/[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 518 by ph10, Tue May 18 15:47:01 2010 UTC revision 574 by ph10, Sat Nov 20 17:47:27 2010 UTC
# Line 261  static const int posix_class_maps[] = { Line 261  static const int posix_class_maps[] = {
261    cbit_xdigit,-1,          0              /* xdigit */    cbit_xdigit,-1,          0              /* xdigit */
262  };  };
263    
264  /* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class  /* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class
265  substitutes must be in the order of the names, defined above, and there are  substitutes must be in the order of the names, defined above, and there are
266  both positive and negative cases. NULL means no substitute. */  both positive and negative cases. NULL means no substitute. */
267    
268  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 272  static const uschar *substitutes[] = { Line 272  static const uschar *substitutes[] = {
272    (uschar *)"\\P{Xsp}",   /* \S */       /* NOTE: Xsp is Perl space */    (uschar *)"\\P{Xsp}",   /* \S */       /* NOTE: Xsp is Perl space */
273    (uschar *)"\\p{Xsp}",   /* \s */    (uschar *)"\\p{Xsp}",   /* \s */
274    (uschar *)"\\P{Xwd}",   /* \W */    (uschar *)"\\P{Xwd}",   /* \W */
275    (uschar *)"\\p{Xwd}"    /* \w */    (uschar *)"\\p{Xwd}"    /* \w */
276  };  };
277    
278  static const uschar *posix_substitutes[] = {  static const uschar *posix_substitutes[] = {
279    (uschar *)"\\p{L}",     /* alpha */    (uschar *)"\\p{L}",     /* alpha */
280    (uschar *)"\\p{Ll}",    /* lower */    (uschar *)"\\p{Ll}",    /* lower */
281    (uschar *)"\\p{Lu}",    /* upper */    (uschar *)"\\p{Lu}",    /* upper */
282    (uschar *)"\\p{Xan}",   /* alnum */    (uschar *)"\\p{Xan}",   /* alnum */
283    NULL,                   /* ascii */    NULL,                   /* ascii */
284    (uschar *)"\\h",        /* blank */    (uschar *)"\\h",        /* blank */
285    NULL,                   /* cntrl */    NULL,                   /* cntrl */
# Line 289  static const uschar *posix_substitutes[] Line 289  static const uschar *posix_substitutes[]
289    NULL,                   /* punct */    NULL,                   /* punct */
290    (uschar *)"\\p{Xps}",   /* space */    /* NOTE: Xps is POSIX space */    (uschar *)"\\p{Xps}",   /* space */    /* NOTE: Xps is POSIX space */
291    (uschar *)"\\p{Xwd}",   /* word */    (uschar *)"\\p{Xwd}",   /* word */
292    NULL,                   /* xdigit */    NULL,                   /* xdigit */
293    /* Negated cases */    /* Negated cases */
294    (uschar *)"\\P{L}",     /* ^alpha */    (uschar *)"\\P{L}",     /* ^alpha */
295    (uschar *)"\\P{Ll}",    /* ^lower */    (uschar *)"\\P{Ll}",    /* ^lower */
296    (uschar *)"\\P{Lu}",    /* ^upper */    (uschar *)"\\P{Lu}",    /* ^upper */
297    (uschar *)"\\P{Xan}",   /* ^alnum */    (uschar *)"\\P{Xan}",   /* ^alnum */
298    NULL,                   /* ^ascii */    NULL,                   /* ^ascii */
299    (uschar *)"\\H",        /* ^blank */    (uschar *)"\\H",        /* ^blank */
300    NULL,                   /* ^cntrl */    NULL,                   /* ^cntrl */
# Line 304  static const uschar *posix_substitutes[] Line 304  static const uschar *posix_substitutes[]
304    NULL,                   /* ^punct */    NULL,                   /* ^punct */
305    (uschar *)"\\P{Xps}",   /* ^space */   /* NOTE: Xps is POSIX space */    (uschar *)"\\P{Xps}",   /* ^space */   /* NOTE: Xps is POSIX space */
306    (uschar *)"\\P{Xwd}",   /* ^word */    (uschar *)"\\P{Xwd}",   /* ^word */
307    NULL                    /* ^xdigit */    NULL                    /* ^xdigit */
308  };  };
309  #define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *))  #define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *))
310  #endif  #endif
311    
312  #define STRING(a)  # a  #define STRING(a)  # a
