/[pcre]/code/branches/pcre16/pcre_compile.c
ViewVC logotype

Diff of /code/branches/pcre16/pcre_compile.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 756 by ph10, Mon Nov 21 10:48:42 2011 UTC revision 794 by zherczeg, Thu Dec 8 07:36:41 2011 UTC
# Line 97  overrun before it actually does run off Line 97  overrun before it actually does run off
97    
98  #define WORK_SIZE_CHECK (COMPILE_WORK_SIZE - 100)  #define WORK_SIZE_CHECK (COMPILE_WORK_SIZE - 100)
99    
100    /* Private flags added to firstchar and reqchar. */
101    
102    #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */
103    #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */
104    
105    /* Repeated character flags. */
106    
107    #define UTF_LENGTH     0x10000000l      /* The char contains its length. */
108    
109  /* Table for handling escaped characters in the range '0'-'z'. Positive returns  /* Table for handling escaped characters in the range '0'-'z'. Positive returns
110  are simple data values; negative values are for special things like \d and so  are simple data values; negative values are for special things like \d and so
# Line 266  substitutes must be in the order of the Line 274  substitutes must be in the order of the
274  both positive and negative cases. NULL means no substitute. */  both positive and negative cases. NULL means no substitute. */
275    
276  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
277  static const pcre_uchar literal_PNd[]  = { '\\', 'P', '{', 'N', 'd', '}', '\0' };  static const pcre_uchar string_PNd[]  = {
278  static const pcre_uchar literal_pNd[]  = { '\\', 'p', '{', 'N', 'd', '}', '\0' };    CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET,
279  static const pcre_uchar literal_PXsp[] = { '\\', 'P', '{', 'X', 's', 'p', '}', '\0' };    CHAR_N, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' };
280  static const pcre_uchar literal_pXsp[] = { '\\', 'p', '{', 'X', 's', 'p', '}', '\0' };  static const pcre_uchar string_pNd[]  = {
281  static const pcre_uchar literal_PXwd[] = { '\\', 'P', '{', 'X', 'w', 'd', '}', '\0' };    CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET,
282  static const pcre_uchar literal_pXwd[] = { '\\', 'p', '{', 'X', 'w', 'd', '}', '\0' };    CHAR_N, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' };
283    static const pcre_uchar string_PXsp[] = {
284      CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET,
285      CHAR_X, CHAR_s, CHAR_p, CHAR_RIGHT_CURLY_BRACKET, '\0' };
286    static const pcre_uchar string_pXsp[] = {
287      CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET,
288      CHAR_X, CHAR_s, CHAR_p, CHAR_RIGHT_CURLY_BRACKET, '\0' };
289    static const pcre_uchar string_PXwd[] = {
290      CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET,
291      CHAR_X, CHAR_w, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' };
292    static const pcre_uchar string_pXwd[] = {
293      CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET,
294      CHAR_X, CHAR_w, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' };
295    
296  static const pcre_uchar *substitutes[] = {  static const pcre_uchar *substitutes[] = {
297    literal_PNd,           /* \D */    string_PNd,           /* \D */
298    literal_pNd,           /* \d */    string_pNd,           /* \d */
299    literal_PXsp,          /* \S */       /* NOTE: Xsp is Perl space */    string_PXsp,          /* \S */       /* NOTE: Xsp is Perl space */
300    literal_pXsp,          /* \s */    string_pXsp,          /* \s */
301    literal_PXwd,          /* \W */    string_PXwd,          /* \W */
302    literal_pXwd           /* \w */    string_pXwd           /* \w */
303  };  };
304    
305  static const pcre_uchar literal_pL[] =   { '\\', 'p', '{', 'L', '}', '\0' };  static const pcre_uchar string_pL[] =   {
306  static const pcre_uchar literal_pLl[] =  { '\\', 'p', '{', 'L', 'l', '}', '\0' };    CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET,
307  static const pcre_uchar literal_pLu[] =  { '\\', 'p', '{', 'L', 'u', '}', '\0' };    CHAR_L, CHAR_RIGHT_CURLY_BRACKET, '\0' };
308  static const pcre_uchar literal_pXan[] = { '\\', 'p', '{', 'X', 'a', 'n', '}', '\0' };  static const pcre_uchar string_pLl[] =  {
309  static const pcre_uchar literal_h[] =    { '\\', 'h', '\0' };    CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET,
310  static const pcre_uchar literal_pXps[] = { '\\', 'p', '{', 'X', 'p', 's', '}', '\0' };    CHAR_L, CHAR_l, CHAR_RIGHT_CURLY_BRACKET, '\0' };
311  static const pcre_uchar literal_PL[] =   { '\\', 'P', '{', 'L', '}', '\0' };  static const pcre_uchar string_pLu[] =  {
312  static const pcre_uchar literal_PLl[] =  { '\\', 'P', '{', 'L', 'l', '}', '\0' };    CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET,
313  static const pcre_uchar literal_PLu[] =  { '\\', 'P', '{', 'L', 'u', '}', '\0' };    CHAR_L, CHAR_u, CHAR_RIGHT_CURLY_BRACKET, '\0' };
314  static const pcre_uchar literal_PXan[] = { '\\', 'P', '{', 'X', 'a', 'n', '}', '\0' };  static const pcre_uchar string_pXan[] = {
315  static const pcre_uchar literal_H[] =    { '\\', 'H', '\0' };    CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET,
316  static const pcre_uchar literal_PXps[] = { '\\', 'P', '{', 'X', 'p', 's', '}', '\0' };    CHAR_X, CHAR_a, CHAR_n, CHAR_RIGHT_CURLY_BRACKET, '\0' };
317    static const pcre_uchar string_h[] =    {
318      CHAR_BACKSLASH, CHAR_h, '\0' };
319    static const pcre_uchar string_pXps[] = {
320      CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET,
321      CHAR_X, CHAR_p, CHAR_s, CHAR_RIGHT_CURLY_BRACKET, '\0' };
322    static const pcre_uchar string_PL[] =   {
323      CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET,
324      CHAR_L, CHAR_RIGHT_CURLY_BRACKET, '\0' };
325    static const pcre_uchar string_PLl[] =  {
326      CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET,
327      CHAR_L, CHAR_l, CHAR_RIGHT_CURLY_BRACKET, '\0' };
328    static const pcre_uchar string_PLu[] =  {
329      CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET,
330      CHAR_L, CHAR_u, CHAR_RIGHT_CURLY_BRACKET, '\0' };
331    static const pcre_uchar string_PXan[] = {
332      CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET,
333      CHAR_X, CHAR_a, CHAR_n, CHAR_RIGHT_CURLY_BRACKET, '\0' };
334    static const pcre_uchar string_H[] =    {
335      CHAR_BACKSLASH, CHAR_H, '\0' };
336    static const pcre_uchar string_PXps[] = {
337      CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET,
338      CHAR_X, CHAR_p, CHAR_s, CHAR_RIGHT_CURLY_BRACKET, '\0' };
339    
340  static const pcre_uchar *posix_substitutes[] = {  static const pcre_uchar *posix_substitutes[] = {
341    literal_pL,            /* alpha */    string_pL,            /* alpha */
342    literal_pLl,           /* lower */    string_pLl,           /* lower */
343    literal_pLu,           /* upper */    string_pLu,           /* upper */
344    literal_pXan,          /* alnum */    string_pXan,          /* alnum */
345    NULL,                  /* ascii */    NULL,                 /* ascii */
346    literal_h,             /* blank */    string_h,             /* blank */
347    NULL,                  /* cntrl */    NULL,                 /* cntrl */
348    literal_pNd,           /* digit */    string_pNd,           /* digit */
349    NULL,                  /* graph */    NULL,                 /* graph */
350    NULL,                  /* print */    NULL,                 /* print */
351    NULL,                  /* punct */    NULL,                 /* punct */
352    literal_pXps,          /* space */    /* NOTE: Xps is POSIX space */    string_pXps,          /* space */    /* NOTE: Xps is POSIX space */
353    literal_pXwd,          /* word */    string_pXwd,          /* word */
354    NULL,                  /* xdigit */    NULL,                 /* xdigit */
355    /* Negated cases */    /* Negated cases */
356    literal_PL,            /* ^alpha */    string_PL,            /* ^alpha */
357    literal_PLl,           /* ^lower */    string_PLl,           /* ^lower */
358    literal_PLu,           /* ^upper */    string_PLu,           /* ^upper */
359    literal_PXan,          /* ^alnum */    string_PXan,          /* ^alnum */
360    NULL,                  /* ^ascii */    NULL,                 /* ^ascii */
361    literal_H,             /* ^blank */    string_H,             /* ^blank */
362    NULL,                  /* ^cntrl */    NULL,                 /* ^cntrl */
363    literal_PNd,           /* ^digit */    string_PNd,           /* ^digit */
364    NULL,                  /* ^graph */    NULL,                 /* ^graph */
365    NULL,                  /* ^print */    NULL,                 /* ^print */
366    NULL,                  /* ^punct */    NULL,                 /* ^punct */
367    literal_PXps,          /* ^space */   /* NOTE: Xps is POSIX space */    string_PXps,          /* ^space */   /* NOTE: Xps is POSIX space */
368    literal_PXwd,          /* ^word */    string_PXwd,          /* ^word */
369    NULL                   /* ^xdigit */    NULL                  /* ^xdigit */
370  };  };
371  #define POSIX_SUBSIZE (sizeof(posix_substitutes) / sizeof(pcre_uchar *))  #define POSIX_SUBSIZE (sizeof(posix_substitutes) / sizeof(pcre_uchar *))
372  #endif  #endif
# Line 432  static const char error_texts[] = Line 474  static const char error_texts[] =
474    "\\k is not followed by a braced, angle-bracketed, or quoted name\0"    "\\k is not followed by a braced, angle-bracketed, or quoted name\0"
475    /* 70 */    /* 70 */
476    "internal error: unknown opcode in find_fixedlength()\0"    "internal error: unknown opcode in find_fixedlength()\0"
477      "Not allowed UTF-8 / UTF-16 code point (>= 0xd800 && <= 0xdfff)\0"
478    ;    ;
479    
480  /* 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 450  For convenience, we use the same bit def Line 493  For convenience, we use the same bit def
493    
494  Then we can use ctype_digit and ctype_xdigit in the code. */  Then we can use ctype_digit and ctype_xdigit in the code. */
495    
496    /* Using a simple comparison for decimal numbers rather than a memory read
497    is much faster, and the resulting code is simpler (the compiler turns it
498    into a subtraction and unsigned comparison). */
499    
500    #define IS_DIGIT(x) ((x) >= CHAR_0 && (x) <= CHAR_9)
501    
502  #ifndef EBCDIC  #ifndef EBCDIC
503    
504  /* This is the "normal" case, for ASCII systems, and EBCDIC systems running in  /* This is the "normal" case, for ASCII systems, and EBCDIC systems running in
505  UTF-8 mode. */  UTF-8 mode. */
506    
507  static const unsigned char digitab[] =  static const pcre_uint8 digitab[] =
508    {    {
509    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
510    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   8- 15 */    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   8- 15 */
# Line 494  static const unsigned char digitab[] = Line 543  static const unsigned char digitab[] =
543    
544  /* This is the "abnormal" case, for EBCDIC systems not running in UTF-8 mode. */  /* This is the "abnormal" case, for EBCDIC systems not running in UTF-8 mode. */
545    
546  static const unsigned char digitab[] =  static const pcre_uint8 digitab[] =
547    {    {
548    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7  0 */    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7  0 */
549    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   8- 15    */    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   8- 15    */
# Line 529  static const unsigned char digitab[] = Line 578  static const unsigned char digitab[] =
578    0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /*  0 - 7  F0 */    0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /*  0 - 7  F0 */
579    0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/*  8 -255    */    0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/*  8 -255    */
580    
581  static const unsigned char ebcdic_chartab[] = { /* chartable partial dup */  static const pcre_uint8 ebcdic_chartab[] = { /* chartable partial dup */
582    0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /*   0-  7 */    0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /*   0-  7 */
583    0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */    0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
584    0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /*  16- 23 */    0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /*  16- 23 */
# Line 617  Returns:    TRUE or FALSE Line 666  Returns:    TRUE or FALSE
666  static BOOL  static BOOL
667  is_counted_repeat(const pcre_uchar *p)  is_counted_repeat(const pcre_uchar *p)
668  {  {
669  if ((digitab[*p++] & ctype_digit) == 0) return FALSE;  if (!IS_DIGIT(*p)) return FALSE;
670  while ((digitab[*p] & ctype_digit) != 0) p++;  p++;
671    while (IS_DIGIT(*p)) p++;
672  if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;  if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
673    
674  if (*p++ != CHAR_COMMA) return FALSE;  if (*p++ != CHAR_COMMA) return FALSE;
675  if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;  if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
676    
677  if ((digitab[*p++] & ctype_digit) == 0) return FALSE;  if (!IS_DIGIT(*p)) return FALSE;
678  while ((digitab[*p] & ctype_digit) != 0) p++;  p++;
679    while (IS_DIGIT(*p)) p++;
680    
681  return (*p == CHAR_RIGHT_CURLY_BRACKET);  return (*p == CHAR_RIGHT_CURLY_BRACKET);
682  }  }
# Line 660  static int Line 711  static int
711  check_escape(const pcre_uchar **ptrptr, int *errorcodeptr, int bracount,  check_escape(const pcre_uchar **ptrptr, int *errorcodeptr, int bracount,
712    int options, BOOL isclass)    int options, BOOL isclass)
713  {  {
714  BOOL utf8 = (options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
715    BOOL utf = (options & PCRE_UTF8) != 0;
716  const pcre_uchar *ptr = *ptrptr + 1;  const pcre_uchar *ptr = *ptrptr + 1;
717  int c, i;  pcre_int32 c;
718    int i;
719    
720  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */
721  ptr--;                            /* Set pointer back to the last byte */  ptr--;                            /* Set pointer back to the last byte */
# Line 676  in a table. A non-zero result is somethi Line 729  in a table. A non-zero result is somethi
729  Otherwise further processing may be required. */  Otherwise further processing may be required. */
730    
731  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
732  else if (c < CHAR_0 || c > CHAR_z) {}                     /* Not alphanumeric */  /* Not alphanumeric */
733    else if (c < CHAR_0 || c > CHAR_z) {}
734  else if ((i = escapes[c - CHAR_0]) != 0) c = i;  else if ((i = escapes[c - CHAR_0]) != 0) c = i;
735    
736  #else           /* EBCDIC coding */  #else           /* EBCDIC coding */
737  else if (c < 'a' || (ebcdic_chartab[c] & 0x0E) == 0) {}   /* Not alphanumeric */  /* Not alphanumeric */
738    else if (c < 'a' || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}
739  else if ((i = escapes[c - 0x48]) != 0)  c = i;  else if ((i = escapes[c - 0x48]) != 0)  c = i;
740  #endif  #endif
741    
# Line 706  else Line 761  else
761        {        {
762        /* In JavaScript, \u must be followed by four hexadecimal numbers.        /* In JavaScript, \u must be followed by four hexadecimal numbers.
763        Otherwise it is a lowercase u letter. */        Otherwise it is a lowercase u letter. */
764        if ((digitab[ptr[1]] & ctype_xdigit) != 0 && (digitab[ptr[2]] & ctype_xdigit) != 0        if (MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0
765             && (digitab[ptr[3]] & ctype_xdigit) != 0 && (digitab[ptr[4]] & ctype_xdigit) != 0)          && MAX_255(ptr[2]) && (digitab[ptr[2]] & ctype_xdigit) != 0
766            && MAX_255(ptr[3]) && (digitab[ptr[3]] & ctype_xdigit) != 0
767            && MAX_255(ptr[4]) && (digitab[ptr[4]] & ctype_xdigit) != 0)
768          {          {
769          c = 0;          c = 0;
770          for (i = 0; i < 4; ++i)          for (i = 0; i < 4; ++i)
# Line 763  else Line 820  else
820        {        {
821        const pcre_uchar *p;        const pcre_uchar *p;
822        for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++)        for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++)
823          if (*p != CHAR_MINUS && (digitab[*p] & ctype_digit) == 0) break;          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;
824        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)
825          {          {
826          c = -ESC_k;          c = -ESC_k;
# Line 781  else Line 838  else
838        }        }
839      else negated = FALSE;      else negated = FALSE;
840    
841        /* The integer range is limited by the machine's int representation. */
842      c = 0;      c = 0;
843      while ((digitab[ptr[1]] & ctype_digit) != 0)      while (IS_DIGIT(ptr[1]))
844          {
845          if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */
846            {
847            c = -1;
848            break;
849            }
850        c = c * 10 + *(++ptr) - CHAR_0;        c = c * 10 + *(++ptr) - CHAR_0;
851          }
852      if (c < 0)   /* Integer overflow */      if (((unsigned int)c) > INT_MAX) /* Integer overflow */
853        {        {
854          while (IS_DIGIT(ptr[1]))
855            ptr++;
856        *errorcodeptr = ERR61;        *errorcodeptr = ERR61;
857        break;        break;
858        }        }
# Line 834  else Line 900  else
900      if (!isclass)      if (!isclass)
901        {        {
902        oldptr = ptr;        oldptr = ptr;
903          /* The integer range is limited by the machine's int representation. */
904        c -= CHAR_0;        c -= CHAR_0;
905        while ((digitab[ptr[1]] & ctype_digit) != 0)        while (IS_DIGIT(ptr[1]))
906            {
907            if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */
908              {
909              c = -1;
910              break;
911              }
912          c = c * 10 + *(++ptr) - CHAR_0;          c = c * 10 + *(++ptr) - CHAR_0;
913        if (c < 0)    /* Integer overflow */          }
914          if (((unsigned int)c) > INT_MAX) /* Integer overflow */
915          {          {
916            while (IS_DIGIT(ptr[1]))
917              ptr++;
918          *errorcodeptr = ERR61;          *errorcodeptr = ERR61;
919          break;          break;
920          }          }
# Line 871  else Line 947  else
947      c -= CHAR_0;      c -= CHAR_0;
948      while(i++ < 2 && ptr[1] >= CHAR_0 && ptr[1] <= CHAR_7)      while(i++ < 2 && ptr[1] >= CHAR_0 && ptr[1] <= CHAR_7)
949          c = c * 8 + *(++ptr) - CHAR_0;          c = c * 8 + *(++ptr) - CHAR_0;
950      if (!utf8 && c > 255) *errorcodeptr = ERR51;      if (!utf && c > 0xff) *errorcodeptr = ERR51;
951      break;      break;
952    
953      /* \x is complicated. \x{ddd} is a character number which can be greater      /* \x is complicated. \x{ddd} is a character number which can be greater
954      than 0xff in utf8 mode, but only if the ddd are hex digits. If not, { is      than 0xff in utf or non-8bit mode, but only if the ddd are hex digits.
