP_POSPLUSI:
7695       if (!inassert) return -1;       if (!inassert) return 0;
7696       if (c < 0) c = scode[1] | REQ_CASELESS;       if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
7697         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7698       break;       break;
7699       }       }
7700    
7701     code += GET(code, 1);     code += GET(code, 1);
7702     }     }
7703  while (*code == OP_ALT);  while (*code == OP_ALT);
7704    
7705    *flags = cflags;
7706  return c;  return c;
7707  }  }
7708    
# Line 7535  Returns:        pointer to compiled data Line 7730  Returns:        pointer to compiled data
7730                  with errorptr and erroroffset set                  with errorptr and erroroffset set
7731  */  */
7732    
7733  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7734  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7735  pcre_compile(const char *pattern, int options, const char **errorptr,  pcre_compile(const char *pattern, int options, const char **errorptr,
7736    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7737  #else  #elif defined COMPILE_PCRE16
7738  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
7739  pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,  pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,
7740    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7741    #elif defined COMPILE_PCRE32
7742    PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
7743    pcre32_compile(PCRE_SPTR32 pattern, int options, const char **errorptr,
7744      int *erroroffset, const unsigned char *tables)
7745  #endif  #endif
7746  {  {
7747  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7748  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7749  #else  #elif defined COMPILE_PCRE16
7750  return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7751    #elif defined COMPILE_PCRE32
7752    return pcre32_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7753  #endif  #endif
7754  }  }
7755    
7756    
7757  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7758  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7759  pcre_compile2(const char *pattern, int options, int *errorcodeptr,  pcre_compile2(const char *pattern, int options, int *errorcodeptr,
7760    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7761  #else  #elif defined COMPILE_PCRE16
7762  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
7763  pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,  pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,
7764    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7765    #elif defined COMPILE_PCRE32
7766    PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
7767    pcre32_compile2(PCRE_SPTR32 pattern, int options, int *errorcodeptr,
7768      const char **errorptr, int *erroroffset, const unsigned char *tables)
7769  #endif  #endif
7770  {  {
7771  REAL_PCRE *re;  REAL_PCRE *re;
7772  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7773  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7774    pcre_int32 firstcharflags, reqcharflags;
7775  int newline;  int newline;
7776  int errorcode = 0;  int errorcode = 0;
7777  int skipatstart = 0;  int skipatstart = 0;
# Line 7646  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7852  while (ptr[skipatstart] == CHAR_LEFT_PAR
7852    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)
7853      { skipatstart += 8; options |= PCRE_UTF16; continue; }      { skipatstart += 8; options |= PCRE_UTF16; continue; }
7854  #endif  #endif
7855    #ifdef COMPILE_PCRE32
7856      if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)
7857        { skipatstart += 8; options |= PCRE_UTF32; continue; }
7858    #endif
7859    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)
7860      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
7861    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)
# Line 7674  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7884  while (ptr[skipatstart] == CHAR_LEFT_PAR
7884    else break;    else break;
7885    }    }
7886    
7887  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF(16|32) have the same value as PCRE_UTF8. */
7888  utf = (options & PCRE_UTF8) != 0;  utf = (options & PCRE_UTF8) != 0;
7889    
7890  /* Can't support UTF unless PCRE has been compiled to include the code. The  /* Can't support UTF unless PCRE has been compiled to include the code. The
# Line 7686  not used here. */ Line 7896  not used here. */
7896  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&
7897       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)
7898    {    {
7899  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7900    errorcode = ERR44;    errorcode = ERR44;
7901  #else  #elif defined COMPILE_PCRE16
7902    errorcode = ERR74;    errorcode = ERR74;
7903    #elif defined COMPILE_PCRE32
7904      errorcode = ERR77;
7905  #endif  #endif
7906    goto PCRE_EARLY_ERROR_RETURN2;    goto PCRE_EARLY_ERROR_RETURN2;
7907    }    }
# Line 7809  ptr += skipatstart; Line 8021  ptr += skipatstart;
8021  code = cworkspace;  code = cworkspace;
8022  *code = OP_BRA;  *code = OP_BRA;
8023  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8024    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8025      cd, &length);
8026  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
8027    
8028  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 7853  re->name_count = cd->names_found; Line 8066  re->name_count = cd->names_found;
8066  re->ref_count = 0;  re->ref_count = 0;
8067  re->tables = (tables == PRIV(default_tables))? NULL : tables;  re->tables = (tables == PRIV(default_tables))? NULL : tables;
8068  re->nullpad = NULL;  re->nullpad = NULL;
8069    #ifdef COMPILE_PCRE32
8070    re->dummy1 = re->dummy2 = 0;
8071    #endif
8072    