313  #define XSTRING(s) STRING(s)  #define XSTRING(s) STRING(s)
# Line 407  static const char error_texts[] = Line 407  static const char error_texts[] =
407    /* 65 */    /* 65 */
408    "different names for subpatterns of the same number are not allowed\0"    "different names for subpatterns of the same number are not allowed\0"
409    "(*MARK) must have an argument\0"    "(*MARK) must have an argument\0"
410    "this version of PCRE is not compiled with PCRE_UCP support\0"    "this version of PCRE is not compiled with PCRE_UCP support\0"
411      "\\c must be followed by an ASCII character\0"
412    ;    ;
413    
414  /* 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 841  else Line 842  else
842      break;      break;
843    
844      /* For \c, a following letter is upper-cased; then the 0x40 bit is flipped.      /* For \c, a following letter is upper-cased; then the 0x40 bit is flipped.
845      This coding is ASCII-specific, but then the whole concept of \cx is      An error is given if the byte following \c is not an ASCII character. This
846        coding is ASCII-specific, but then the whole concept of \cx is
847      ASCII-specific. (However, an EBCDIC equivalent has now been added.) */      ASCII-specific. (However, an EBCDIC equivalent has now been added.) */
848    
849      case CHAR_c:      case CHAR_c:
# Line 851  else Line 853  else
853        *errorcodeptr = ERR2;        *errorcodeptr = ERR2;
854        break;        break;
855        }        }
856    #ifndef EBCDIC    /* ASCII/UTF-8 coding */
857  #ifndef EBCDIC  /* ASCII/UTF-8 coding */      if (c > 127)  /* Excludes all non-ASCII in either mode */
858          {
859          *errorcodeptr = ERR68;
860          break;
861          }
862      if (c >= CHAR_a && c <= CHAR_z) c -= 32;      if (c >= CHAR_a && c <= CHAR_z) c -= 32;
863      c ^= 0x40;      c ^= 0x40;
864  #else           /* EBCDIC coding */  #else             /* EBCDIC coding */
865      if (c >= CHAR_a && c <= CHAR_z) c += 64;      if (c >= CHAR_a && c <= CHAR_z) c += 64;
866      c ^= 0xC0;      c ^= 0xC0;
867  #endif  #endif
# Line 1110  Arguments: Line 1116  Arguments:
1116    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1117    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1118    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1119      utf8         TRUE if we are in UTF-8 mode
1120    count        pointer to the current capturing subpattern number (updated)    count        pointer to the current capturing subpattern number (updated)
1121    
1122  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 1117  Returns:       the number of the named s Line 1124  Returns:       the number of the named s
1124    
1125  static int  static int
1126  find_parens_sub(uschar **ptrptr, compile_data *cd, const uschar *name, int lorn,  find_parens_sub(uschar **ptrptr, compile_data *cd, const uschar *name, int lorn,
1127    BOOL xmode, int *count)    BOOL xmode, BOOL utf8, int *count)
1128  {  {
1129  uschar *ptr = *ptrptr;  uschar *ptr = *ptrptr;
1130  int start_count = *count;  int start_count = *count;
# Line 1129  dealing with. The very first call may no Line 1136  dealing with. The very first call may no
1136    
1137  if (ptr[0] == CHAR_LEFT_PARENTHESIS)  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1138    {    {
1139    if (ptr[1] == CHAR_QUESTION_MARK &&    /* Handle specials such as (*SKIP) or (*UTF8) etc. */
1140        ptr[2] == CHAR_VERTICAL_LINE)  
1141      if (ptr[1] == CHAR_ASTERISK) ptr += 2;
1142    
1143      /* Handle a normal, unnamed capturing parenthesis. */
1144    
1145      else if (ptr[1] != CHAR_QUESTION_MARK)
1146        {
1147        *count += 1;
1148        if (name == NULL && *count == lorn) return *count;
1149        ptr++;
1150        }
1151    
1152      /* All cases now have (? at the start. Remember when we are in a group