955      treated as a data character. */      If not, { is treated as a data character. */
956    
957      case CHAR_x:      case CHAR_x:
958      if ((options & PCRE_JAVASCRIPT_COMPAT) != 0)      if ((options & PCRE_JAVASCRIPT_COMPAT) != 0)
959        {        {
960        /* In JavaScript, \x must be followed by two hexadecimal numbers.        /* In JavaScript, \x must be followed by two hexadecimal numbers.
961        Otherwise it is a lowercase x letter. */        Otherwise it is a lowercase x letter. */
962        if ((digitab[ptr[1]] & ctype_xdigit) != 0 && (digitab[ptr[2]] & ctype_xdigit) != 0)        if (MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0
963            && MAX_255(ptr[2]) && (digitab[ptr[2]] & ctype_xdigit) != 0)
964          {          {
965          c = 0;          c = 0;
966          for (i = 0; i < 2; ++i)          for (i = 0; i < 2; ++i)
# Line 904  else Line 981  else
981      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
982        {        {
983        const pcre_uchar *pt = ptr + 2;        const pcre_uchar *pt = ptr + 2;
       int count = 0;  
984    
985        c = 0;        c = 0;
986        while ((digitab[*pt] & ctype_xdigit) != 0)        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)
987          {          {
988          register int cc = *pt++;          register int cc = *pt++;
989          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */
         count++;  
990    
991  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
992          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
# Line 920  else Line 995  else
995          if (cc >= CHAR_a && cc <= CHAR_z) cc += 64;  /* Convert to upper case */          if (cc >= CHAR_a && cc <= CHAR_z) cc += 64;  /* Convert to upper case */
996          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
997  #endif  #endif
998    
999    #ifdef COMPILE_PCRE8
1000            if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }
1001    #else
1002    #ifdef COMPILE_PCRE16
1003            if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }
1004    #endif
1005    #endif
1006            }
1007    
1008          if (c < 0)
1009            {
1010            while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;
1011            *errorcodeptr = ERR34;
1012          }          }
1013    
1014        if (*pt == CHAR_RIGHT_CURLY_BRACKET)        if (*pt == CHAR_RIGHT_CURLY_BRACKET)
1015          {          {
1016          if (c < 0 || count > (utf8? 8 : 2)) *errorcodeptr = ERR34;          if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR71;
1017          ptr = pt;          ptr = pt;
1018          break;          break;
1019          }          }
# Line 936  else Line 1025  else
1025      /* Read just a single-byte hex-defined char */      /* Read just a single-byte hex-defined char */
1026    
1027      c = 0;      c = 0;
1028      while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0)      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)
1029        {        {
1030        int cc;                                  /* Some compilers don't like */        int cc;                                  /* Some compilers don't like */
1031        cc = *(++ptr);                           /* ++ in initializers */        cc = *(++ptr);                           /* ++ in initializers */
# Line 1038  get_ucp(const pcre_uchar **ptrptr, BOOL Line 1127  get_ucp(const pcre_uchar **ptrptr, BOOL
1127  {  {
1128  int c, i, bot, top;  int c, i, bot, top;
1129  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
1130  char name[32];  pcre_uchar name[32];
1131    
1132  c = *(++ptr);  c = *(++ptr);
1133  if (c == 0) goto ERROR_RETURN;  if (c == 0) goto ERROR_RETURN;
# Line 1079  else Line 1168  else
1168  /* Search for a recognized property name using binary chop */  /* Search for a recognized property name using binary chop */
1169    
1170  bot = 0;  bot = 0;
1171  top = _pcre_utt_size;  top = PRIV(utt_size);
1172    
1173  while (bot < top)  while (bot < top)
1174    {    {
1175    i = (bot + top) >> 1;    i = (bot + top) >> 1;
1176    c = strcmp(name, _pcre_utt_names + _pcre_utt[i].name_offset);    c = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);
1177    if (c == 0)    if (c == 0)
1178      {      {
1179      *dptr = _pcre_utt[i].value;      *dptr = PRIV(utt)[i].value;
1180      return _pcre_utt[i].type;      return PRIV(utt)[i].type;
1181      }      }
1182    if (c > 0) bot = i + 1; else top = i;    if (c > 0) bot = i + 1; else top = i;
1183    }    }
# Line 1135  int max = -1; Line 1224  int max = -1;
1224  /* 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
1225  an integer overflow. */  an integer overflow. */
1226    
1227  while ((digitab[*p] & ctype_digit) != 0) min = min * 10 + *p++ - CHAR_0;  while (IS_DIGIT(*p)) min = min * 10 + *p++ - CHAR_0;
1228  if (min < 0 || min > 65535)  if (min < 0 || min > 65535)
1229    {    {
1230    *errorcodeptr = ERR5;    *errorcodeptr = ERR5;
# Line 1150  if (*p == CHAR_RIGHT_CURLY_BRACKET) max Line 1239  if (*p == CHAR_RIGHT_CURLY_BRACKET) max
1239    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
1240      {      {
1241      max = 0;      max = 0;
1242      while((digitab[*p] & ctype_digit) != 0) max = max * 10 + *p++ - CHAR_0;      while(IS_DIGIT(*p)) max = max * 10 + *p++ - CHAR_0;
1243      if (max < 0 || max > 65535)      if (max < 0 || max > 65535)
1244        {        {
1245        *errorcodeptr = ERR5;        *errorcodeptr = ERR5;
# Line 1205  Arguments: Line 1294  Arguments:
1294    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1295    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1296    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1297    utf8         TRUE if we are in UTF-8 mode    utf          TRUE if we are in UTF-8 / UTF-16 mode
1298    count        pointer to the current capturing subpattern number (updated)    count        pointer to the current capturing subpattern number (updated)
1299    
1300  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 1213  Returns:       the number of the named s Line 1302  Returns:       the number of the named s
1302    
1303  static int  static int
1304  find_parens_sub(pcre_uchar **ptrptr, compile_data *cd, const pcre_uchar *name, int lorn,  find_parens_sub(pcre_uchar **ptrptr, compile_data *cd, const pcre_uchar *name, int lorn,
1305    BOOL xmode, BOOL utf8, int *count)    BOOL xmode, BOOL utf, int *count)
1306  {  {
1307  pcre_uchar *ptr = *ptrptr;  pcre_uchar *ptr = *ptrptr;
1308  int start_count = *count;  int start_count = *count;
# Line 1290  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1379  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1379        thisname = ptr;        thisname = ptr;
1380        while (*ptr != term) ptr++;        while (*ptr != term) ptr++;
1381        if (name != NULL && lorn == ptr - thisname &&        if (name != NULL && lorn == ptr - thisname &&
1382            strncmp((const char *)name, (const char *)thisname, lorn) == 0)            STRNCMP_UC_UC(name, thisname, lorn) == 0)
1383          return *count;          return *count;
1384        term++;        term++;
1385        }        }
# Line 1333  for (; ptr < cd->end_pattern; ptr++) Line 1422  for (; ptr < cd->end_pattern; ptr++)
1422          {          {
1423          if (ptr[2] == CHAR_E)          if (ptr[2] == CHAR_E)
1424            ptr+= 2;            ptr+= 2;
1425          else if (strncmp((const char *)ptr+2,          else if (STRNCMP_UC_C8(ptr + 2,
1426                   STR_Q STR_BACKSLASH STR_E, 3) == 0)                   STR_Q STR_BACKSLASH STR_E, 3) == 0)
1427            ptr += 4;            ptr += 4;
1428          else          else
# Line 1381  for (; ptr < cd->end_pattern; ptr++) Line 1470  for (; ptr < cd->end_pattern; ptr++)
1470        {        {
1471        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
1472        ptr++;        ptr++;
1473  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1474        if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;        if (utf) FORWARDCHAR(ptr);
1475  #endif  #endif
1476        }        }
1477      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == 0) goto FAIL_EXIT;
# Line 1393  for (; ptr < cd->end_pattern; ptr++) Line 1482  for (; ptr < cd->end_pattern; ptr++)
1482    
1483    if (*ptr == CHAR_LEFT_PARENTHESIS)    if (*ptr == CHAR_LEFT_PARENTHESIS)
1484      {      {
1485      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, count);      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);
1486      if (rc > 0) return rc;      if (rc > 0) return rc;
1487      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == 0) goto FAIL_EXIT;
1488      }      }
# Line 1439  Arguments: Line 1528  Arguments:
1528    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1529    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1530    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1531    utf8         TRUE if we are in UTF-8 mode    utf          TRUE if we are in UTF-8 / UTF-16 mode
1532    
1533  Returns:       the number of the found subpattern, or -1 if not found  Returns:       the number of the found subpattern, or -1 if not found
1534  */  */
1535    
1536  static int  static int
1537  find_parens(compile_data *cd, const pcre_uchar *name, int lorn, BOOL xmode,  find_parens(compile_data *cd, const pcre_uchar *name, int lorn, BOOL xmode,
1538    BOOL utf8)    BOOL utf)
1539  {  {
1540  pcre_uchar *ptr = (pcre_uchar *)cd->start_pattern;  pcre_uchar *ptr = (pcre_uchar *)cd->start_pattern;
1541  int count = 0;  int count = 0;
# Line 1459  matching closing parens. That is why we Line 1548  matching closing parens. That is why we
1548    
1549  for (;;)  for (;;)
1550    {    {
1551    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, &count);    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);
1552    if (rc > 0 || *ptr++ == 0) break;    if (rc > 0 || *ptr++ == 0) break;
1553    }    }
1554    
# Line 1498  for (;;) Line 1587  for (;;)
1587      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1588      if (!skipassert) return code;      if (!skipassert) return code;
1589      do code += GET(code, 1); while (*code == OP_ALT);      do code += GET(code, 1); while (*code == OP_ALT);
1590      code += _pcre_OP_lengths[*code];      code += PRIV(OP_lengths)[*code];
1591      break;      break;
1592    
1593      case OP_WORD_BOUNDARY:      case OP_WORD_BOUNDARY:
# Line 1512  for (;;) Line 1601  for (;;)
1601      case OP_RREF:      case OP_RREF:
1602      case OP_NRREF:      case OP_NRREF:
1603      case OP_DEF:      case OP_DEF:
1604      code += _pcre_OP_lengths[*code];      code += PRIV(OP_lengths)[*code];
1605      break;      break;
1606    
1607      default:      default:
# Line 1542  and doing the check at the end; a flag s Line 1631  and doing the check at the end; a flag s
1631    
1632  Arguments:  Arguments:
1633    code     points to the start of the pattern (the bracket)    code     points to the start of the pattern (the bracket)
1634    utf8     TRUE in UTF-8 mode    utf      TRUE in UTF-8 / UTF-16 mode
1635    atend    TRUE if called when the pattern is complete    atend    TRUE if called when the pattern is complete
1636    cd       the "compile data" structure    cd       the "compile data" structure
1637    
# Line 1554  Returns:   the fixed length, Line 1643  Returns:   the fixed length,
1643  */  */
1644    
1645  static int  static int
1646  find_fixedlength(pcre_uchar *code, BOOL utf8, BOOL atend, compile_data *cd)  find_fixedlength(pcre_uchar *code, BOOL utf, BOOL atend, compile_data *cd)
1647  {  {
1648  int length = -1;  int length = -1;
1649    
# Line 1581  for (;;) Line 1670  for (;;)
1670      case OP_ONCE:      case OP_ONCE:
1671      case OP_ONCE_NC:      case OP_ONCE_NC:
1672      case OP_COND:      case OP_COND:
1673      d = find_fixedlength(cc + ((op == OP_CBRA)? 2:0), utf8, atend, cd);      d = find_fixedlength(cc + ((op == OP_CBRA)? IMM2_SIZE : 0), utf, atend, cd);
1674      if (d < 0) return d;      if (d < 0) return d;
1675      branchlength += d;      branchlength += d;
1676      do cc += GET(cc, 1); while (*cc == OP_ALT);      do cc += GET(cc, 1); while (*cc == OP_ALT);
# Line 1615  for (;;) Line 1704  for (;;)
1704      cs = ce = (pcre_uchar *)cd->start_code + GET(cc, 1);  /* Start subpattern */      cs = ce = (pcre_uchar *)cd->start_code + GET(cc, 1);  /* Start subpattern */
1705      do ce += GET(ce, 1); while (*ce == OP_ALT);           /* End subpattern */      do ce += GET(ce, 1); while (*ce == OP_ALT);           /* End subpattern */
1706      if (cc > cs && cc < ce) return -1;                    /* Recursion */      if (cc > cs && cc < ce) return -1;                    /* Recursion */
1707      d = find_fixedlength(cs + 2, utf8, atend, cd);      d = find_fixedlength(cs + 2, utf, atend, cd);
1708      if (d < 0) return d;      if (d < 0) return d;
1709      branchlength += d;      branchlength += d;
1710      cc += 1 + LINK_SIZE;      cc += 1 + LINK_SIZE;
# Line 1636  for (;;) Line 1725  for (;;)
1725      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
1726      case OP_SKIP_ARG:      case OP_SKIP_ARG:
1727      case OP_THEN_ARG:      case OP_THEN_ARG:
1728      cc += cc[1] + _pcre_OP_lengths[*cc];      cc += cc[1] + PRIV(OP_lengths)[*cc];
1729      break;      break;
1730    
1731      case OP_CALLOUT:      case OP_CALLOUT:
# Line 1663  for (;;) Line 1752  for (;;)
1752      case OP_SOM:      case OP_SOM:
1753      case OP_THEN:      case OP_THEN:
1754      case OP_WORD_BOUNDARY:      case OP_WORD_BOUNDARY:
1755      cc += _pcre_OP_lengths[*cc];      cc += PRIV(OP_lengths)[*cc];
1756      break;      break;
1757    
1758      /* Handle literal characters */      /* Handle literal characters */
# Line 1674  for (;;) Line 1763  for (;;)
1763      case OP_NOTI:      case OP_NOTI:
1764      branchlength++;      branchlength++;
1765      cc += 2;      cc += 2;
1766  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1767      if (utf8 && cc[-1] >= 0xc0) cc += _pcre_utf8_table4[cc[-1] & 0x3f];      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1768  #endif  #endif
1769      break;      break;
1770    
# Line 1687  for (;;) Line 1776  for (;;)
1776      case OP_NOTEXACT:      case OP_NOTEXACT:
1777      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1778      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1779      cc += 4;      cc += 2 + IMM2_SIZE;
1780  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1781      if (utf8 && cc[-1] >= 0xc0) cc += _pcre_utf8_table4[cc[-1] & 0x3f];      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1782  #endif  #endif
1783      break;      break;
1784    
1785      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1786      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1787      if (cc[3] == OP_PROP || cc[3] == OP_NOTPROP) cc += 2;      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP) cc += 2;
1788      cc += 4;      cc += 1 + IMM2_SIZE + 1;
1789      break;      break;
1790    
1791      /* Handle single-char matchers */      /* Handle single-char matchers */
# Line 1730  for (;;) Line 1819  for (;;)
1819    
1820      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1821    
1822  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || defined COMPILE_PCRE16
1823      case OP_XCLASS:      case OP_XCLASS:
1824      cc += GET(cc, 1) - 33;      cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];
1825      /* Fall through */      /* Fall through */
1826  #endif  #endif
1827    
1828      case OP_CLASS:      case OP_CLASS:
1829      case OP_NCLASS:      case OP_NCLASS:
1830      cc += 33;      cc += PRIV(OP_lengths)[OP_CLASS];
1831    
1832      switch (*cc)      switch (*cc)
1833        {        {
# Line 1752  for (;;) Line 1841  for (;;)
1841    
1842        case OP_CRRANGE:        case OP_CRRANGE:
1843        case OP_CRMINRANGE:        case OP_CRMINRANGE:
1844        if (GET2(cc,1) != GET2(cc,3)) return -1;        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;
1845        branchlength += GET2(cc,1);        branchlength += GET2(cc,1);
1846        cc += 5;        cc += 1 + 2 * IMM2_SIZE;
1847        break;        break;
1848    
1849        default:        default:
# Line 1869  length. Line 1958  length.
1958    
1959  Arguments:  Arguments:
1960    code        points to start of expression    code        points to start of expression
1961    utf8        TRUE in UTF-8 mode    utf         TRUE in UTF-8 / UTF-16 mode
1962    number      the required bracket number or negative to find a lookbehind    number      the required bracket number or negative to find a lookbehind
1963    
1964  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
1965  */  */
1966    
1967  const pcre_uchar *  const pcre_uchar *
1968  _pcre_find_bracket(const pcre_uchar *code, BOOL utf8, int number)  PRIV(find_bracket)(const pcre_uchar *code, BOOL utf, int number)
1969  {  {
1970  for (;;)  for (;;)
1971    {    {
# Line 1895  for (;;) Line 1984  for (;;)
1984    else if (c == OP_REVERSE)    else if (c == OP_REVERSE)
1985      {      {
1986      if (number < 0) return (pcre_uchar *)code;      if (number < 0) return (pcre_uchar *)code;
1987      code += _pcre_OP_lengths[c];      code += PRIV(OP_lengths)[c];
1988      }      }
1989    
1990    /* Handle capturing bracket */    /* Handle capturing bracket */
# Line 1905  for (;;) Line 1994  for (;;)
1994      {      {
1995      int n = GET2(code, 1+LINK_SIZE);      int n = GET2(code, 1+LINK_SIZE);
1996      if (n == number) return (pcre_uchar *)code;      if (n == number) return (pcre_uchar *)code;
1997      code += _pcre_OP_lengths[c];      code += PRIV(OP_lengths)[c];
1998      }      }
1999    
2000    /* Otherwise, we can get the item's length from the table, except that for    /* Otherwise, we can get the item's length from the table, except that for
# Line 1933  for (;;) Line 2022  for (;;)
2022        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2023        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2024        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
2025        if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;        if (code[1 + IMM2_SIZE] == OP_PROP
2026            || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;
2027        break;        break;
2028    
2029        case OP_MARK:        case OP_MARK:
# Line 1949  for (;;) Line 2039  for (;;)
2039    
2040      /* Add in the fixed length from the table */      /* Add in the fixed length from the table */
2041    
2042      code += _pcre_OP_lengths[c];      code += PRIV(OP_lengths)[c];
2043    
2044    /* In UTF-8 mode, opcodes that are followed by a character may be followed by    /* In UTF-8 mode, opcodes that are followed by a character may be followed by
2045    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
2046    arrange to skip the extra bytes. */    arrange to skip the extra bytes. */
2047    
2048  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2049      if (utf8) switch(c)      if (utf) switch(c)
2050        {        {
2051        case OP_CHAR:        case OP_CHAR:
2052        case OP_CHARI:        case OP_CHARI:
# Line 1986  for (;;) Line 2076  for (;;)
2076        case OP_MINQUERYI:        case OP_MINQUERYI:
2077        case OP_POSQUERY:        case OP_POSQUERY:
2078        case OP_POSQUERYI:        case OP_POSQUERYI:
2079        if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];        if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]);
2080        break;        break;
2081        }        }
2082  #else  #else
2083      (void)(utf8);  /* Keep compiler happy by referencing function argument */      (void)(utf);  /* Keep compiler happy by referencing function argument */