8073  /* The starting points of the name/number translation table and of the code are  /* The starting points of the name/number translation table and of the code are
8074  passed around in the compile data block. The start/end pattern and initial  passed around in the compile data block. The start/end pattern and initial
# Line 7884  ptr = (const pcre_uchar *)pattern + skip Line 8100  ptr = (const pcre_uchar *)pattern + skip
8100  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
8101  *code = OP_BRA;  *code = OP_BRA;
8102  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
8103    &firstchar, &reqchar, NULL, cd, NULL);    &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
8104  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
8105  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
8106  re->max_lookbehind = cd->max_lookbehind;  re->max_lookbehind = cd->max_lookbehind;
8107  re->flags = cd->external_flags | PCRE_MODE;  re->flags = cd->external_flags | PCRE_MODE;
8108    
8109  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept)
8110      {
8111      reqchar = 0;              /* Must disable after (*ACCEPT) */
8112      reqcharflags = REQ_NONE;
8113      }
8114    
8115  /* If not reached end of pattern on success, there's an excess bracket. */  /* If not reached end of pattern on success, there's an excess bracket. */
8116    
# Line 7940  if (errorcode == 0 && re->top_backref > Line 8160  if (errorcode == 0 && re->top_backref >
8160    
8161  /* If there were any lookbehind assertions that contained OP_RECURSE  /* If there were any lookbehind assertions that contained OP_RECURSE
8162  (recursions or subroutine calls), a flag is set for them to be checked here,  (recursions or subroutine calls), a flag is set for them to be checked here,
8163  because they may contain forward references. Actual recursions can't be fixed  because they may contain forward references. Actual recursions cannot be fixed
8164  length, but subroutine calls can. It is done like this so that those without  length, but subroutine calls can. It is done like this so that those without
8165  OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The  OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The
8166  exceptional ones forgo this. We scan the pattern to check that they are fixed  exceptional ones forgo this. We scan the pattern to check that they are fixed
# Line 8011  if ((re->options & PCRE_ANCHORED) == 0) Line 8231  if ((re->options & PCRE_ANCHORED) == 0)
8231    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
8232    else    else
8233      {      {
8234      if (firstchar < 0)      if (firstcharflags < 0)
8235        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
8236      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstcharflags >= 0)   /* Remove caseless flag for non-caseable chars */
8237        {        {
8238  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8239        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8240  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
8241        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8242    #elif defined COMPILE_PCRE32
8243          re->first_char = firstchar;
8244  #endif  #endif
8245  #endif        if ((firstcharflags & REQ_CASELESS) != 0)
       if ((firstchar & REQ_CASELESS) != 0)  
8246          {          {
8247  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8248          /* We ignore non-ASCII first chars in 8 bit mode. */          /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8054  if ((re->options & PCRE_ANCHORED) == 0) Line 8274  if ((re->options & PCRE_ANCHORED) == 0)
8274  variable length item in the regex. Remove the caseless flag for non-caseable  variable length item in the regex. Remove the caseless flag for non-caseable
8275  bytes. */  bytes. */
8276    
8277  if (reqchar >= 0 &&  if (reqcharflags >= 0 &&
8278       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0))
8279    {    {
8280  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8281    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8282  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
8283    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8284    #elif defined COMPILE_PCRE32
8285      re->req_char = reqchar;
8286  #endif  #endif
8287  #endif    if ((reqcharflags & REQ_CASELESS) != 0)
   if ((reqchar & REQ_CASELESS) != 0)  
8288      {      {
8289  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8290      /* We ignore non-ASCII first chars in 8 bit mode. */      /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8114  if ((re->flags & PCRE_REQCHSET) != 0) Line 8334  if ((re->flags & PCRE_REQCHSET) != 0)
8334      else printf("Req char = \\x%02x%s\n", ch, caseless);      else printf("Req char = \\x%02x%s\n", ch, caseless);
8335    }    }
8336    
8337  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8338  pcre_printint((pcre *)re, stdout, TRUE);  pcre_printint((pcre *)re, stdout, TRUE);
8339  #else  #elif defined COMPILE_PCRE16
8340  pcre16_printint((pcre *)re, stdout, TRUE);  pcre16_printint((pcre *)re, stdout, TRUE);
8341    #elif defined COMPILE_PCRE32
8342    pcre32_printint((pcre *)re, stdout, TRUE);
8343  #endif  #endif
8344    
8345  /* This check is done here in the debugging case so that the code that  /* This check is done here in the debugging case so that the code that
# Line 8133  if (code - codestart > length) Line 8355  if (code - codestart > length)
8355    }    }
8356  #endif   /* PCRE_DEBUG */  #endif   /* PCRE_DEBUG */
8357    
8358  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8359  return (pcre *)re;  return (pcre *)re;
8360  #else  #elif defined COMPILE_PCRE16
8361  return (pcre16 *)re;  return (pcre16 *)re;
8362    #elif defined COMPILE_PCRE32
8363    return (pcre32 *)re;
8364  #endif  #endif
8365  }  }
8366    

Legend:
Removed from v.1045  
changed lines
  Added in v.1133

  ViewVC Help
Powered by ViewVC 1.1.5