1153      where the parenthesis numbers are duplicated. */
1154    
1155      else if (ptr[2] == CHAR_VERTICAL_LINE)
1156      {      {
1157      ptr += 3;      ptr += 3;
1158      dup_parens = TRUE;      dup_parens = TRUE;
1159      }      }
1160    
1161    /* Handle a normal, unnamed capturing parenthesis */    /* Handle comments; all characters are allowed until a ket is reached. */
1162    
1163    else if (ptr[1] != CHAR_QUESTION_MARK && ptr[1] != CHAR_ASTERISK)    else if (ptr[2] == CHAR_NUMBER_SIGN)
1164      {      {
1165      *count += 1;      for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
1166      if (name == NULL && *count == lorn) return *count;      goto FAIL_EXIT;
     ptr++;  
1167      }      }
1168    
1169    /* Handle a condition. If it is an assertion, just carry on so that it    /* Handle a condition. If it is an assertion, just carry on so that it
1170    is processed as normal. If not, skip to the closing parenthesis of the    is processed as normal. If not, skip to the closing parenthesis of the
1171    condition (there can't be any nested parens. */    condition (there can't be any nested parens). */
1172    
1173    else if (ptr[2] == CHAR_LEFT_PARENTHESIS)    else if (ptr[2] == CHAR_LEFT_PARENTHESIS)
1174      {      {
# Line 1159  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1180  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1180        }        }
1181      }      }
1182    
1183    /* We have either (? or (* and not a condition */    /* Start with (? but not a condition. */
1184    
1185    else    else
1186      {      {
# Line 1264  for (; *ptr != 0; ptr++) Line 1285  for (; *ptr != 0; ptr++)
1285    
1286    if (xmode && *ptr == CHAR_NUMBER_SIGN)    if (xmode && *ptr == CHAR_NUMBER_SIGN)
1287      {      {
1288      while (*(++ptr) != 0 && *ptr != CHAR_NL) {};      ptr++;
1289        while (*ptr != 0)
1290          {
1291          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
1292          ptr++;
1293    #ifdef SUPPORT_UTF8
1294          if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
1295    #endif
1296          }
1297      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == 0) goto FAIL_EXIT;
1298      continue;      continue;
1299      }      }
# Line 1273  for (; *ptr != 0; ptr++) Line 1302  for (; *ptr != 0; ptr++)
1302    
1303    if (*ptr == CHAR_LEFT_PARENTHESIS)    if (*ptr == CHAR_LEFT_PARENTHESIS)
1304      {      {
1305      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, count);      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, count);
1306      if (rc > 0) return rc;      if (rc > 0) return rc;
1307      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == 0) goto FAIL_EXIT;
1308      }      }
# Line 1281  for (; *ptr != 0; ptr++) Line 1310  for (; *ptr != 0; ptr++)
1310    else if (*ptr == CHAR_RIGHT_PARENTHESIS)    else if (*ptr == CHAR_RIGHT_PARENTHESIS)
1311      {      {
1312      if (dup_parens && *count < hwm_count) *count = hwm_count;      if (dup_parens && *count < hwm_count) *count = hwm_count;
1313      *ptrptr = ptr;      goto FAIL_EXIT;
     return -1;  
1314      }      }
1315    
1316    else if (*ptr == CHAR_VERTICAL_LINE && dup_parens)    else if (*ptr == CHAR_VERTICAL_LINE && dup_parens)
# Line 1320  Arguments: Line 1348  Arguments:
1348    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1349    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1350    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1351      utf8         TRUE if we are in UTF-8 mode
1352    
1353  Returns:       the number of the found subpattern, or -1 if not found  Returns:       the number of the found subpattern, or -1 if not found
1354  */  */
1355    
1356  static int  static int
1357  find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode)  find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode,