2084  #endif  #endif
2085      }      }
2086    }    }
# Line 2007  instance of OP_RECURSE. Line 2097  instance of OP_RECURSE.
2097    
2098  Arguments:  Arguments:
2099    code        points to start of expression    code        points to start of expression
2100    utf8        TRUE in UTF-8 mode    utf         TRUE in UTF-8 / UTF-16 mode
2101    
2102  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
2103  */  */
2104    
2105  static const pcre_uchar *  static const pcre_uchar *
2106  find_recurse(const pcre_uchar *code, BOOL utf8)  find_recurse(const pcre_uchar *code, BOOL utf)
2107  {  {
2108  for (;;)  for (;;)
2109    {    {
# Line 2052  for (;;) Line 2142  for (;;)
2142        case OP_TYPEUPTO:        case OP_TYPEUPTO:
2143        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2144        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2145        if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;        if (code[1 + IMM2_SIZE] == OP_PROP
2146            || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;
2147        break;        break;
2148    
2149        case OP_MARK:        case OP_MARK:
# Line 2068  for (;;) Line 2159  for (;;)
2159    
2160      /* Add in the fixed length from the table */      /* Add in the fixed length from the table */
2161    
2162      code += _pcre_OP_lengths[c];      code += PRIV(OP_lengths)[c];
2163    
2164      /* In UTF-8 mode, opcodes that are followed by a character may be followed      /* In UTF-8 mode, opcodes that are followed by a character may be followed
2165      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
2166      to arrange to skip the extra bytes. */      to arrange to skip the extra bytes. */
2167    
2168  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2169      if (utf8) switch(c)      if (utf) switch(c)
2170        {        {
2171        case OP_CHAR:        case OP_CHAR:
2172        case OP_CHARI:        case OP_CHARI:
# Line 2105  for (;;) Line 2196  for (;;)
2196        case OP_MINQUERYI:        case OP_MINQUERYI:
2197        case OP_POSQUERY:        case OP_POSQUERY:
2198        case OP_POSQUERYI:        case OP_POSQUERYI:
2199        if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];        if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]);
2200        break;        break;
2201        }        }
2202  #else  #else
2203      (void)(utf8);  /* Keep compiler happy by referencing function argument */      (void)(utf);  /* Keep compiler happy by referencing function argument */
2204  #endif  #endif
2205      }      }
2206    }    }
# Line 2132  bracket whose current branch will alread Line 2223  bracket whose current branch will alread
2223  Arguments:  Arguments:
2224    code        points to start of search    code        points to start of search
2225    endcode     points to where to stop    endcode     points to where to stop
2226    utf8        TRUE if in UTF8 mode    utf         TRUE if in UTF-8 / UTF-16 mode
2227    cd          contains pointers to tables etc.    cd          contains pointers to tables etc.
2228    
2229  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2140  Returns:      TRUE if what is matched co Line 2231  Returns:      TRUE if what is matched co
2231    
2232  static BOOL  static BOOL
2233  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,
2234    BOOL utf8, compile_data *cd)    BOOL utf, compile_data *cd)
2235  {  {
2236  register int c;  register int c;
2237  for (code = first_significant_code(code + _pcre_OP_lengths[*code], TRUE);  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
2238       code < endcode;       code < endcode;
2239       code = first_significant_code(code + _pcre_OP_lengths[c], TRUE))       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))
2240    {    {
2241    const pcre_uchar *ccode;    const pcre_uchar *ccode;
2242    
# Line 2188  for (code = first_significant_code(code Line 2279  for (code = first_significant_code(code
2279    
2280      do      do
2281        {        {
2282        if (could_be_empty_branch(scode, endcode, utf8, cd))        if (could_be_empty_branch(scode, endcode, utf, cd))
2283          {          {
2284          empty_branch = TRUE;          empty_branch = TRUE;
2285          break;          break;
# Line 2206  for (code = first_significant_code(code Line 2297  for (code = first_significant_code(code
2297    if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO ||    if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO ||
2298        c == OP_BRAPOSZERO)        c == OP_BRAPOSZERO)
2299      {      {
2300      code += _pcre_OP_lengths[c];      code += PRIV(OP_lengths)[c];
2301      do code += GET(code, 1); while (*code == OP_ALT);      do code += GET(code, 1); while (*code == OP_ALT);
2302      c = *code;      c = *code;
2303      continue;      continue;
# Line 2244  for (code = first_significant_code(code Line 2335  for (code = first_significant_code(code
2335        empty_branch = FALSE;        empty_branch = FALSE;
2336        do        do
2337          {          {
2338          if (!empty_branch && could_be_empty_branch(code, endcode, utf8, cd))          if (!empty_branch && could_be_empty_branch(code, endcode, utf, cd))
2339            empty_branch = TRUE;            empty_branch = TRUE;
2340          code += GET(code, 1);          code += GET(code, 1);
2341          }          }
# Line 2262  for (code = first_significant_code(code Line 2353  for (code = first_significant_code(code
2353      {      {
2354      /* Check for quantifiers after a class. XCLASS is used for classes that      /* Check for quantifiers after a class. XCLASS is used for classes that
2355      cannot be represented just by a bit map. This includes negated single      cannot be represented just by a bit map. This includes negated single
2356      high-valued characters. The length in _pcre_OP_lengths[] is zero; the      high-valued characters. The length in PRIV(OP_lengths)[] is zero; the
2357      actual length is stored in the compiled code, so we must update "code"      actual length is stored in the compiled code, so we must update "code"
2358      here. */      here. */
2359    
# Line 2274  for (code = first_significant_code(code Line 2365  for (code = first_significant_code(code
2365    
2366      case OP_CLASS:      case OP_CLASS:
2367      case OP_NCLASS:      case OP_NCLASS:
2368      ccode = code + 33;      ccode = code + PRIV(OP_lengths)[OP_CLASS];
2369    
2370  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2371      CHECK_CLASS_REPEAT:      CHECK_CLASS_REPEAT:
# Line 2349  for (code = first_significant_code(code Line 2440  for (code = first_significant_code(code
2440      case OP_TYPEUPTO:      case OP_TYPEUPTO:
2441      case OP_TYPEMINUPTO:      case OP_TYPEMINUPTO:
2442      case OP_TYPEPOSUPTO:      case OP_TYPEPOSUPTO:
2443      if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;      if (code[1 + IMM2_SIZE] == OP_PROP
2444          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;
2445      break;      break;
2446    
2447      /* End of branch */      /* End of branch */
# Line 2364  for (code = first_significant_code(code Line 2456  for (code = first_significant_code(code
2456      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
2457      MINUPTO, and POSUPTO may be followed by a multibyte character */      MINUPTO, and POSUPTO may be followed by a multibyte character */
2458    
2459  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2460      case OP_STAR:      case OP_STAR:
2461      case OP_STARI:      case OP_STARI:
2462      case OP_MINSTAR:      case OP_MINSTAR:
# Line 2377  for (code = first_significant_code(code Line 2469  for (code = first_significant_code(code
2469      case OP_MINQUERYI:      case OP_MINQUERYI:
2470      case OP_POSQUERY:      case OP_POSQUERY:
2471      case OP_POSQUERYI:      case OP_POSQUERYI:
2472      if (utf8 && code[1] >= 0xc0) code += _pcre_utf8_table4[code[1] & 0x3f];      if (utf && HAS_EXTRALEN(code[1])) code += GET_EXTRALEN(code[1]);
2473      break;      break;
2474    
2475      case OP_UPTO:      case OP_UPTO:
# Line 2386  for (code = first_significant_code(code Line 2478  for (code = first_significant_code(code
2478      case OP_MINUPTOI:      case OP_MINUPTOI:
2479      case OP_POSUPTO:      case OP_POSUPTO:
2480      case OP_POSUPTOI:      case OP_POSUPTOI:
2481      if (utf8 && code[3] >= 0xc0) code += _pcre_utf8_table4[code[3] & 0x3f];      if (utf && HAS_EXTRALEN(code[1 + IMM2_SIZE])) code += GET_EXTRALEN(code[1 + IMM2_SIZE]);
2482      break;      break;
2483  #endif  #endif
2484    
# Line 2430  Arguments: Line 2522  Arguments:
2522    code        points to start of the recursion    code        points to start of the recursion
2523    endcode     points to where to stop (current RECURSE item)    endcode     points to where to stop (current RECURSE item)
2524    bcptr       points to the chain of current (unclosed) branch starts    bcptr       points to the chain of current (unclosed) branch starts
2525    utf8        TRUE if in UTF-8 mode    utf         TRUE if in UTF-8 / UTF-16 mode
2526    cd          pointers to tables etc    cd          pointers to tables etc
2527    
2528  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2438  Returns:      TRUE if what is matched co Line 2530  Returns:      TRUE if what is matched co
2530    
2531  static BOOL  static BOOL
2532  could_be_empty(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty(const pcre_uchar *code, const pcre_uchar *endcode,
2533    branch_chain *bcptr, BOOL utf8, compile_data *cd)    branch_chain *bcptr, BOOL utf, compile_data *cd)
2534  {  {
2535  while (bcptr != NULL && bcptr->current_branch >= code)  while (bcptr != NULL && bcptr->current_branch >= code)
2536    {    {
2537    if (!could_be_empty_branch(bcptr->current_branch, endcode, utf8, cd))    if (!could_be_empty_branch(bcptr->current_branch, endcode, utf, cd))
2538      return FALSE;      return FALSE;
2539    bcptr = bcptr->outer;    bcptr = bcptr->outer;
2540    }    }
# Line 2545  register int yield = 0; Line 2637  register int yield = 0;
2637  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
2638    {    {
2639    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
2640      strncmp((const char *)ptr, pn, len) == 0) return yield;      STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;
2641    pn += posix_name_lengths[yield] + 1;    pn += posix_name_lengths[yield] + 1;
2642    yield++;    yield++;
2643    }    }
# Line 2577  value in the reference (which is a group Line 2669  value in the reference (which is a group
2669  Arguments:  Arguments:
2670    group      points to the start of the group    group      points to the start of the group
2671    adjust     the amount by which the group is to be moved    adjust     the amount by which the group is to be moved
2672    utf8       TRUE in UTF-8 mode    utf        TRUE in UTF-8 / UTF-16 mode
2673    cd         contains pointers to tables etc.    cd         contains pointers to tables etc.
2674    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
2675    
# Line 2585  Returns:     nothing Line 2677  Returns:     nothing
2677  */  */
2678    
2679  static void  static void
2680  adjust_recurse(pcre_uchar *group, int adjust, BOOL utf8, compile_data *cd,  adjust_recurse(pcre_uchar *group, int adjust, BOOL utf, compile_data *cd,
2681    pcre_uchar *save_hwm)    pcre_uchar *save_hwm)
2682  {  {
2683  pcre_uchar *ptr = group;  pcre_uchar *ptr = group;
2684    
2685  while ((ptr = (pcre_uchar *)find_recurse(ptr, utf8)) != NULL)  while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) != NULL)
2686    {    {
2687    int offset;    int offset;
2688    pcre_uchar *hc;    pcre_uchar *hc;
# Line 2750  switch(ptype) Line 2842  switch(ptype)
2842            prop->chartype == ucp_Lt) == negated;            prop->chartype == ucp_Lt) == negated;
2843    
2844    case PT_GC:    case PT_GC:
2845    return (pdata == _pcre_ucp_gentype[prop->chartype]) == negated;    return (pdata == PRIV(ucp_gentype)[prop->chartype]) == negated;
2846    
2847    case PT_PC:    case PT_PC:
2848    return (pdata == prop->chartype) == negated;    return (pdata == prop->chartype) == negated;
# Line 2761  switch(ptype) Line 2853  switch(ptype)
2853    /* These are specials */    /* These are specials */
2854    
2855    case PT_ALNUM:    case PT_ALNUM:
2856    return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2857            _pcre_ucp_gentype[prop->chartype] == ucp_N) == negated;            PRIV(ucp_gentype)[prop->chartype] == ucp_N) == negated;
2858    
2859    case PT_SPACE:    /* Perl space */    case PT_SPACE:    /* Perl space */
2860    return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2861            c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)            c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2862            == negated;            == negated;
2863    
2864    case PT_PXSPACE:  /* POSIX space */    case PT_PXSPACE:  /* POSIX space */
2865    return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2866            c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||            c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2867            c == CHAR_FF || c == CHAR_CR)            c == CHAR_FF || c == CHAR_CR)
2868            == negated;            == negated;
2869    
2870    case PT_WORD:    case PT_WORD:
2871    return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2872            _pcre_ucp_gentype[prop->chartype] == ucp_N ||            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2873            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
2874    }    }
2875  return FALSE;  return FALSE;
# Line 2796  sense to automatically possessify the re Line 2888  sense to automatically possessify the re
2888    
2889  Arguments:  Arguments:
2890    previous      pointer to the repeated opcode    previous      pointer to the repeated opcode
2891    utf8          TRUE in UTF-8 mode    utf           TRUE in UTF-8 / UTF-16 mode
2892    ptr           next character in pattern    ptr           next character in pattern
2893    options       options bits    options       options bits
2894    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
# Line 2805  Returns:        TRUE if possessifying is Line 2897  Returns:        TRUE if possessifying is
2897  */  */
2898    
2899  static BOOL  static BOOL
2900  check_auto_possessive(const pcre_uchar *previous, BOOL utf8,  check_auto_possessive(const pcre_uchar *previous, BOOL utf,
2901    const pcre_uchar *ptr, int options, compile_data *cd)    const pcre_uchar *ptr, int options, compile_data *cd)
2902  {  {
2903  int c, next;  pcre_int32 c, next;
2904  int op_code = *previous++;  int op_code = *previous++;
2905    
2906  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
# Line 2825  if ((options & PCRE_EXTENDED) != 0) Line 2917  if ((options & PCRE_EXTENDED) != 0)
2917          {          {
2918          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2919          ptr++;          ptr++;
2920  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2921          if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;          if (utf) FORWARDCHAR(ptr);
2922  #endif  #endif
2923          }          }
2924        }        }
# Line 2844  if (*ptr == CHAR_BACKSLASH) Line 2936  if (*ptr == CHAR_BACKSLASH)
2936    if (temperrorcode != 0) return FALSE;    if (temperrorcode != 0) return FALSE;
2937    ptr++;    /* Point after the escape sequence */    ptr++;    /* Point after the escape sequence */
2938    }    }
2939    else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)
 else if ((cd->ctypes[*ptr] & ctype_meta) == 0)  
2940    {    {
2941  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2942    if (utf8) { GETCHARINC(next, ptr); } else    if (utf) { GETCHARINC(next, ptr); } else
2943  #endif  #endif
2944    next = *ptr++;    next = *ptr++;
2945    }    }
   
2946  else return FALSE;  else return FALSE;
2947    
2948  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
# Line 2869  if ((options & PCRE_EXTENDED) != 0) Line 2959  if ((options & PCRE_EXTENDED) != 0)
2959          {          {
2960          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2961          ptr++;          ptr++;
2962  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2963          if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;          if (utf) FORWARDCHAR(ptr);
2964  #endif  #endif
2965          }          }
2966        }        }
# Line 2881  if ((options & PCRE_EXTENDED) != 0) Line 2971  if ((options & PCRE_EXTENDED) != 0)
2971  /* If the next thing is itself optional, we have to give up. */  /* If the next thing is itself optional, we have to give up. */
2972    
2973  if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||  if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
2974    strncmp((char *)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)
2975      return FALSE;      return FALSE;
2976    
2977  /* Now compare the next item with the previous opcode. First, handle cases when  /* Now compare the next item with the previous opcode. First, handle cases when
# Line 2909  if (next >= 0) switch(op_code) Line 2999  if (next >= 0) switch(op_code)
2999  #endif  #endif
3000    if (c == next) return FALSE;    if (c == next) return FALSE;
3001  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3002    if (utf8)    if (utf)
3003      {      {
3004      unsigned int othercase;      unsigned int othercase;
3005      if (next < 128) othercase = cd->fcc[next]; else      if (next < 128) othercase = cd->fcc[next]; else
# Line 2934  if (next >= 0) switch(op_code) Line 3024  if (next >= 0) switch(op_code)
3024    case OP_NOTI:    case OP_NOTI:
3025    if ((c = *previous) == next) return TRUE;    if ((c = *previous) == next) return TRUE;
3026  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3027    if (utf8)    if (utf)
3028      {      {
3029      unsigned int othercase;      unsigned int othercase;
3030      if (next < 128) othercase = cd->fcc[next]; else      if (next < 128) othercase = cd->fcc[next]; else
# Line 3143  switch(op_code) Line 3233  switch(op_code)
3233        to the original \d etc. At this point, ptr will point to a zero byte. */        to the original \d etc. At this point, ptr will point to a zero byte. */
3234    
3235        if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||        if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
3236          strncmp((char *)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)
3237            return FALSE;            return FALSE;
3238    
3239        /* Do the property check. */        /* Do the property check. */
# Line 3221  Arguments: Line 3311  Arguments:
3311    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3312    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3313    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3314    firstbyteptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)
3315    reqbyteptr     set to the last literal character required, else < 0    reqcharptr     set to the last literal character required, else < 0
3316    bcptr          points to current branch chain    bcptr          points to current branch chain
3317    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3318    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3235  Returns:         TRUE on success Line 3325  Returns:         TRUE on success
3325    
3326  static BOOL  static BOOL
3327  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3328    const pcre_uchar **ptrptr, int *errorcodeptr, int *firstbyteptr,    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,
3329    int *reqbyteptr, branch_chain *bcptr, int cond_depth, compile_data *cd,    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,
3330    int *lengthptr)    compile_data *cd, int *lengthptr)
3331  {  {
3332  int repeat_type, op_type;  int repeat_type, op_type;
3333  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3334  int bravalue = 0;  int bravalue = 0;
3335  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3336  int firstbyte, reqbyte;  pcre_int32 firstchar, reqchar;
3337  int zeroreqbyte, zerofirstbyte;  pcre_int32 zeroreqchar, zerofirstchar;
3338  int req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3339  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3340  int after_manual_callout = 0;  int after_manual_callout = 0;
3341  int length_prevgroup = 0;  int length_prevgroup = 0;
# Line 3255  pcre_uchar *last_code = code; Line 3345  pcre_uchar *last_code = code;
3345  pcre_uchar *orig_code = code;  pcre_uchar *orig_code = code;
3346  pcre_uchar *tempcode;  pcre_uchar *tempcode;
3347  BOOL inescq = FALSE;  BOOL inescq = FALSE;
3348  BOOL groupsetfirstbyte = FALSE;  BOOL groupsetfirstchar = FALSE;
3349  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
3350  const pcre_uchar *tempptr;  const pcre_uchar *tempptr;
3351  const pcre_uchar *nestptr = NULL;  const pcre_uchar *nestptr = NULL;
3352  pcre_uchar *previous = NULL;  pcre_uchar *previous = NULL;
3353  pcre_uchar *previous_callout = NULL;  pcre_uchar *previous_callout = NULL;
3354  pcre_uchar *save_hwm = NULL;  pcre_uchar *save_hwm = NULL;
3355  pcre_uchar classbits[32];  pcre_uint8 classbits[32];
3356    
3357  /* We can fish out the UTF-8 setting once and for all into a BOOL, but we  /* We can fish out the UTF-8 setting once and for all into a BOOL, but we
3358  must not do this for other options (e.g. PCRE_EXTENDED) because they may change  must not do this for other options (e.g. PCRE_EXTENDED) because they may change
3359  dynamically as we process the pattern. */  dynamically as we process the pattern. */
3360    
3361  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3362  BOOL class_utf8;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
3363  BOOL utf8 = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3364  pcre_uint8 *class_utf8data;  pcre_uchar utf_chars[6];
 pcre_uint8 *class_utf8data_base;  
 pcre_uint8 utf8_char[6];  
3365  #else  #else
3366  BOOL utf8 = FALSE;  BOOL utf = FALSE;
3367    #endif
3368    
3369    /* Helper variables for OP_XCLASS opcode (for characters > 255). */
3370    
3371    #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3372    BOOL xclass;
3373    pcre_uchar *class_uchardata;
3374    pcre_uchar *class_uchardata_base;
3375  #endif  #endif
3376    
3377  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 3289  greedy_non_default = greedy_default ^ 1; Line 3385  greedy_non_default = greedy_default ^ 1;
3385    
3386  /* Initialize no first byte, no required byte. REQ_UNSET means "no char  /* Initialize no first byte, no required byte. REQ_UNSET means "no char
3387  matching encountered yet". It gets changed to REQ_NONE if we hit something that  matching encountered yet". It gets changed to REQ_NONE if we hit something that
3388  matches a non-fixed char first char; reqbyte just remains unset if we never  matches a non-fixed char first char; reqchar just remains unset if we never
3389  find one.  find one.