1358      BOOL utf8)
1359  {  {
1360  uschar *ptr = (uschar *)cd->start_pattern;  uschar *ptr = (uschar *)cd->start_pattern;
1361  int count = 0;  int count = 0;
# Line 1338  matching closing parens. That is why we Line 1368  matching closing parens. That is why we
1368    
1369  for (;;)  for (;;)
1370    {    {
1371    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, &count);    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, &count);
1372    if (rc > 0 || *ptr++ == 0) break;    if (rc > 0 || *ptr++ == 0) break;
1373    }    }
1374    
# Line 1711  for (;;) Line 1741  for (;;)
1741        case OP_MARK:        case OP_MARK:
1742        case OP_PRUNE_ARG:        case OP_PRUNE_ARG:
1743        case OP_SKIP_ARG:        case OP_SKIP_ARG:
       case OP_THEN_ARG:  
1744        code += code[1];        code += code[1];
1745        break;        break;
1746    
1747          case OP_THEN_ARG:
1748          code += code[1+LINK_SIZE];
1749          break;
1750        }        }
1751    
1752      /* Add in the fixed length from the table */      /* Add in the fixed length from the table */
# Line 1814  for (;;) Line 1847  for (;;)
1847        case OP_MARK:        case OP_MARK:
1848        case OP_PRUNE_ARG:        case OP_PRUNE_ARG:
1849        case OP_SKIP_ARG:        case OP_SKIP_ARG:
       case OP_THEN_ARG:  
1850        code += code[1];        code += code[1];
1851        break;        break;
1852    
1853          case OP_THEN_ARG:
1854          code += code[1+LINK_SIZE];
1855          break;
1856        }        }
1857    
1858      /* Add in the fixed length from the table */      /* Add in the fixed length from the table */
# Line 2092  for (code = first_significant_code(code Line 2128  for (code = first_significant_code(code
2128      case OP_MARK:      case OP_MARK:
2129      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
2130      case OP_SKIP_ARG:      case OP_SKIP_ARG:
     case OP_THEN_ARG:  
2131      code += code[1];      code += code[1];
2132      break;      break;
2133    
2134        case OP_THEN_ARG:
2135        code += code[1+LINK_SIZE];
2136        break;
2137    
2138      /* None of the remaining opcodes are required to match a character. */      /* None of the remaining opcodes are required to match a character. */
2139    
2140      default:      default:
# Line 2316  auto_callout(uschar *code, const uschar Line 2355  auto_callout(uschar *code, const uschar
2355  {  {
2356  *code++ = OP_CALLOUT;  *code++ = OP_CALLOUT;
2357  *code++ = 255;  *code++ = 255;
2358  PUT(code, 0, ptr - cd->start_pattern);  /* Pattern offset */  PUT(code, 0, (int)(ptr - cd->start_pattern));  /* Pattern offset */
2359  PUT(code, LINK_SIZE, 0);                /* Default length */  PUT(code, LINK_SIZE, 0);                       /* Default length */
2360  return code + 2*LINK_SIZE;  return code + 2*LINK_SIZE;
2361  }  }
2362    
# Line 2342  Returns:             nothing Line 2381  Returns:             nothing
2381  static void  static void
2382  complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd)  complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd)
2383  {  {
2384  int length = ptr - cd->start_pattern - GET(previous_callout, 2);  int length = (int)(ptr - cd->start_pattern - GET(previous_callout, 2));
2385  PUT(previous_callout, 2 + LINK_SIZE, length);  PUT(previous_callout, 2 + LINK_SIZE, length);
2386  }  }
2387    
# Line 2392  for (++c; c <= d; c++) Line 2431  for (++c; c <= d; c++)
2431    
2432  return TRUE;  return TRUE;
2433  }  }
2434    
2435    
2436    
2437    /*************************************************
2438    *        Check a character and a property        *
2439    *************************************************/
2440    
2441    /* This function is called by check_auto_possessive() when a property item
2442    is adjacent to a fixed character.
2443    
2444    Arguments:
2445      c            the character
2446      ptype        the property type
2447      pdata        the data for the type
2448      negated      TRUE if it's a negated property (\P or \p{^)
2449    
2450    Returns:       TRUE if auto-possessifying is OK
2451    */
2452    
2453    static BOOL
2454    check_char_prop(int c, int ptype, int pdata, BOOL negated)
2455    {
2456    const ucd_record *prop = GET_UCD(c);
2457    switch(ptype)
2458      {
2459      case PT_LAMP:
2460      return (prop->chartype == ucp_Lu ||
2461              prop->chartype == ucp_Ll ||
2462              prop->chartype == ucp_Lt) == negated;
2463    
2464      case PT_GC:
2465      return (pdata == _pcre_ucp_gentype[prop->chartype]) == negated;
2466    
2467      case PT_PC:
2468      return (pdata == prop->chartype) == negated;
2469    
2470      case PT_SC:
2471      return (pdata == prop->script) == negated;
2472    
2473      /* These are specials */
2474    
2475      case PT_ALNUM:
2476      return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||
2477              _pcre_ucp_gentype[prop->chartype] == ucp_N) == negated;
2478    
2479      case PT_SPACE:    /* Perl space */
2480      return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
2481              c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2482              == negated;
2483    
2484      case PT_PXSPACE:  /* POSIX space */
2485      return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
2486              c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2487              c == CHAR_FF || c == CHAR_CR)
2488              == negated;
2489    
2490      case PT_WORD:
2491      return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||
2492              _pcre_ucp_gentype[prop->chartype] == ucp_N ||
2493              c == CHAR_UNDERSCORE) == negated;
2494      }
2495    return FALSE;
2496    }
2497  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2498    
2499    
# Line 2405  whether the next thing could possibly ma Line 2507  whether the next thing could possibly ma
2507  sense to automatically possessify the repeated item.  sense to automatically possessify the repeated item.
2508    
2509  Arguments:  Arguments:
2510    op_code       the repeated op code    previous      pointer to the repeated opcode
   this          data for this item, depends on the opcode  
2511    utf8          TRUE in UTF-8 mode    utf8          TRUE in UTF-8 mode
   utf8_char     used for utf8 character bytes, NULL if not relevant  
2512    ptr           next character in pattern    ptr           next character in pattern
2513    options       options bits    options       options bits
2514    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
# Line 2417  Returns:        TRUE if possessifying is Line 2517  Returns:        TRUE if possessifying is
2517  */  */
2518    
2519  static BOOL  static BOOL
2520  check_auto_possessive(int op_code, int item, BOOL utf8, uschar *utf8_char,  check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr,
2521    const uschar *ptr, int options, compile_data *cd)    int options, compile_data *cd)
2522  {  {
2523  int next;  int c, next;
2524    int op_code = *previous++;
2525    
2526  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
2527    
# Line 2431  if ((options & PCRE_EXTENDED) != 0) Line 2532  if ((options & PCRE_EXTENDED) != 0)
2532      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
2533      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
2534        {        {
2535        while (*(++ptr) != 0)        ptr++;
2536          while (*ptr != 0)
2537            {
2538          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2539            ptr++;
2540    #ifdef SUPPORT_UTF8
2541            if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
2542    #endif
2543            }
2544        }        }
2545      else break;      else break;
2546      }      }
# Line 2468  if ((options & PCRE_EXTENDED) != 0) Line 2576  if ((options & PCRE_EXTENDED) != 0)
2576      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
2577      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
2578        {        {
2579        while (*(++ptr) != 0)        ptr++;
2580          while (*ptr != 0)
2581            {