3390    
3391  When we hit a repeat whose minimum is zero, we may have to adjust these values  When we hit a repeat whose minimum is zero, we may have to adjust these values
3392  to take the zero repeat into account. This is implemented by setting them to  to take the zero repeat into account. This is implemented by setting them to
3393  zerofirstbyte and zeroreqbyte when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3394  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3395    
3396  firstbyte = reqbyte = zerofirstbyte = zeroreqbyte = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;
3397    
3398  /* The variable req_caseopt contains either the REQ_CASELESS value or zero,  /* The variable req_caseopt contains either the REQ_CASELESS value
3399  according to the current setting of the caseless flag. REQ_CASELESS is a bit  or zero, according to the current setting of the caseless flag. The
3400  value > 255. It is added into the firstbyte or reqbyte variables to record the  REQ_CASELESS leaves the lower 28 bit empty. It is added into the
3401  case status of the value. This is used only for ASCII characters. */  firstchar or reqchar variables to record the case status of the
3402    value. This is used only for ASCII characters. */
3403    
3404  req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;  req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS:0;
3405    
3406  /* Switch on next character until the end of the branch */  /* Switch on next character until the end of the branch */
3407    
# Line 3322  for (;; ptr++) Line 3419  for (;; ptr++)
3419    int recno;    int recno;
3420    int refsign;    int refsign;
3421    int skipbytes;    int skipbytes;
3422    int subreqbyte;    int subreqchar;
3423    int subfirstbyte;    int subfirstchar;
3424    int terminator;    int terminator;
3425    int mclength;    int mclength;
3426    int tempbracount;    int tempbracount;
3427    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3428    
3429    /* Get next byte in the pattern */    /* Get next character in the pattern */
3430    
3431    c = *ptr;    c = *ptr;
3432    
# Line 3374  for (;; ptr++) Line 3471  for (;; ptr++)
3471        }        }
3472    
3473      *lengthptr += (int)(code - last_code);      *lengthptr += (int)(code - last_code);
3474      DPRINTF(("length=%d added %d c=%c\n", *lengthptr, (int)(code - last_code),      DPRINTF(("length=%d added %d c=%c (0x%x)\n", *lengthptr,
3475        c));        (int)(code - last_code), c, c));
3476    
3477      /* If "previous" is set and it is not at the start of the work space, move      /* If "previous" is set and it is not at the start of the work space, move
3478      it back to there, in order to avoid filling up the work space. Otherwise,      it back to there, in order to avoid filling up the work space. Otherwise,
# Line 3385  for (;; ptr++) Line 3482  for (;; ptr++)
3482        {        {
3483        if (previous > orig_code)        if (previous > orig_code)
3484          {          {
3485          memmove(orig_code, previous, code - previous);          memmove(orig_code, previous, IN_UCHARS(code - previous));
3486          code -= previous - orig_code;          code -= previous - orig_code;
3487          previous = orig_code;          previous = orig_code;
3488          }          }
# Line 3461  for (;; ptr++) Line 3558  for (;; ptr++)
3558          {          {
3559          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
3560          ptr++;          ptr++;
3561  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3562          if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;          if (utf) FORWARDCHAR(ptr);
3563  #endif  #endif
3564          }          }
3565        if (*ptr != 0) continue;        if (*ptr != 0) continue;
# Line 3486  for (;; ptr++) Line 3583  for (;; ptr++)
3583      case 0:                        /* The branch terminates at string end */      case 0:                        /* The branch terminates at string end */
3584      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3585      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3586      *firstbyteptr = firstbyte;      *firstcharptr = firstchar;
3587      *reqbyteptr = reqbyte;      *reqcharptr = reqchar;
3588      *codeptr = code;      *codeptr = code;
3589      *ptrptr = ptr;      *ptrptr = ptr;
3590      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3511  for (;; ptr++) Line 3608  for (;; ptr++)
3608      previous = NULL;      previous = NULL;
3609      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3610        {        {
3611        if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
3612        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3613        }        }
3614      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3523  for (;; ptr++) Line 3620  for (;; ptr++)
3620      break;      break;
3621    
3622      /* There can never be a first char if '.' is first, whatever happens about      /* There can never be a first char if '.' is first, whatever happens about
3623      repeats. The value of reqbyte doesn't change either. */      repeats. The value of reqchar doesn't change either. */
3624    
3625      case CHAR_DOT:      case CHAR_DOT:
3626      if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
3627      zerofirstbyte = firstbyte;      zerofirstchar = firstchar;
3628      zeroreqbyte = reqbyte;      zeroreqchar = reqchar;
3629      previous = code;      previous = code;
3630      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
3631      break;      break;
# Line 3583  for (;; ptr++) Line 3680  for (;; ptr++)
3680          {          {
3681          if (ptr[1] == CHAR_E)          if (ptr[1] == CHAR_E)
3682            ptr++;            ptr++;
3683          else if (strncmp((const char *)ptr+1,          else if (STRNCMP_UC_C8(ptr + 1, STR_Q STR_BACKSLASH STR_E, 3) == 0)
                           STR_Q STR_BACKSLASH STR_E, 3) == 0)  
3684            ptr += 3;            ptr += 3;
3685          else          else
3686            break;            break;
# Line 3603  for (;; ptr++) Line 3699  for (;; ptr++)
3699          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
3700        {        {
3701        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
3702        if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
3703        zerofirstbyte = firstbyte;        zerofirstchar = firstchar;
3704        break;        break;
3705        }        }
3706    
# Line 3628  for (;; ptr++) Line 3724  for (;; ptr++)
3724    
3725      memset(classbits, 0, 32 * sizeof(pcre_uint8));      memset(classbits, 0, 32 * sizeof(pcre_uint8));
3726    
3727  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3728      class_utf8 = FALSE;                       /* No chars >= 256 */      xclass = FALSE;                           /* No chars >= 256 */
3729      class_utf8data = code + LINK_SIZE + 2;    /* For UTF-8 items */      class_uchardata = code + LINK_SIZE + 2;   /* For UTF-8 items */
3730      class_utf8data_base = class_utf8data;     /* For resetting in pass 1 */      class_uchardata_base = class_uchardata;   /* For resetting in pass 1 */
3731  #endif  #endif
3732    
3733      /* Process characters until ] is reached. By writing this as a "do" it      /* Process characters until ] is reached. By writing this as a "do" it
# Line 3642  for (;; ptr++) Line 3738  for (;; ptr++)
3738        {        {
3739        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
3740    
3741  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3742        if (utf8 && c > 127)        if (utf && HAS_EXTRALEN(c))
3743          {                           /* Braces are required because the */          {                           /* Braces are required because the */
3744          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
3745          }          }
3746    #endif
3747    
3748        /* In the pre-compile phase, accumulate the length of any UTF-8 extra  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3749          /* In the pre-compile phase, accumulate the length of any extra
3750        data and reset the pointer. This is so that very large classes that        data and reset the pointer. This is so that very large classes that
3751        contain a zillion UTF-8 characters no longer overwrite the work space        contain a zillion > 255 characters no longer overwrite the work space
3752        (which is on the stack). */        (which is on the stack). */
3753    
3754        if (lengthptr != NULL)        if (lengthptr != NULL)
3755          {          {
3756          *lengthptr += class_utf8data - class_utf8data_base;          *lengthptr += class_uchardata - class_uchardata_base;
3757          class_utf8data = class_utf8data_base;          class_uchardata = class_uchardata_base;
3758          }          }
   
3759  #endif  #endif
3760    
3761        /* Inside \Q...\E everything is literal except \E */        /* Inside \Q...\E everything is literal except \E */
# Line 3859  for (;; ptr++) Line 3956  for (;; ptr++)
3956              SETBIT(classbits, 0x09); /* VT */              SETBIT(classbits, 0x09); /* VT */
3957              SETBIT(classbits, 0x20); /* SPACE */              SETBIT(classbits, 0x20); /* SPACE */
3958              SETBIT(classbits, 0xa0); /* NSBP */              SETBIT(classbits, 0xa0); /* NSBP */
3959  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3960              if (utf8)              if (utf)
3961                {                {
3962                class_utf8 = TRUE;                xclass = TRUE;
3963                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3964                class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x1680, class_uchardata);
3965                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3966                class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x180e, class_uchardata);
3967                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
3968                class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2000, class_uchardata);
3969                class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x200A, class_uchardata);
3970                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3971                class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x202f, class_uchardata);
3972                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3973                class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x205f, class_uchardata);
3974                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3975                class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x3000, class_uchardata);
3976                }                }
3977  #endif  #endif
3978              continue;              continue;
# Line 3894  for (;; ptr++) Line 3991  for (;; ptr++)
3991                classbits[c] |= x;                classbits[c] |= x;
3992                }                }
3993    
3994  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3995              if (utf8)              if (utf)
3996                {                {
3997                class_utf8 = TRUE;                xclass = TRUE;
3998                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
3999                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);
4000                class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x167f, class_uchardata);
4001                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4002                class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x1681, class_uchardata);
4003                class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x180d, class_uchardata);
4004                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4005                class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x180f, class_uchardata);
4006                class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x1fff, class_uchardata);
4007                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4008                class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x200B, class_uchardata);
4009                class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x202e, class_uchardata);
4010                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4011                class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2030, class_uchardata);
4012                class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x205e, class_uchardata);
4013                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4014                class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2060, class_uchardata);
4015                class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2fff, class_uchardata);
4016                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4017                class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x3001, class_uchardata);
4018                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);
4019                }                }
4020  #endif  #endif
4021              continue;              continue;
# Line 3929  for (;; ptr++) Line 4026  for (;; ptr++)
4026              SETBIT(classbits, 0x0c); /* FF */              SETBIT(classbits, 0x0c); /* FF */
4027              SETBIT(classbits, 0x0d); /* CR */              SETBIT(classbits, 0x0d); /* CR */
4028              SETBIT(classbits, 0x85); /* NEL */              SETBIT(classbits, 0x85); /* NEL */
4029  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4030              if (utf8)              if (utf)
4031                {                {
4032                class_utf8 = TRUE;                xclass = TRUE;
4033                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4034                class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2028, class_uchardata);
4035                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2029, class_uchardata);
4036                }                }
4037  #endif  #endif
4038              continue;              continue;
# Line 3957  for (;; ptr++) Line 4054  for (;; ptr++)
4054                classbits[c] |= x;                classbits[c] |= x;
4055                }                }
4056    
4057  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4058              if (utf8)              if (utf)
4059                {                {
4060                class_utf8 = TRUE;                xclass = TRUE;
4061                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4062                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);
4063                class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2027, class_uchardata);
4064                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4065                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2029, class_uchardata);
4066                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);
4067                }                }
4068  #endif  #endif
4069              continue;              continue;
# Line 3979  for (;; ptr++) Line 4076  for (;; ptr++)
4076                int pdata;                int pdata;
4077                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
4078                if (ptype < 0) goto FAILED;                if (ptype < 0) goto FAILED;
4079                class_utf8 = TRUE;                xclass = TRUE;
4080                *class_utf8data++ = ((-c == ESC_p) != negated)?                *class_uchardata++ = ((-c == ESC_p) != negated)?
4081                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4082                *class_utf8data++ = ptype;                *class_uchardata++ = ptype;
4083                *class_utf8data++ = pdata;                *class_uchardata++ = pdata;
4084                class_charcount -= 2;   /* Not a < 256 character */                class_charcount -= 2;   /* Not a < 256 character */
4085                continue;                continue;
4086                }                }
# Line 4005  for (;; ptr++) Line 4102  for (;; ptr++)
4102            }            }
4103    
4104          /* Fall through if we have a single character (c >= 0). This may be          /* Fall through if we have a single character (c >= 0). This may be
4105          greater than 256 in UTF-8 mode. */          greater than 256 mode. */
4106    
4107          }   /* End of backslash handling */          }   /* End of backslash handling */
4108    
# Line 4054  for (;; ptr++) Line 4151  for (;; ptr++)
4151            }            }
4152    
4153  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4154          if (utf8)          if (utf)
4155            {                           /* Braces are required because the */            {                           /* Braces are required because the */
4156            GETCHARLEN(d, ptr, ptr);    /* macro generates multiple statements */            GETCHARLEN(d, ptr, ptr);    /* macro generates multiple statements */
4157            }            }
# Line 4103  for (;; ptr++) Line 4200  for (;; ptr++)
4200          matching for characters > 127 is available only if UCP support is          matching for characters > 127 is available only if UCP support is
4201          available. */          available. */
4202    
4203  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4204          if (utf8 && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))          if (utf && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))
4205    #elif !(defined COMPILE_PCRE8)
4206            if (d > 255)
4207    #endif
4208    #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
4209            {            {
4210            class_utf8 = TRUE;            xclass = TRUE;
4211    
4212            /* With UCP support, we can find the other case equivalents of            /* With UCP support, we can find the other case equivalents of
4213            the relevant characters. There may be several ranges. Optimize how            the relevant characters. There may be several ranges. Optimize how
# Line 4139  for (;; ptr++) Line 4240  for (;; ptr++)
4240    
4241                if (occ == ocd)                if (occ == ocd)
4242                  {                  {
4243                  *class_utf8data++ = XCL_SINGLE;                  *class_uchardata++ = XCL_SINGLE;
4244                  }                  }
4245                else                else
4246                  {                  {
4247                  *class_utf8data++ = XCL_RANGE;                  *class_uchardata++ = XCL_RANGE;
4248                  class_utf8data += _pcre_ord2utf8(occ, class_utf8data);                  class_uchardata += PRIV(ord2utf)(occ, class_uchardata);
4249                  }                  }
4250                class_utf8data += _pcre_ord2utf8(ocd, class_utf8data);                class_uchardata += PRIV(ord2utf)(ocd, class_uchardata);
4251                }                }
4252              }              }
4253  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
# Line 4154  for (;; ptr++) Line 4255  for (;; ptr++)
4255            /* Now record the original range, possibly modified for UCP caseless            /* Now record the original range, possibly modified for UCP caseless
4256            overlapping ranges. */            overlapping ranges. */
4257    
4258            *class_utf8data++ = XCL_RANGE;            *class_uchardata++ = XCL_RANGE;
4259            class_utf8data += _pcre_ord2utf8(c, class_utf8data);  #ifdef SUPPORT_UTF
4260            class_utf8data += _pcre_ord2utf8(d, class_utf8data);            class_uchardata += PRIV(ord2utf)(c, class_uchardata);
4261              class_uchardata += PRIV(ord2utf)(d, class_uchardata);
4262    #else
4263              *class_uchardata++ = c;
4264              *class_uchardata++ = d;
4265    #endif
4266    
4267            /* With UCP support, we are done. Without UCP support, there is no            /* With UCP support, we are done. Without UCP support, there is no
4268            caseless matching for UTF-8 characters > 127; we can use the bit map            caseless matching for UTF characters > 127; we can use the bit map
4269            for the smaller ones. */            for the smaller ones. As for 16 bit characters without UTF, we
4270              can still use  */
4271    
4272  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4273            continue;    /* With next character in the class */            continue;    /* With next character in the class */
4274  #else  #else
4275    #ifdef SUPPORT_UTF
4276            if ((options & PCRE_CASELESS) == 0 || c > 127) continue;            if ((options & PCRE_CASELESS) == 0 || c > 127) continue;
   
4277            /* Adjust upper limit and fall through to set up the map */            /* Adjust upper limit and fall through to set up the map */
   
4278            d = 127;            d = 127;
4279    #else
4280              if (c > 255) continue;
4281              /* Adjust upper limit and fall through to set up the map */
4282              d = 255;
4283    #endif  /* SUPPORT_UTF */
4284  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
4285            }            }
4286  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 || COMPILE_PCRE16 */
4287    
4288          /* We use the bit map for all cases when not in UTF-8 mode; else          /* We use the bit map for 8 bit mode, or when the characters fall
4289          ranges that lie entirely within 0-127 when there is UCP support; else          partially or entirely to [0-255] ([0-127] for UCP) ranges. */
         for partial ranges without UCP support. */  
4290    
4291          class_charcount += d - c + 1;          class_charcount += d - c + 1;
4292          class_lastchar = d;          class_lastchar = d;
# Line 4205  for (;; ptr++) Line 4314  for (;; ptr++)
4314    
4315        /* Handle a character that cannot go in the bit map */        /* Handle a character that cannot go in the bit map */
4316    
4317  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4318        if (utf8 && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))        if (utf && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))
4319          {  #elif !(defined COMPILE_PCRE8)
4320          class_utf8 = TRUE;        if (c > 255)
4321          *class_utf8data++ = XCL_SINGLE;  #endif
4322          class_utf8data += _pcre_ord2utf8(c, class_utf8data);  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
4323            {
4324            xclass = TRUE;
4325            *class_uchardata++ = XCL_SINGLE;
4326    #ifdef SUPPORT_UTF
4327            class_uchardata += PRIV(ord2utf)(c, class_uchardata);
4328    #else
4329            *class_uchardata++ = c;
4330    #endif
4331    
4332  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4333          if ((options & PCRE_CASELESS) != 0)          if ((options & PCRE_CASELESS) != 0)
# Line 4218  for (;; ptr++) Line 4335  for (;; ptr++)
4335            unsigned int othercase;            unsigned int othercase;
4336            if ((othercase = UCD_OTHERCASE(c)) != c)            if ((othercase = UCD_OTHERCASE(c)) != c)
4337              {              {
4338              *class_utf8data++ = XCL_SINGLE;              *class_uchardata++ = XCL_SINGLE;
4339              class_utf8data += _pcre_ord2utf8(othercase, class_utf8data);              class_uchardata += PRIV(ord2utf)(othercase, class_uchardata);
4340              }              }
4341            }            }
4342  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
4343    
4344          }          }
4345        else        else
4346  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF || COMPILE_PCRE16 */
   
4347        /* Handle a single-byte character */        /* Handle a single-byte character */
4348          {          {
4349          classbits[c/8] |= (1 << (c&7));          classbits[c/8] |= (1 << (c&7));
# Line 4239  for (;; ptr++) Line 4355  for (;; ptr++)
4355          class_charcount++;          class_charcount++;
4356          class_lastchar = c;          class_lastchar = c;
4357          }          }
4358    
4359        }        }
4360    
4361      /* Loop until ']' reached. This "while" is the end of the "do" far above.      /* Loop until ']' reached. This "while" is the end of the "do" far above.
# Line 4271  for (;; ptr++) Line 4388  for (;; ptr++)
4388      The optimization throws away the bit map. We turn the item into a      The optimization throws away the bit map. We turn the item into a
4389      1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative.      1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative.
4390      Note that OP_NOT[I] does not support multibyte characters. In the positive      Note that OP_NOT[I] does not support multibyte characters. In the positive
4391      case, it can cause firstbyte to be set. Otherwise, there can be no first      case, it can cause firstchar to be set. Otherwise, there can be no first
4392      char if this item is first, whatever repeat count may follow. In the case      char if this item is first, whatever repeat count may follow. In the case
4393      of reqbyte, save the previous value for reinstating. */      of reqchar, save the previous value for reinstating. */
4394    
4395  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4396      if (class_charcount == 1 && !class_utf8 &&      if (class_charcount == 1 && !xclass &&
4397        (!utf8 || !negate_class || class_lastchar < 128))        (!utf || !negate_class || class_lastchar < 128))
4398  #else  #else
4399      if (class_charcount == 1)      if (class_charcount == 1)
4400  #endif  #endif
4401        {        {
4402        zeroreqbyte = reqbyte;        zeroreqchar = reqchar;
4403    
4404        /* The OP_NOT[I] opcodes work on one-byte characters only. */        /* The OP_NOT[I] opcodes work on one-byte characters only. */
4405    
4406        if (negate_class)        if (negate_class)
4407          {          {
4408          if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
4409          zerofirstbyte = firstbyte;          zerofirstchar = firstchar;
4410          *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
4411          *code++ = class_lastchar;          *code++ = class_lastchar;
4412          break;          break;
# Line 4299  for (;; ptr++) Line 4416  for (;; ptr++)
4416        then we can handle this with the normal one-character code. */        then we can handle this with the normal one-character code. */
4417    
4418  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4419        if (utf8 && class_lastchar > 127)        if (utf && class_lastchar > 127)
4420          mclength = _pcre_ord2utf8(class_lastchar, mcbuffer);          mclength = PRIV(ord2utf)(class_lastchar, mcbuffer);
4421        else        else
4422  #endif  #endif
4423          {          {
# Line 4312  for (;; ptr++) Line 4429  for (;; ptr++)
4429    
4430      /* The general case - not the one-char optimization. If this is the first      /* The general case - not the one-char optimization. If this is the first
4431      thing in the branch, there can be no first char setting, whatever the      thing in the branch, there can be no first char setting, whatever the
4432      repeat count. Any reqbyte setting must remain unchanged after any kind of      repeat count. Any reqchar setting must remain unchanged after any kind of
4433      repeat. */      repeat. */
4434    
4435      if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
4436      zerofirstbyte = firstbyte;      zerofirstchar = firstchar;
4437      zeroreqbyte = reqbyte;      zeroreqchar = reqchar;
4438    
4439      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4440      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 4327  for (;; ptr++) Line 4444  for (;; ptr++)
4444      be listed) there are no characters < 256, we can omit the bitmap in the      be listed) there are no characters < 256, we can omit the bitmap in the
4445      actual compiled code. */      actual compiled code. */
4446    
4447  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4448      if (class_utf8 && (!should_flip_negation || (options & PCRE_UCP) != 0))      if (xclass && (!should_flip_negation || (options & PCRE_UCP) != 0))
4449    #endif
4450    #ifndef COMPILE_PCRE8
4451        if (xclass && !should_flip_negation)
4452    #endif
4453    #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
4454        {        {
4455        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */        *class_uchardata++ = XCL_END;    /* Marks the end of extra data */
4456        *code++ = OP_XCLASS;        *code++ = OP_XCLASS;
4457        code += LINK_SIZE;        code += LINK_SIZE;
4458        *code = negate_class? XCL_NOT : 0;        *code = negate_class? XCL_NOT:0;
4459    
4460        /* If the map is required, move up the extra data to make room for it;        /* If the map is required, move up the extra data to make room for it;
4461        otherwise just move the code pointer to the end of the extra data. */        otherwise just move the code pointer to the end of the extra data. */
# Line 4341  for (;; ptr++) Line 4463  for (;; ptr++)
4463        if (class_charcount > 0)        if (class_charcount > 0)
4464          {          {
4465          *code++ |= XCL_MAP;          *code++ |= XCL_MAP;
4466          memmove(code + 32, code, class_utf8data - code);          memmove(code + (32 / sizeof(pcre_uchar)), code,
4467              IN_UCHARS(class_uchardata - code));
4468          memcpy(code, classbits, 32);          memcpy(code, classbits, 32);
4469          code = class_utf8data + 32;          code = class_uchardata + (32 / sizeof(pcre_uchar));
4470          }          }
4471        else code = class_utf8data;        else code = class_uchardata;
4472    
4473        /* Now fill in the complete length of the item */        /* Now fill in the complete length of the item */
4474    
# Line 4361  for (;; ptr++) Line 4484  for (;; ptr++)
4484      negating it if necessary. */      negating it if necessary. */
4485    
4486      *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;      *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;
4487      if (negate_class)      if (lengthptr == NULL)    /* Save time in the pre-compile phase */
       {  
       if (lengthptr == NULL)    /* Save time in the pre-compile phase */  
         for (c = 0; c < 32; c++) code[c] = ~classbits[c];  
       }  
     else  
4488        {        {
4489          if (negate_class)
4490            for (c = 0; c < 32; c++) classbits[c] = ~classbits[c];
4491        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
4492        }        }
4493      code += 32;      code += 32 / sizeof(pcre_uchar);
4494      break;      break;
4495    
4496    
# Line 4407  for (;; ptr++) Line 4527  for (;; ptr++)
4527    
4528      if (repeat_min == 0)      if (repeat_min == 0)
4529        {        {
4530        firstbyte = zerofirstbyte;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4531        reqbyte = zeroreqbyte;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4532        }        }
4533    
4534      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4450  for (;; ptr++) Line 4570  for (;; ptr++)
4570    
4571      if (*previous == OP_RECURSE)      if (*previous == OP_RECURSE)
4572        {        {
4573        memmove(previous + 1 + LINK_SIZE, previous, 1 + LINK_SIZE);        memmove(previous + 1 + LINK_SIZE, previous, IN_UCHARS(1 + LINK_SIZE));
4574        *previous = OP_ONCE;        *previous = OP_ONCE;
4575        PUT(previous, 1, 2 + 2*LINK_SIZE);        PUT(previous, 1, 2 + 2*LINK_SIZE);
4576        previous[2 + 2*LINK_SIZE] = OP_KET;        previous[2 + 2*LINK_SIZE] = OP_KET;
# Line 4473  for (;; ptr++) Line 4593  for (;; ptr++)
4593    
4594      /* If previous was a character match, abolish the item and generate a      /* If previous was a character match, abolish the item and generate a
4595      repeat item instead. If a char item has a minumum of more than one, ensure      repeat item instead. If a char item has a minumum of more than one, ensure
4596      that it is set in reqbyte - it might not be if a sequence such as x{3} is      that it is set in reqchar - it might not be if a sequence such as x{3} is
4597      the first thing in a branch because the x will have gone into firstbyte      the first thing in a branch because the x will have gone into firstchar
4598      instead.  */      instead.  */
4599    
4600      if (*previous == OP_CHAR || *previous == OP_CHARI)      if (*previous == OP_CHAR || *previous == OP_CHARI)
4601        {        {
4602        op_type = (*previous == OP_CHAR)? 0 : OP_STARI - OP_STAR;        op_type = (*previous == OP_CHAR)? 0 : OP_STARI - OP_STAR;
4603    
4604        /* Deal with UTF-8 characters that take up more than one byte. It's        /* Deal with UTF characters that take up more than one character. It's
4605        easier to write this out separately than try to macrify it. Use c to        easier to write this out separately than try to macrify it. Use c to
4606        hold the length of the character in bytes, plus 0x80 to flag that it's a        hold the length of the character in bytes, plus UTF_LENGTH to flag that
4607        length rather than a small character. */        it's a length rather than a small character. */
4608    
4609  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4610        if (utf8 && (code[-1] & 0x80) != 0)        if (utf && NOT_FIRSTCHAR(code[-1]))
4611          {          {
4612          pcre_uchar *lastchar = code - 1;          pcre_uchar *lastchar = code - 1;
4613          while((*lastchar & 0xc0) == 0x80) lastchar--;          BACKCHAR(lastchar);
4614          c = code - lastchar;            /* Length of UTF-8 character */          c = code - lastchar;            /* Length of UTF-8 character */
4615          memcpy(utf8_char, lastchar, c); /* Save the char */          memcpy(utf_chars, lastchar, IN_UCHARS(c)); /* Save the char */
4616          c |= 0x80;                      /* Flag c as a length */          c |= UTF_LENGTH;                /* Flag c as a length */
4617          }          }
4618        else        else
4619  #endif  #endif /* SUPPORT_UTF */
   
       /* Handle the case of a single byte - either with no UTF8 support, or  
       with UTF-8 disabled, or for a UTF-8 character < 128. */  
4620    
4621          /* Handle the case of a single charater - either with no UTF support, or
4622          with UTF disabled, or for a single character UTF character. */
4623          {          {
4624          c = code[-1];          c = code[-1];
4625          if (repeat_min > 1) reqbyte = c | req_caseopt | cd->req_varyopt;          if (repeat_min > 1) reqchar = c | req_caseopt | cd->req_varyopt;
4626          }          }
4627    
4628        /* 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 4513  for (;; ptr++) Line 4632  for (;; ptr++)
4632    
4633        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4634            repeat_max < 0 &&            repeat_max < 0 &&
4635            check_auto_possessive(previous, utf8, ptr + 1, options, cd))            check_auto_possessive(previous, utf, ptr + 1, options, cd))
4636          {          {
4637          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4638          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4534  for (;; ptr++) Line 4653  for (;; ptr++)
4653        c = previous[1];        c = previous[1];
4654        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4655            repeat_max < 0 &&            repeat_max < 0 &&
4656            check_auto_possessive(previous, utf8, ptr + 1, options, cd))            check_auto_possessive(previous, utf, ptr + 1, options, cd))
4657          {          {
4658          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4659          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4558  for (;; ptr++) Line 4677  for (;; ptr++)
4677    
4678        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4679            repeat_max < 0 &&            repeat_max < 0 &&
4680            check_auto_possessive(previous, utf8, ptr + 1, options, cd))            check_auto_possessive(previous, utf, ptr + 1, options, cd))
4681          {          {
4682          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4683          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4638  for (;; ptr++) Line 4757  for (;; ptr++)
4757          we have to insert the character for the previous code. For a repeated          we have to insert the character for the previous code. For a repeated
4758          Unicode property match, there are two extra bytes that define the          Unicode property match, there are two extra bytes that define the
4759          required property. In UTF-8 mode, long characters have their length in          required property. In UTF-8 mode, long characters have their length in
4760          c, with the 0x80 bit as a flag. */          c, with the UTF_LENGTH bit as a flag. */
4761    
4762          if (repeat_max < 0)          if (repeat_max < 0)
4763            {            {
4764  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4765            if (utf8 && c >= 128)            if (utf && (c & UTF_LENGTH) != 0)
4766              {              {
4767              memcpy(code, utf8_char, c & 7);              memcpy(code, utf_chars, IN_UCHARS(c & 7));
4768              code += c & 7;              code += c & 7;
4769              }              }
4770            else            else
# Line 4667  for (;; ptr++) Line 4786  for (;; ptr++)
4786    
4787          else if (repeat_max != repeat_min)          else if (repeat_max != repeat_min)
4788            {            {
4789  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4790            if (utf8 && c >= 128)            if (utf && (c & UTF_LENGTH) != 0)
4791              {              {
4792              memcpy(code, utf8_char, c & 7);              memcpy(code, utf_chars, IN_UCHARS(c & 7));
4793              code += c & 7;              code += c & 7;
4794              }              }
4795            else            else
# Line 4697  for (;; ptr++) Line 4816  for (;; ptr++)
4816    
4817        /* The character or character type itself comes last in all cases. */        /* The character or character type itself comes last in all cases. */
4818    
4819  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4820        if (utf8 && c >= 128)        if (utf && (c & UTF_LENGTH) != 0)
4821          {          {
4822          memcpy(code, utf8_char, c & 7);          memcpy(code, utf_chars, IN_UCHARS(c & 7));
4823          code += c & 7;          code += c & 7;
4824          }          }
4825        else        else
# Line 4724  for (;; ptr++) Line 4843  for (;; ptr++)
4843    
4844      else if (*previous == OP_CLASS ||      else if (*previous == OP_CLASS ||
4845               *previous == OP_NCLASS ||               *previous == OP_NCLASS ||
4846  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF8 || !defined COMPILE_PCRE8
4847               *previous == OP_XCLASS ||               *previous == OP_XCLASS ||
4848  #endif  #endif
4849               *previous == OP_REF ||               *previous == OP_REF ||
# Line 4827  for (;; ptr++) Line 4946  for (;; ptr++)
4946          if (repeat_max <= 1)    /* Covers 0, 1, and unlimited */          if (repeat_max <= 1)    /* Covers 0, 1, and unlimited */
4947            {            {
4948            *code = OP_END;            *code = OP_END;
4949            adjust_recurse(previous, 1, utf8, cd, save_hwm);            adjust_recurse(previous, 1, utf, cd, save_hwm);
4950            memmove(previous+1, previous, len);            memmove(previous + 1, previous, IN_UCHARS(len));
4951            code++;            code++;
4952            if (repeat_max == 0)            if (repeat_max == 0)
4953              {              {
# Line 4851  for (;; ptr++) Line 4970  for (;; ptr++)
4970            {            {
4971            int offset;            int offset;
4972            *code = OP_END;            *code = OP_END;
4973            adjust_recurse(previous, 2 + LINK_SIZE, utf8, cd, save_hwm);            adjust_recurse(previous, 2 + LINK_SIZE, utf, cd, save_hwm);
4974            memmove(previous + 2 + LINK_SIZE, previous, len);            memmove(previous + 2 + LINK_SIZE, previous, IN_UCHARS(len));
4975            code += 2 + LINK_SIZE;            code += 2 + LINK_SIZE;
4976            *previous++ = OP_BRAZERO + repeat_type;            *previous++ = OP_BRAZERO + repeat_type;
4977            *previous++ = OP_BRA;            *previous++ = OP_BRA;
# Line 4902  for (;; ptr++) Line 5021  for (;; ptr++)
5021    
5022            else            else
5023              {              {
5024              if (groupsetfirstbyte && reqbyte < 0) reqbyte = firstbyte;              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;
5025              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5026                {                {
5027                pcre_uchar *hc;                pcre_uchar *hc;
5028                pcre_uchar *this_hwm = cd->hwm;                pcre_uchar *this_hwm = cd->hwm;
5029                memcpy(code, previous, len);                memcpy(code, previous, IN_UCHARS(len));
5030                for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)                for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)
5031                  {                  {
5032                  PUT(cd->hwm, 0, GET(hc, 0) + len);                  PUT(cd->hwm, 0, GET(hc, 0) + len);
# Line 4974  for (;; ptr++) Line 5093  for (;; ptr++)
5093              PUTINC(code, 0, offset);              PUTINC(code, 0, offset);
5094              }              }
5095    
5096            memcpy(code, previous, len);            memcpy(code, previous, IN_UCHARS(len));
5097            for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)            for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)
5098              {              {
5099              PUT(cd->hwm, 0, GET(hc, 0) + len + ((i != 0)? 2+LINK_SIZE : 1));              PUT(cd->hwm, 0, GET(hc, 0) + len + ((i != 0)? 2+LINK_SIZE : 1));
# Line 5053  for (;; ptr++) Line 5172  for (;; ptr++)
5172              pcre_uchar *scode = bracode;              pcre_uchar *scode = bracode;
5173              do              do
5174                {                {
5175                if (could_be_empty_branch(scode, ketcode, utf8, cd))                if (could_be_empty_branch(scode, ketcode, utf, cd))
5176                  {                  {
5177                  *bracode += OP_SBRA - OP_BRA;                  *bracode += OP_SBRA - OP_BRA;
5178                  break;                  break;
# Line 5076  for (;; ptr++) Line 5195  for (;; ptr++)
5195                {                {
5196                int nlen = (int)(code - bracode);                int nlen = (int)(code - bracode);
5197                *code = OP_END;                *code = OP_END;
5198                adjust_recurse(bracode, 1 + LINK_SIZE, utf8, cd, save_hwm);                adjust_recurse(bracode, 1 + LINK_SIZE, utf, cd, save_hwm);
5199                memmove(bracode + 1+LINK_SIZE, bracode, nlen);                memmove(bracode + 1 + LINK_SIZE, bracode, IN_UCHARS(nlen));
5200                code += 1 + LINK_SIZE;                code += 1 + LINK_SIZE;
5201                nlen += 1 + LINK_SIZE;                nlen += 1 + LINK_SIZE;
5202                *bracode = OP_BRAPOS;                *bracode = OP_BRAPOS;
# Line 5146  for (;; ptr++) Line 5265  for (;; ptr++)
5265        int len;        int len;
5266    
5267        if (*tempcode == OP_TYPEEXACT)        if (*tempcode == OP_TYPEEXACT)
5268          tempcode += _pcre_OP_lengths[*tempcode] +          tempcode += PRIV(OP_lengths)[*tempcode] +
5269            ((tempcode[3] == OP_PROP || tempcode[3] == OP_NOTPROP)? 2 : 0);            ((tempcode[1 + IMM2_SIZE] == OP_PROP
5270              || tempcode[1 + IMM2_SIZE] == OP_NOTPROP)? 2 : 0);
5271    
5272        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)
5273          {          {
5274          tempcode += _pcre_OP_lengths[*tempcode];          tempcode += PRIV(OP_lengths)[*tempcode];
5275  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5276          if (utf8 && tempcode[-1] >= 0xc0)          if (utf && HAS_EXTRALEN(tempcode[-1]))
5277            tempcode += _pcre_utf8_table4[tempcode[-1] & 0x3f];            tempcode += GET_EXTRALEN(tempcode[-1]);
5278  #endif  #endif
5279          }          }
5280    
# Line 5191  for (;; ptr++) Line 5311  for (;; ptr++)
5311    
5312          default:          default:
5313          *code = OP_END;          *code = OP_END;
5314          adjust_recurse(tempcode, 1 + LINK_SIZE, utf8, cd, save_hwm);          adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, save_hwm);
5315          memmove(tempcode + 1+LINK_SIZE, tempcode, len);          memmove(tempcode + 1 + LINK_SIZE, tempcode, IN_UCHARS(len));
5316          code += 1 + LINK_SIZE;          code += 1 + LINK_SIZE;
5317          len += 1 + LINK_SIZE;          len += 1 + LINK_SIZE;
5318          tempcode[0] = OP_ONCE;          tempcode[0] = OP_ONCE;
# Line 5204  for (;; ptr++) Line 5324  for (;; ptr++)
5324        }        }
5325    
5326      /* In all case we no longer have a previous item. We also set the      /* In all case we no longer have a previous item. We also set the
5327      "follows varying string" flag for subsequently encountered reqbytes if      "follows varying string" flag for subsequently encountered reqchars if
5328      it isn't already set and we have just passed a varying length item. */      it isn't already set and we have just passed a varying length item. */
5329    
5330      END_REPEAT:      END_REPEAT:
# Line 5261  for (;; ptr++) Line 5381  for (;; ptr++)
5381        for (i = 0; i < verbcount; i++)        for (i = 0; i < verbcount; i++)
5382          {          {
5383          if (namelen == verbs[i].len &&          if (namelen == verbs[i].len &&
5384              strncmp((char *)name, vn, namelen) == 0)              STRNCMP_UC_C8(name, vn, namelen) == 0)
5385            {            {
5386            /* Check for open captures before ACCEPT and convert it to            /* Check for open captures before ACCEPT and convert it to
5387            ASSERT_ACCEPT if in an assertion. */            ASSERT_ACCEPT if in an assertion. */
# Line 5282  for (;; ptr++) Line 5402  for (;; ptr++)
5402                }                }
5403              *code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;              *code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5404    
5405              /* Do not set firstbyte after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5406              if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
5407              }              }
5408    
5409            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5309  for (;; ptr++) Line 5429  for (;; ptr++)
5429              *code = verbs[i].op_arg;              *code = verbs[i].op_arg;
5430              if (*code++ == OP_THEN_ARG) cd->external_flags |= PCRE_HASTHEN;              if (*code++ == OP_THEN_ARG) cd->external_flags |= PCRE_HASTHEN;
5431              *code++ = arglen;              *code++ = arglen;
5432              memcpy(code, arg, arglen);              memcpy(code, arg, IN_UCHARS(arglen));
5433              code += arglen;              code += arglen;
5434              *code++ = 0;              *code++ = 0;
5435              }              }
# Line 5386  for (;; ptr++) Line 5506  for (;; ptr++)
5506            break;            break;
5507    
5508          /* Most other conditions use OP_CREF (a couple change to OP_RREF          /* Most other conditions use OP_CREF (a couple change to OP_RREF
5509          below), and all need to skip 3 bytes at the start of the group. */          below), and all need to skip 1+IMM2_SIZE bytes at the start of the group. */
5510    
5511          code[1+LINK_SIZE] = OP_CREF;          code[1+LINK_SIZE] = OP_CREF;
5512          skipbytes = 3;          skipbytes = 1+IMM2_SIZE;
5513          refsign = -1;          refsign = -1;
5514    
5515          /* Check for a test for recursion in a named group. */          /* Check for a test for recursion in a named group. */
# Line 5436  for (;; ptr++) Line 5556  for (;; ptr++)
5556          while ((cd->ctypes[*ptr] & ctype_word) != 0)          while ((cd->ctypes[*ptr] & ctype_word) != 0)
5557            {            {
5558            if (recno >= 0)            if (recno >= 0)
5559              recno = ((digitab[*ptr] & ctype_digit) != 0)?              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;
               recno * 10 + *ptr - CHAR_0 : -1;  
5560            ptr++;            ptr++;
5561            }            }
5562          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
# Line 5485  for (;; ptr++) Line 5604  for (;; ptr++)
5604          slot = cd->name_table;          slot = cd->name_table;
5605          for (i = 0; i < cd->names_found; i++)          for (i = 0; i < cd->names_found; i++)
5606            {            {
5607            if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break;            if (STRNCMP_UC_UC(name, slot+IMM2_SIZE, namelen) == 0) break;
5608            slot += cd->name_entry_size;            slot += cd->name_entry_size;
5609            }            }
5610    
# Line 5501  for (;; ptr++) Line 5620  for (;; ptr++)
5620          /* Search the pattern for a forward reference */          /* Search the pattern for a forward reference */
5621    
5622          else if ((i = find_parens(cd, name, namelen,          else if ((i = find_parens(cd, name, namelen,
5623                          (options & PCRE_EXTENDED) != 0, utf8)) > 0)                          (options & PCRE_EXTENDED) != 0, utf)) > 0)
5624            {            {
5625            PUT2(code, 2+LINK_SIZE, i);            PUT2(code, 2+LINK_SIZE, i);
5626            code[1+LINK_SIZE]++;            code[1+LINK_SIZE]++;
# Line 5527  for (;; ptr++) Line 5646  for (;; ptr++)
5646            recno = 0;            recno = 0;
5647            for (i = 1; i < namelen; i++)            for (i = 1; i < namelen; i++)
5648              {              {
5649              if ((digitab[name[i]] & ctype_digit) == 0)              if (!IS_DIGIT(name[i]))
5650                {                {
5651                *errorcodeptr = ERR15;                *errorcodeptr = ERR15;
5652                goto FAILED;                goto FAILED;
# Line 5542  for (;; ptr++) Line 5661  for (;; ptr++)
5661          /* Similarly, check for the (?(DEFINE) "condition", which is always          /* Similarly, check for the (?(DEFINE) "condition", which is always
5662          false. */          false. */
5663    
5664          else if (namelen == 6 && strncmp((char *)name, STRING_DEFINE, 6) == 0)          else if (namelen == 6 && STRNCMP_UC_C8(name, STRING_DEFINE, 6) == 0)
5665            {            {
5666            code[1+LINK_SIZE] = OP_DEF;            code[1+LINK_SIZE] = OP_DEF;
5667            skipbytes = 1;            skipbytes = 1;
# Line 5627  for (;; ptr++) Line 5746  for (;; ptr++)
5746          *code++ = OP_CALLOUT;          *code++ = OP_CALLOUT;
5747            {            {
5748            int n = 0;            int n = 0;
5749            while ((digitab[*(++ptr)] & ctype_digit) != 0)            ptr++;
5750              n = n * 10 + *ptr - CHAR_0;            while(IS_DIGIT(*ptr))
5751                n = n * 10 + *ptr++ - CHAR_0;
5752            if (*ptr != CHAR_RIGHT_PARENTHESIS)            if (*ptr != CHAR_RIGHT_PARENTHESIS)
5753              {              {
5754              *errorcodeptr = ERR39;              *errorcodeptr = ERR39;
# Line 5690  for (;; ptr++) Line 5810  for (;; ptr++)
5810                *errorcodeptr = ERR49;                *errorcodeptr = ERR49;
5811                goto FAILED;                goto FAILED;
5812                }                }
5813              if (namelen + 3 > cd->name_entry_size)              if (namelen + IMM2_SIZE + 1 > cd->name_entry_size)
5814                {                {
5815                cd->name_entry_size = namelen + 3;                cd->name_entry_size = namelen + IMM2_SIZE + 1;
5816                if (namelen > MAX_NAME_SIZE)                if (namelen > MAX_NAME_SIZE)
5817                  {                  {
5818                  *errorcodeptr = ERR48;                  *errorcodeptr = ERR48;
# Line 5721  for (;; ptr++) Line 5841  for (;; ptr++)
5841    
5842              for (i = 0; i < cd->names_found; i++)              for (i = 0; i < cd->names_found; i++)
5843                {                {
5844                int crc = memcmp(name, slot+2, namelen);                int crc = memcmp(name, slot+IMM2_SIZE, IN_UCHARS(namelen));
5845                if (crc == 0)                if (crc == 0)
5846                  {                  {
5847                  if (slot[2+namelen] == 0)                  if (slot[IMM2_SIZE+namelen] == 0)
5848                    {                    {
5849                    if (GET2(slot, 0) != cd->bracount + 1 &&                    if (GET2(slot, 0) != cd->bracount + 1 &&
5850                        (options & PCRE_DUPNAMES) == 0)                        (options & PCRE_DUPNAMES) == 0)
# Line 5745  for (;; ptr++) Line 5865  for (;; ptr++)
5865                if (crc < 0)                if (crc < 0)
5866                  {                  {
5867                  memmove(slot + cd->name_entry_size, slot,                  memmove(slot + cd->name_entry_size, slot,
5868                    (cd->names_found - i) * cd->name_entry_size);                    IN_UCHARS((cd->names_found - i) * cd->name_entry_size));
5869                  break;                  break;
5870                  }                  }
5871    
# Line 5776  for (;; ptr++) Line 5896  for (;; ptr++)
5896                }                }
5897    
5898              PUT2(slot, 0, cd->bracount + 1);              PUT2(slot, 0, cd->bracount + 1);
5899              memcpy(slot + 2, name, namelen);              memcpy(slot + IMM2_SIZE, name, IN_UCHARS(namelen));
5900              slot[2+namelen] = 0;              slot[IMM2_SIZE + namelen] = 0;
5901              }              }
5902            }            }
5903    
# Line 5845  for (;; ptr++) Line 5965  for (;; ptr++)
5965            temp = cd->end_pattern;            temp = cd->end_pattern;
5966            cd->end_pattern = ptr;            cd->end_pattern = ptr;
5967            recno = find_parens(cd, name, namelen,            recno = find_parens(cd, name, namelen,
5968              (options & PCRE_EXTENDED) != 0, utf8);              (options & PCRE_EXTENDED) != 0, utf);
5969            cd->end_pattern = temp;            cd->end_pattern = temp;
5970            if (recno < 0) recno = 0;    /* Forward ref; set dummy number */            if (recno < 0) recno = 0;    /* Forward ref; set dummy number */
5971            }            }
# Line 5860  for (;; ptr++) Line 5980  for (;; ptr++)
5980            slot = cd->name_table;            slot = cd->name_table;
5981            for (i = 0; i < cd->names_found; i++)            for (i = 0; i < cd->names_found; i++)
5982              {              {
5983              if (strncmp((char *)name, (char *)slot+2, namelen) == 0 &&              if (STRNCMP_UC_UC(name, slot+IMM2_SIZE, namelen) == 0 &&
5984                  slot[2+namelen] == 0)                  slot[IMM2_SIZE+namelen] == 0)
5985                break;                break;
5986              slot += cd->name_entry_size;              slot += cd->name_entry_size;
5987              }              }
# Line 5872  for (;; ptr++) Line 5992  for (;; ptr++)
5992              }              }
5993            else if ((recno =                /* Forward back reference */            else if ((recno =                /* Forward back reference */
5994                      find_parens(cd, name, namelen,                      find_parens(cd, name, namelen,
5995                        (options & PCRE_EXTENDED) != 0, utf8)) <= 0)                        (options & PCRE_EXTENDED) != 0, utf)) <= 0)
5996              {              {
5997              *errorcodeptr = ERR15;              *errorcodeptr = ERR15;
5998              goto FAILED;              goto FAILED;
# Line 5911  for (;; ptr++) Line 6031  for (;; ptr++)
6031            if ((refsign = *ptr) == CHAR_PLUS)            if ((refsign = *ptr) == CHAR_PLUS)
6032              {              {
6033              ptr++;              ptr++;
6034              if ((digitab[*ptr] & ctype_digit) == 0)              if (!IS_DIGIT(*ptr))
6035                {                {
6036                *errorcodeptr = ERR63;                *errorcodeptr = ERR63;
6037                goto FAILED;                goto FAILED;
# Line 5919  for (;; ptr++) Line 6039  for (;; ptr++)
6039              }              }
6040            else if (refsign == CHAR_MINUS)            else if (refsign == CHAR_MINUS)
6041              {              {
6042              if ((digitab[ptr[1]] & ctype_digit) == 0)              if (!IS_DIGIT(ptr[1]))
6043                goto OTHER_CHAR_AFTER_QUERY;                goto OTHER_CHAR_AFTER_QUERY;
6044              ptr++;              ptr++;
6045              }              }
6046    
6047            recno = 0;            recno = 0;
6048            while((digitab[*ptr] & ctype_digit) != 0)            while(IS_DIGIT(*ptr))
6049              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6050    
6051            if (*ptr != terminator)            if (*ptr != terminator)
# Line 5976  for (;; ptr++) Line 6096  for (;; ptr++)
6096              {              {
6097              *code = OP_END;              *code = OP_END;
6098              if (recno != 0)              if (recno != 0)
6099                called = _pcre_find_bracket(cd->start_code, utf8, recno);                called = PRIV(find_bracket)(cd->start_code, utf, recno);
6100    
6101              /* Forward reference */              /* Forward reference */
6102    
6103              if (called == NULL)              if (called == NULL)
6104                {                {
6105                if (find_parens(cd, NULL, recno,                if (find_parens(cd, NULL, recno,
6106                      (options & PCRE_EXTENDED) != 0, utf8) < 0)                      (options & PCRE_EXTENDED) != 0, utf) < 0)
6107                  {                  {
6108                  *errorcodeptr = ERR15;                  *errorcodeptr = ERR15;
6109                  goto FAILED;                  goto FAILED;
# Line 6007  for (;; ptr++) Line 6127  for (;; ptr++)
6127              conditional subpatterns will be picked up then. */              conditional subpatterns will be picked up then. */
6128    
6129              else if (GET(called, 1) == 0 && cond_depth <= 0 &&              else if (GET(called, 1) == 0 && cond_depth <= 0 &&
6130                       could_be_empty(called, code, bcptr, utf8, cd))                       could_be_empty(called, code, bcptr, utf, cd))
6131                {                {
6132                *errorcodeptr = ERR40;                *errorcodeptr = ERR40;
6133                goto FAILED;                goto FAILED;
# Line 6023  for (;; ptr++) Line 6143  for (;; ptr++)
6143    
6144          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6145    
6146          if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
6147          continue;          continue;
6148    
6149    
# Line 6080  for (;; ptr++) Line 6200  for (;; ptr++)
6200          both phases.          both phases.
6201    
6202          If we are not at the pattern start, reset the greedy defaults and the          If we are not at the pattern start, reset the greedy defaults and the
6203          case value for firstbyte and reqbyte. */          case value for firstchar and reqchar. */
6204    
6205          if (*ptr == CHAR_RIGHT_PARENTHESIS)          if (*ptr == CHAR_RIGHT_PARENTHESIS)
6206            {            {
# Line 6093  for (;; ptr++) Line 6213  for (;; ptr++)
6213              {              {
6214              greedy_default = ((newoptions & PCRE_UNGREEDY) != 0);              greedy_default = ((newoptions & PCRE_UNGREEDY) != 0);
6215              greedy_non_default = greedy_default ^ 1;              greedy_non_default = greedy_default ^ 1;
6216              req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;              req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS:0;
6217              }              }
6218    
6219            /* Change options at this level, and pass them back for use            /* Change options at this level, and pass them back for use
# Line 6130  for (;; ptr++) Line 6250  for (;; ptr++)
6250        NUMBERED_GROUP:        NUMBERED_GROUP:
6251        cd->bracount += 1;        cd->bracount += 1;
6252        PUT2(code, 1+LINK_SIZE, cd->bracount);        PUT2(code, 1+LINK_SIZE, cd->bracount);
6253        skipbytes = 2;        skipbytes = IMM2_SIZE;
6254        }        }
6255    
6256      /* Process nested bracketed regex. Assertions used not to be repeatable,      /* Process nested bracketed regex. Assertions used not to be repeatable,
# Line 6156  for (;; ptr++) Line 6276  for (;; ptr++)
6276           skipbytes,                       /* Skip over bracket number */           skipbytes,                       /* Skip over bracket number */
6277           cond_depth +           cond_depth +
6278             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6279           &subfirstbyte,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6280           &subreqbyte,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6281           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6282           cd,                              /* Tables block */           cd,                              /* Tables block */
6283           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6208  for (;; ptr++) Line 6328  for (;; ptr++)
6328          }          }
6329    
6330        /* A "normal" conditional group. If there is just one branch, we must not        /* A "normal" conditional group. If there is just one branch, we must not
6331        make use of its firstbyte or reqbyte, because this is equivalent to an        make use of its firstchar or reqchar, because this is equivalent to an
6332        empty second branch. */        empty second branch. */
6333    
6334        else        else
# Line 6218  for (;; ptr++) Line 6338  for (;; ptr++)
6338            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6339            goto FAILED;            goto FAILED;
6340            }            }
6341          if (condcount == 1) subfirstbyte = subreqbyte = REQ_NONE;          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;
6342          }          }
6343        }        }
6344    
# Line 6262  for (;; ptr++) Line 6382  for (;; ptr++)
6382      /* Handle updating of the required and first characters for other types of      /* Handle updating of the required and first characters for other types of
6383      group. Update for normal brackets of all kinds, and conditions with two      group. Update for normal brackets of all kinds, and conditions with two
6384      branches (see code above). If the bracket is followed by a quantifier with      branches (see code above). If the bracket is followed by a quantifier with
6385      zero repeat, we have to back off. Hence the definition of zeroreqbyte and      zero repeat, we have to back off. Hence the definition of zeroreqchar and
6386      zerofirstbyte outside the main loop so that they can be accessed for the      zerofirstchar outside the main loop so that they can be accessed for the
6387      back off. */      back off. */
6388    
6389      zeroreqbyte = reqbyte;      zeroreqchar = reqchar;
6390      zerofirstbyte = firstbyte;      zerofirstchar = firstchar;
6391      groupsetfirstbyte = FALSE;      groupsetfirstchar = FALSE;
6392    
6393      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
6394        {        {
6395        /* If we have not yet set a firstbyte in this branch, take it from the        /* If we have not yet set a firstchar in this branch, take it from the
6396        subpattern, remembering that it was set here so that a repeat of more        subpattern, remembering that it was set here so that a repeat of more
6397        than one can replicate it as reqbyte if necessary. If the subpattern has        than one can replicate it as reqchar if necessary. If the subpattern has
6398        no firstbyte, set "none" for the whole branch. In both cases, a zero        no firstchar, set "none" for the whole branch. In both cases, a zero
6399        repeat forces firstbyte to "none". */        repeat forces firstchar to "none". */
6400    
6401        if (firstbyte == REQ_UNSET)        if (firstchar == REQ_UNSET)
6402          {          {
6403          if (subfirstbyte >= 0)          if (subfirstchar >= 0)
6404            {            {
6405            firstbyte = subfirstbyte;            firstchar = subfirstchar;
6406            groupsetfirstbyte = TRUE;            groupsetfirstchar = TRUE;
6407            }            }
6408          else firstbyte = REQ_NONE;          else firstchar = REQ_NONE;
6409          zerofirstbyte = REQ_NONE;          zerofirstchar = REQ_NONE;
6410          }          }
6411    
6412        /* If firstbyte was previously set, convert the subpattern's firstbyte        /* If firstchar was previously set, convert the subpattern's firstchar
6413        into reqbyte 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
6414        existence beforehand. */        existence beforehand. */
6415    
6416        else if (subfirstbyte >= 0 && subreqbyte < 0)        else if (subfirstchar >= 0 && subreqchar < 0)
6417          subreqbyte = subfirstbyte | tempreqvary;          subreqchar = subfirstchar | tempreqvary;
6418    
6419        /* 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
6420        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6421    
6422        if (subreqbyte >= 0) reqbyte = subreqbyte;        if (subreqchar >= 0) reqchar = subreqchar;
6423        }        }
6424    
6425      /* For a forward assertion, we take the reqbyte, if set. This can be      /* For a forward assertion, we take the reqchar, if set. This can be
6426      helpful if the pattern that follows the assertion doesn't set a different      helpful if the pattern that follows the assertion doesn't set a different
6427      char. For example, it's useful for /(?=abcde).+/. We can't set firstbyte      char. For example, it's useful for /(?=abcde).+/. We can't set firstchar
6428      for an assertion, however because it leads to incorrect effect for patterns      for an assertion, however because it leads to incorrect effect for patterns
6429      such as /(?=a)a.+/ when the "real" "a" would then become a reqbyte instead      such as /(?=a)a.+/ when the "real" "a" would then become a reqchar instead
6430      of a firstbyte. 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
6431      firstbyte, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6432    
6433      else if (bravalue == OP_ASSERT && subreqbyte >= 0) reqbyte = subreqbyte;      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;
6434      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6435    
6436    
# Line 6343  for (;; ptr++) Line 6463  for (;; ptr++)
6463        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6464        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6465    
6466        if (firstbyte == REQ_UNSET && -c > ESC_b && -c < ESC_Z)        if (firstchar == REQ_UNSET && -c > ESC_b && -c < ESC_Z)
6467          firstbyte = REQ_NONE;          firstchar = REQ_NONE;
6468    
6469        /* 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. */
6470    
6471        zerofirstbyte = firstbyte;        zerofirstchar = firstchar;
6472        zeroreqbyte = reqbyte;        zeroreqchar = reqchar;
6473    
6474        /* \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'
6475        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
# Line 6400  for (;; ptr++) Line 6520  for (;; ptr++)
6520          /* Test a signed number in angle brackets or quotes. */          /* Test a signed number in angle brackets or quotes. */
6521    
6522          p = ptr + 2;          p = ptr + 2;
6523          while ((digitab[*p] & ctype_digit) != 0) p++;          while (IS_DIGIT(*p)) p++;
6524          if (*p != terminator)          if (*p != terminator)
6525            {            {
6526            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
# Line 6428  for (;; ptr++) Line 6548  for (;; ptr++)
6548          goto NAMED_REF_OR_RECURSE;          goto NAMED_REF_OR_RECURSE;
6549          }          }
6550    
6551        /* Back references are handled specially; must disable firstbyte if        /* Back references are handled specially; must disable firstchar if
6552        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
6553        ':' later. */        ':' later. */
6554    
# Line 6438  for (;; ptr++) Line 6558  for (;; ptr++)
6558          recno = -c - ESC_REF;          recno = -c - ESC_REF;
6559    
6560          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6561          if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
6562          previous = code;          previous = code;
6563          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6564          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6505  for (;; ptr++) Line 6625  for (;; ptr++)
6625    
6626            {            {
6627            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
6628            *code++ = (!utf8 && c == -ESC_C)? OP_ALLANY : -c;            *code++ = (!utf && c == -ESC_C)? OP_ALLANY : -c;
6629            }            }
6630          }          }
6631        continue;        continue;
# Line 6516  for (;; ptr++) Line 6636  for (;; ptr++)
6636      handle it as a data character. */      handle it as a data character. */
6637    
6638  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
6639      if (utf8 && c > 127)      if (utf && c > 127)
6640        mclength = _pcre_ord2utf8(c, mcbuffer);        mclength = PRIV(ord2utf)(c, mcbuffer);
6641      else      else
6642  #endif  #endif
6643    
# Line 6538  for (;; ptr++) Line 6658  for (;; ptr++)
6658      mclength = 1;      mclength = 1;
6659      mcbuffer[0] = c;      mcbuffer[0] = c;
6660    
6661  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6662      if (utf8 && c >= 0xc0)      if (utf && HAS_EXTRALEN(c))
6663        {        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));
       while ((ptr[1] & 0xc0) == 0x80)  
         mcbuffer[mclength++] = *(++ptr);  
       }  
6664  #endif  #endif
6665    
6666      /* At this point we have the character's bytes in mcbuffer, and the length      /* At this point we have the character's bytes in mcbuffer, and the length
# Line 6561  for (;; ptr++) Line 6678  for (;; ptr++)
6678    
6679      /* Set the first and required bytes appropriately. If no previous first      /* Set the first and required bytes appropriately. If no previous first
6680      byte, set it from this character, but revert to none on a zero repeat.      byte, set it from this character, but revert to none on a zero repeat.
6681      Otherwise, leave the firstbyte value alone, and don't change it on a zero      Otherwise, leave the firstchar value alone, and don't change it on a zero
6682      repeat. */      repeat. */
6683    
6684      if (firstbyte == REQ_UNSET)      if (firstchar == REQ_UNSET)
6685        {        {
6686        zerofirstbyte = REQ_NONE;        zerofirstchar = REQ_NONE;
6687        zeroreqbyte = reqbyte;        zeroreqchar = reqchar;
6688    
6689        /* If the character is more than one byte long, we can set firstbyte        /* If the character is more than one byte long, we can set firstchar
6690        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
6691    
6692        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
6693          {          {
6694          firstbyte = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
6695          if (mclength != 1) reqbyte = code[-1] | cd->req_varyopt;          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;
6696          }          }
6697        else firstbyte = reqbyte = REQ_NONE;        else firstchar = reqchar = REQ_NONE;
6698        }        }
6699    
6700      /* firstbyte was previously set; we can set reqbyte only if the length is      /* firstchar was previously set; we can set reqchar only if the length is
6701      1 or the matching is caseful. */      1 or the matching is caseful. */
6702    
6703      else      else
6704        {        {
6705        zerofirstbyte = firstbyte;        zerofirstchar = firstchar;
6706        zeroreqbyte = reqbyte;        zeroreqchar = reqchar;
6707        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
6708          reqbyte = code[-1] | req_caseopt | cd->req_varyopt;          reqchar = code[-1] | req_caseopt | cd->req_varyopt;
6709        }        }
6710    
6711      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6628  Arguments: Line 6745  Arguments:
6745    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
6746    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
6747    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
6748    firstbyteptr   place to put the first required character, or a negative number    firstcharptr   place to put the first required character, or a negative number
6749    reqbyteptr     place to put the last required character, or a negative number    reqcharptr     place to put the last required character, or a negative number
6750    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
6751    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
6752    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 6641  Returns:         TRUE on success Line 6758  Returns:         TRUE on success
6758  static BOOL  static BOOL
6759  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
6760    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
6761    int cond_depth, int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr,    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,
6762    compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
6763  {  {
6764  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
6765  pcre_uchar *code = *codeptr;  pcre_uchar *code = *codeptr;
# Line 6651  pcre_uchar *start_bracket = code; Line 6768  pcre_uchar *start_bracket = code;
6768  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
6769  open_capitem capitem;  open_capitem capitem;
6770  int capnumber = 0;  int capnumber = 0;
6771  int firstbyte, reqbyte;  pcre_int32 firstchar, reqchar;
6772  int branchfirstbyte, branchreqbyte;  pcre_int32 branchfirstchar, branchreqchar;
6773  int length;  int length;
6774  int orig_bracount;  int orig_bracount;
6775  int max_bracount;  int max_bracount;
# Line 6661  branch_chain bc; Line 6778  branch_chain bc;
6778  bc.outer = bcptr;  bc.outer = bcptr;
6779  bc.current_branch = code;  bc.current_branch = code;
6780    
6781  firstbyte = reqbyte = REQ_UNSET;  firstchar = reqchar = REQ_UNSET;
6782    
6783  /* 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
6784  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 6720  for (;;) Line 6837  for (;;)
6837    /* Now compile the branch; in the pre-compile phase its length gets added    /* Now compile the branch; in the pre-compile phase its length gets added
6838    into the length. */    into the length. */
6839    
6840    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstbyte,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
6841          &branchreqbyte, &bc, cond_depth, cd,          &branchreqchar, &bc, cond_depth, cd,
6842          (lengthptr == NULL)? NULL : &length))          (lengthptr == NULL)? NULL : &length))
6843      {      {
6844      *ptrptr = ptr;      *ptrptr = ptr;
# Line 6737  for (;;) Line 6854  for (;;)
6854    
6855    if (lengthptr == NULL)    if (lengthptr == NULL)
6856      {      {
6857      /* If this is the first branch, the firstbyte and reqbyte values for the      /* If this is the first branch, the firstchar and reqchar values for the
6858      branch become the values for the regex. */      branch become the values for the regex. */
6859    
6860      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
6861        {        {
6862        firstbyte = branchfirstbyte;        firstchar = branchfirstchar;
6863        reqbyte = branchreqbyte;        reqchar = branchreqchar;
6864        }        }
6865    
6866      /* If this is not the first branch, the first char and reqbyte have to      /* If this is not the first branch, the first char and reqchar have to
6867      match the values from all the previous branches, except that if the      match the values from all the previous branches, except that if the
6868      previous value for reqbyte didn't have REQ_VARY set, it can still match,      previous value for reqchar didn't have REQ_VARY set, it can still match,
6869      and we set REQ_VARY for the regex. */      and we set REQ_VARY for the regex. */
6870    
6871      else      else
6872        {        {
6873        /* If we previously had a firstbyte, but it doesn't match the new branch,        /* If we previously had a firstchar, but it doesn't match the new branch,
6874        we have to abandon the firstbyte for the regex, but if there was        we have to abandon the firstchar for the regex, but if there was
6875        previously no reqbyte, it takes on the value of the old firstbyte. */        previously no reqchar, it takes on the value of the old firstchar. */
6876    
6877        if (firstbyte >= 0 && firstbyte != branchfirstbyte)        if (firstchar >= 0 && firstchar != branchfirstchar)
6878          {          {
6879          if (reqbyte < 0) reqbyte = firstbyte;          if (reqchar < 0) reqchar = firstchar;
6880          firstbyte = REQ_NONE;          firstchar = REQ_NONE;
6881          }          }
6882    
6883        /* If we (now or from before) have no firstbyte, a firstbyte from the        /* If we (now or from before) have no firstchar, a firstchar from the
6884        branch becomes a reqbyte if there isn't a branch reqbyte. */        branch becomes a reqchar if there isn't a branch reqchar. */
6885    
6886        if (firstbyte < 0 && branchfirstbyte >= 0 && branchreqbyte < 0)        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)
6887            branchreqbyte = branchfirstbyte;            branchreqchar = branchfirstchar;
6888    
6889        /* Now ensure that the reqbytes match */        /* Now ensure that the reqchars match */
6890    
6891        if ((reqbyte & ~REQ_VARY) != (branchreqbyte & ~REQ_VARY))        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))
6892          reqbyte = REQ_NONE;          reqchar = REQ_NONE;
6893        else reqbyte |= branchreqbyte;   /* To "or" REQ_VARY */        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */
6894        }        }
6895    
6896      /* 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 6843  for (;;) Line 6960  for (;;)
6960        if (cd->open_caps->flag)        if (cd->open_caps->flag)
6961          {          {
6962          memmove(start_bracket + 1 + LINK_SIZE, start_bracket,          memmove(start_bracket + 1 + LINK_SIZE, start_bracket,
6963            code - start_bracket);            IN_UCHARS(code - start_bracket));
6964          *start_bracket = OP_ONCE;          *start_bracket = OP_ONCE;
6965          code += 1 + LINK_SIZE;          code += 1 + LINK_SIZE;
6966          PUT(start_bracket, 1, (int)(code - start_bracket));          PUT(start_bracket, 1, (int)(code - start_bracket));
# Line 6863  for (;;) Line 6980  for (;;)
6980    
6981      *codeptr = code;      *codeptr = code;
6982      *ptrptr = ptr;      *ptrptr = ptr;
6983      *firstbyteptr = firstbyte;      *firstcharptr = firstchar;
6984      *reqbyteptr = reqbyte;      *reqcharptr = reqchar;
6985      if (lengthptr != NULL)      if (lengthptr != NULL)
6986        {        {
6987        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 6950  is_anchored(register const pcre_uchar *c Line 7067  is_anchored(register const pcre_uchar *c
7067  {  {
7068  do {  do {
7069     const pcre_uchar *scode = first_significant_code(     const pcre_uchar *scode = first_significant_code(
7070       code + _pcre_OP_lengths[*code], FALSE);       code + PRIV(OP_lengths)[*code], FALSE);
7071     register int op = *scode;     register int op = *scode;
7072    
7073     /* Non-capturing brackets */     /* Non-capturing brackets */
# Line 7027  is_startline(const pcre_uchar *code, uns Line 7144  is_startline(const pcre_uchar *code, uns
7144  {  {
7145  do {  do {
7146     const pcre_uchar *scode = first_significant_code(     const pcre_uchar *scode = first_significant_code(
7147       code + _pcre_OP_lengths[*code], FALSE);       code + PRIV(OP_lengths)[*code], FALSE);
7148     register int op = *scode;     register int op = *scode;
7149    
7150     /* If we are at the start of a conditional assertion group, *both* the     /* If we are at the start of a conditional assertion group, *both* the
# Line 7038  do { Line 7155  do {
7155     if (op == OP_COND)     if (op == OP_COND)
7156       {       {
7157       scode += 1 + LINK_SIZE;       scode += 1 + LINK_SIZE;
7158       if (*scode == OP_CALLOUT) scode += _pcre_OP_lengths[OP_CALLOUT];       if (*scode == OP_CALLOUT) scode += PRIV(OP_lengths)[OP_CALLOUT];
7159       switch (*scode)       switch (*scode)
7160         {         {
7161         case OP_CREF:         case OP_CREF:
# Line 7131  register int c = -1; Line 7248  register int c = -1;
7248  do {  do {
7249     int d;     int d;
7250     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7251               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? 2:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7252     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7253       TRUE);       TRUE);
7254     register int op = *scode;     register int op = *scode;
# Line 7157  do { Line 7274  do {
7274       break;       break;
7275    
7276       case OP_EXACT:       case OP_EXACT:
7277       scode += 2;       scode += IMM2_SIZE;
7278       /* Fall through */       /* Fall through */
7279    
7280       case OP_CHAR:       case OP_CHAR:
# Line 7170  do { Line 7287  do {
7287       break;       break;
7288    
7289       case OP_EXACTI:       case OP_EXACTI:
7290       scode += 2;       scode += IMM2_SIZE;
7291       /* Fall through */       /* Fall through */
7292    
7293       case OP_CHARI:       case OP_CHARI:
# Line 7213  Returns:        pointer to compiled data Line 7330  Returns:        pointer to compiled data
7330                  with errorptr and erroroffset set                  with errorptr and erroroffset set
7331  */  */
7332    
7333    #ifdef COMPILE_PCRE8
7334  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7335  pcre_compile(const char *pattern, int options, const char **errorptr,  pcre_compile(const char *pattern, int options, const char **errorptr,
7336    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7337    #else
7338    PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7339    pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,
7340      int *erroroffset, const unsigned char *tables)
7341    #endif
7342  {  {
7343    #ifdef COMPILE_PCRE8
7344  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7345    #else
7346    return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7347    #endif
7348  }  }
7349    
7350    
7351    #ifdef COMPILE_PCRE8
7352  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7353  pcre_compile2(const char *pattern, int options, int *errorcodeptr,  pcre_compile2(const char *pattern, int options, int *errorcodeptr,
7354    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7355    #else
7356    PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7357    pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,
7358      const char **errorptr, int *erroroffset, const unsigned char *tables)
7359    #endif
7360  {  {
7361  real_pcre *re;  real_pcre *re;
7362  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7363  int firstbyte, reqbyte, newline;  pcre_int32 firstchar, reqchar;
7364    int newline;
7365  int errorcode = 0;  int errorcode = 0;
7366  int skipatstart = 0;  int skipatstart = 0;
7367  BOOL utf8;  BOOL utf;
7368  size_t size;  size_t size;
7369  pcre_uchar *code;  pcre_uchar *code;
7370  const pcre_uchar *codestart;  const pcre_uchar *codestart;
# Line 7275  if (erroroffset == NULL) Line 7409  if (erroroffset == NULL)
7409    
7410  /* Set up pointers to the individual character tables */  /* Set up pointers to the individual character tables */
7411    
7412  if (tables == NULL) tables = _pcre_default_tables;  if (tables == NULL) tables = PRIV(default_tables);
7413  cd->lcc = tables + lcc_offset;  cd->lcc = tables + lcc_offset;
7414  cd->fcc = tables + fcc_offset;  cd->fcc = tables + fcc_offset;
7415  cd->cbits = tables + cbits_offset;  cd->cbits = tables + cbits_offset;
# Line 7298  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7432  while (ptr[skipatstart] == CHAR_LEFT_PAR
7432    int newnl = 0;    int newnl = 0;
7433    int newbsr = 0;    int newbsr = 0;
7434    
7435    if (strncmp((char *)(ptr+skipatstart+2), STRING_UTF8_RIGHTPAR, 5) == 0)  #ifdef COMPILE_PCRE8
7436      if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 5) == 0)
7437      { skipatstart += 7; options |= PCRE_UTF8; continue; }      { skipatstart += 7; options |= PCRE_UTF8; continue; }
7438    else if (strncmp((char *)(ptr+skipatstart+2), STRING_UCP_RIGHTPAR, 4) == 0)  #endif
7439    #ifdef COMPILE_PCRE16
7440      if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)
7441        { skipatstart += 8; options |= PCRE_UTF16; continue; }
7442    #endif
7443      else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)
7444      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
7445    else if (strncmp((char *)(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)
7446      { skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; }      { skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; }
7447    
7448    if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CR_RIGHTPAR, 3) == 0)
7449      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }
7450    else if (strncmp((char *)(ptr+skipatstart+2), STRING_LF_RIGHTPAR, 3)  == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LF_RIGHTPAR, 3)  == 0)
7451      { skipatstart += 5; newnl = PCRE_NEWLINE_LF; }      { skipatstart += 5; newnl = PCRE_NEWLINE_LF; }
7452    else if (strncmp((char *)(ptr+skipatstart+2), STRING_CRLF_RIGHTPAR, 5)  == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CRLF_RIGHTPAR, 5)  == 0)
7453      { skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; }      { skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; }
7454    else if (strncmp((char *)(ptr+skipatstart+2), STRING_ANY_RIGHTPAR, 4) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_ANY_RIGHTPAR, 4) == 0)
7455      { skipatstart += 6; newnl = PCRE_NEWLINE_ANY; }      { skipatstart += 6; newnl = PCRE_NEWLINE_ANY; }
7456    else if (strncmp((char *)(ptr+skipatstart+2), STRING_ANYCRLF_RIGHTPAR, 8) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_ANYCRLF_RIGHTPAR, 8) == 0)
7457      { skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; }      { skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; }
7458    
7459    else if (strncmp((char *)(ptr+skipatstart+2), STRING_BSR_ANYCRLF_RIGHTPAR, 12) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_BSR_ANYCRLF_RIGHTPAR, 12) == 0)
7460      { skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; }      { skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; }
7461    else if (strncmp((char *)(ptr+skipatstart+2), STRING_BSR_UNICODE_RIGHTPAR, 12) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_BSR_UNICODE_RIGHTPAR, 12) == 0)
7462      { skipatstart += 14; newbsr = PCRE_BSR_UNICODE; }      { skipatstart += 14; newbsr = PCRE_BSR_UNICODE; }
7463    
7464    if (newnl != 0)    if (newnl != 0)
# Line 7328  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7468  while (ptr[skipatstart] == CHAR_LEFT_PAR
7468    else break;    else break;
7469    }    }
7470    
7471  utf8 = (options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
7472    utf = (options & PCRE_UTF8) != 0;
7473    
7474  /* Can't support UTF8 unless PCRE has been compiled to include the code. The  /* Can't support UTF8 unless PCRE has been compiled to include the code. The
7475  return of an error code from _pcre_valid_utf8() is a new feature, introduced in  return of an error code from PRIV(valid_utf)() is a new feature, introduced in
7476  release 8.13. It is passed back from pcre_[dfa_]exec(), but at the moment is  release 8.13. It is passed back from pcre_[dfa_]exec(), but at the moment is
7477  not used here. */  not used here. */
7478    
7479  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
7480  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0 &&  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&
7481       (errorcode = _pcre_valid_utf8((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)
7482    {    {
7483    errorcode = ERR44;    errorcode = ERR44;
7484    goto PCRE_EARLY_ERROR_RETURN2;    goto PCRE_EARLY_ERROR_RETURN2;
7485    }    }
7486  #else  #else
7487  if (utf8)  if (utf)
7488    {    {
7489    errorcode = ERR32;    errorcode = ERR32;
7490    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
# Line 7419  cd->backref_map = 0; Line 7560  cd->backref_map = 0;
7560  /* Reflect pattern for debugging output */  /* Reflect pattern for debugging output */
7561    
7562  DPRINTF(("------------------------------------------------------------------\n"));  DPRINTF(("------------------------------------------------------------------\n"));
7563  DPRINTF(("%s\n", pattern));  #ifdef PCRE_DEBUG
7564    print_puchar(stdout, (PCRE_PUCHAR)pattern);
7565    #endif
7566    DPRINTF(("\n"));
7567    
7568  /* Pretend to compile the pattern while actually just accumulating the length  /* Pretend to compile the pattern while actually just accumulating the length
7569  of memory required. This behaviour is triggered by passing a non-NULL final  of memory required. This behaviour is triggered by passing a non-NULL final
# Line 7436  cd->start_workspace = cworkspace; Line 7580  cd->start_workspace = cworkspace;
7580  cd->start_code = cworkspace;  cd->start_code = cworkspace;
7581  cd->hwm = cworkspace;  cd->hwm = cworkspace;
7582  cd->start_pattern = (const pcre_uchar *)pattern;  cd->start_pattern = (const pcre_uchar *)pattern;
7583  cd->end_pattern = (const pcre_uchar *)(pattern + strlen(pattern));  cd->end_pattern = (const pcre_uchar *)(pattern + STRLEN_UC((const pcre_uchar *)pattern));
7584  cd->req_varyopt = 0;  cd->req_varyopt = 0;
7585  cd->external_options = options;  cd->external_options = options;
7586  cd->external_flags = 0;  cd->external_flags = 0;
# Line 7452  ptr += skipatstart; Line 7596  ptr += skipatstart;
7596  code = cworkspace;  code = cworkspace;
7597  *code = OP_BRA;  *code = OP_BRA;
7598  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
7599    FALSE, 0, 0, &firstbyte, &reqbyte, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);
7600  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
7601    
7602  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
7603    cd->hwm - cworkspace));    (int)(cd->hwm - cworkspace)));
7604    
7605  if (length > MAX_PATTERN_SIZE)  if (length > MAX_PATTERN_SIZE)
7606    {    {
# Line 7469  externally provided function. Integer ov Line 7613  externally provided function. Integer ov
7613  because nowadays we limit the maximum value of cd->names_found and  because nowadays we limit the maximum value of cd->names_found and
7614  cd->name_entry_size. */  cd->name_entry_size. */
7615    
7616  size = length + sizeof(real_pcre) + cd->names_found * (cd->name_entry_size + 3);  size = sizeof(real_pcre) + (length + cd->names_found * cd->name_entry_size) * sizeof(pcre_uchar);
7617  re = (real_pcre *)(pcre_malloc)(size);  re = (real_pcre *)(pcre_malloc)(size);
7618    
7619  if (re == NULL)  if (re == NULL)
# Line 7489  re->size = (int)size; Line 7633  re->size = (int)size;
7633  re->options = cd->external_options;  re->options = cd->external_options;
7634  re->flags = cd->external_flags;  re->flags = cd->external_flags;
7635  re->dummy1 = 0;  re->dummy1 = 0;
7636  re->first_byte = 0;  re->first_char = 0;
7637  re->req_byte = 0;  re->req_char = 0;
7638  re->name_table_offset = sizeof(real_pcre);  re->name_table_offset = sizeof(real_pcre) / sizeof(pcre_uchar);
7639  re->name_entry_size = cd->name_entry_size;  re->name_entry_size = cd->name_entry_size;
7640  re->name_count = cd->names_found;  re->name_count = cd->names_found;
7641  re->ref_count = 0;  re->ref_count = 0;
7642  re->tables = (tables == _pcre_default_tables)? NULL : tables;  re->tables = (tables == PRIV(default_tables))? NULL : tables;
7643  re->nullpad = NULL;  re->nullpad = NULL;
7644    
7645  /* 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
# Line 7526  ptr = (const pcre_uchar *)pattern + skip Line 7670  ptr = (const pcre_uchar *)pattern + skip
7670  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
7671  *code = OP_BRA;  *code = OP_BRA;
7672  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
7673    &firstbyte, &reqbyte, NULL, cd, NULL);    &firstchar, &reqchar, NULL, cd, NULL);
7674  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
7675  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
7676  re->flags = cd->external_flags;  re->flags = cd->external_flags;
7677    
7678  if (cd->had_accept) reqbyte = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */
7679    
7680  /* 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. */
7681    
# Line 7555  while (errorcode == 0 && cd->hwm > cwork Line 7699  while (errorcode == 0 && cd->hwm > cwork
7699    cd->hwm -= LINK_SIZE;    cd->hwm -= LINK_SIZE;
7700    offset = GET(cd->hwm, 0);    offset = GET(cd->hwm, 0);
7701    recno = GET(codestart, offset);    recno = GET(codestart, offset);
7702    groupptr = _pcre_find_bracket(codestart, utf8, recno);    groupptr = PRIV(find_bracket)(codestart, utf, recno);
7703    if (groupptr == NULL) errorcode = ERR53;    if (groupptr == NULL) errorcode = ERR53;
7704      else PUT(((pcre_uchar *)codestart), offset, (int)(groupptr - codestart));      else PUT(((pcre_uchar *)codestart), offset, (int)(groupptr - codestart));
7705    }    }
# Line 7582  if (cd->check_lookbehind) Line 7726  if (cd->check_lookbehind)
7726    of zero, but that is a pathological case, and it does no harm.) When we find    of zero, but that is a pathological case, and it does no harm.) When we find
7727    one, we temporarily terminate the branch it is in while we scan it. */    one, we temporarily terminate the branch it is in while we scan it. */
7728    
7729    for (cc = (pcre_uchar *)_pcre_find_bracket(codestart, utf8, -1);    for (cc = (pcre_uchar *)PRIV(find_bracket)(codestart, utf, -1);
7730         cc != NULL;         cc != NULL;
7731         cc = (pcre_uchar *)_pcre_find_bracket(cc, utf8, -1))         cc = (pcre_uchar *)PRIV(find_bracket)(cc, utf, -1))
7732      {      {
7733      if (GET(cc, 1) == 0)      if (GET(cc, 1) == 0)
7734        {        {
# Line 7637  if ((re->options & PCRE_ANCHORED) == 0) Line 7781  if ((re->options & PCRE_ANCHORED) == 0)
7781      re->options |= PCRE_ANCHORED;      re->options |= PCRE_ANCHORED;
7782    else    else
7783      {      {
7784      if (firstbyte < 0)      if (firstchar < 0)
7785        firstbyte = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, FALSE);
7786      if (firstbyte >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */
7787        {        {
7788        int ch = firstbyte & 255;  #ifdef COMPILE_PCRE8
7789        re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&        re->first_char = firstchar & 0xff;
7790           cd->fcc[ch] == ch)? ch : firstbyte;  #else
7791    #ifdef COMPILE_PCRE16
7792          re->first_char = firstchar & 0xffff;
7793    #endif
7794    #endif
7795          if ((firstchar & REQ_CASELESS) != 0)
7796            {
7797    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
7798            /* We ignore non-ASCII first chars in 8 bit mode. */
7799            if (utf)
7800              {
7801              if (re->first_char < 128)
7802                {
7803                if (cd->fcc[re->first_char] != re->first_char)
7804                  re->flags |= PCRE_FCH_CASELESS;
7805                }
7806              else if (UCD_OTHERCASE(re->first_char) != re->first_char)
7807                re->flags |= PCRE_FCH_CASELESS;
7808              }
7809            else
7810    #endif
7811            if (MAX_255(re->first_char)
7812                && cd->fcc[re->first_char] != re->first_char)
7813              re->flags |= PCRE_FCH_CASELESS;
7814            }
7815    
7816        re->flags |= PCRE_FIRSTSET;        re->flags |= PCRE_FIRSTSET;
7817        }        }
7818      else if (is_startline(codestart, 0, cd->backref_map))      else if (is_startline(codestart, 0, cd->backref_map))
# Line 7655  if ((re->options & PCRE_ANCHORED) == 0) Line 7824  if ((re->options & PCRE_ANCHORED) == 0)
7824  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
7825  bytes. */  bytes. */
7826    
7827  if (reqbyte >= 0 &&  if (reqchar >= 0 &&
7828       ((re->options & PCRE_ANCHORED) == 0 || (reqbyte & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))
7829    {    {
7830    int ch = reqbyte & 255;  #ifdef COMPILE_PCRE8
7831    re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&    re->req_char = reqchar & 0xff;
7832      cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;  #else
7833    #ifdef COMPILE_PCRE16
7834      re->req_char = reqchar & 0xffff;
7835    #endif
7836    #endif
7837      if ((reqchar & REQ_CASELESS) != 0)
7838        {
7839    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
7840        /* We ignore non-ASCII first chars in 8 bit mode. */
7841        if (utf)
7842          {
7843          if (re->req_char < 128)
7844            {
7845            if (cd->fcc[re->req_char] != re->req_char)
7846              re->flags |= PCRE_RCH_CASELESS;
7847            }
7848          else if (UCD_OTHERCASE(re->req_char) != re->req_char)
7849            re->flags |= PCRE_RCH_CASELESS;
7850          }
7851        else
7852    #endif
7853        if (MAX_255(re->req_char) && cd->fcc[re->req_char] != re->req_char)
7854          re->flags |= PCRE_RCH_CASELESS;
7855        }
7856    
7857    re->flags |= PCRE_REQCHSET;    re->flags |= PCRE_REQCHSET;
7858    }    }
7859    
# Line 7675  printf("Options=%08x\n", re->options); Line 7868  printf("Options=%08x\n", re->options);
7868    
7869  if ((re->flags & PCRE_FIRSTSET) != 0)  if ((re->flags & PCRE_FIRSTSET) != 0)
7870    {    {
7871    int ch = re->first_byte & 255;    pcre_uchar ch = re->first_char;
7872    const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?    const char *caseless =
7873      "" : " (caseless)";      ((re->flags & PCRE_FCH_CASELESS) == 0)? "" : " (caseless)";
7874    if (isprint(ch)) printf("First char = %c%s\n", ch, caseless);    if (PRINTABLE(ch)) printf("First char = %c%s\n", ch, caseless);
7875      else printf("First char = \\x%02x%s\n", ch, caseless);      else printf("First char = \\x%02x%s\n", ch, caseless);
7876    }    }
7877    
7878  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
7879    {    {
7880    int ch = re->req_byte & 255;    pcre_uchar ch = re->req_char;
7881    const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?    const char *caseless =
7882      "" : " (caseless)";      ((re->flags & PCRE_RCH_CASELESS) == 0)? "" : " (caseless)";
7883    if (isprint(ch)) printf("Req char = %c%s\n", ch, caseless);    if (PRINTABLE(ch)) printf("Req char = %c%s\n", ch, caseless);
7884      else printf("Req char = \\x%02x%s\n", ch, caseless);      else printf("Req char = \\x%02x%s\n", ch, caseless);
7885    }    }
7886    

Legend:
Removed from v.756  
changed lines
  Added in v.794

  ViewVC Help
Powered by ViewVC 1.1.5