2582          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2583            ptr++;
2584    #ifdef SUPPORT_UTF8
2585            if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
2586    #endif
2587            }
2588        }        }
2589      else break;      else break;
2590      }      }
# Line 2481  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 2596  if (*ptr == CHAR_ASTERISK || *ptr == CHA
2596    strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)    strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
2597      return FALSE;      return FALSE;
2598    
2599  /* Now compare the next item with the previous opcode. If the previous is a  /* Now compare the next item with the previous opcode. First, handle cases when
2600  positive single character match, "item" either contains the character or, if  the next item is a character. */
 "item" is greater than 127 in utf8 mode, the character's bytes are in  
 utf8_char. */  
   
   
 /* Handle cases when the next item is a character. */  
2601    
2602  if (next >= 0) switch(op_code)  if (next >= 0) switch(op_code)
2603    {    {
2604    case OP_CHAR:    case OP_CHAR:
2605  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2606    if (utf8 && item > 127) { GETCHAR(item, utf8_char); }    GETCHARTEST(c, previous);
2607  #else  #else
2608    (void)(utf8_char);  /* Keep compiler happy by referencing function argument */    c = *previous;
2609  #endif  #endif
2610    return item != next;    return c != next;
2611    
2612    /* For CHARNC (caseless character) we must check the other case. If we have    /* For CHARNC (caseless character) we must check the other case. If we have
2613    Unicode property support, we can use it to test the other case of    Unicode property support, we can use it to test the other case of
# Line 2505  if (next >= 0) switch(op_code) Line 2615  if (next >= 0) switch(op_code)
2615    
2616    case OP_CHARNC:    case OP_CHARNC:
2617  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2618    if (utf8 && item > 127) { GETCHAR(item, utf8_char); }    GETCHARTEST(c, previous);
2619    #else
2620      c = *previous;
2621  #endif  #endif
2622    if (item == next) return FALSE;    if (c == next) return FALSE;
2623  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2624    if (utf8)    if (utf8)
2625      {      {
# Line 2518  if (next >= 0) switch(op_code) Line 2630  if (next >= 0) switch(op_code)
2630  #else  #else
2631      othercase = NOTACHAR;      othercase = NOTACHAR;
2632  #endif  #endif
2633      return (unsigned int)item != othercase;      return (unsigned int)c != othercase;
2634      }      }
2635    else    else
2636  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
2637    return (item != cd->fcc[next]);  /* Non-UTF-8 mode */    return (c != cd->fcc[next]);  /* Non-UTF-8 mode */
2638    
2639    /* For OP_NOT, "item" must be a single-byte character. */    /* For OP_NOT, its data is always a single-byte character. */
2640    
2641    case OP_NOT:    case OP_NOT:
2642    if (item == next) return TRUE;    if ((c = *previous) == next) return TRUE;
2643    if ((options & PCRE_CASELESS) == 0) return FALSE;    if ((options & PCRE_CASELESS) == 0) return FALSE;
2644  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2645    if (utf8)    if (utf8)
# Line 2539  if (next >= 0) switch(op_code) Line 2651  if (next >= 0) switch(op_code)
2651  #else  #else
2652      othercase = NOTACHAR;      othercase = NOTACHAR;
2653  #endif  #endif
2654      return (unsigned int)item == othercase;      return (unsigned int)c == othercase;
2655      }      }
2656    else    else
2657  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
2658    return (item == cd->fcc[next]);  /* Non-UTF-8 mode */    return (c == cd->fcc[next]);  /* Non-UTF-8 mode */
2659    
2660      /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
2661      When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
2662    
2663    case OP_DIGIT:    case OP_DIGIT:
2664    return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;    return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;
# Line 2586  if (next >= 0) switch(op_code) Line 2701  if (next >= 0) switch(op_code)
2701      case 0x202f:      case 0x202f:
2702      case 0x205f:      case 0x205f:
2703      case 0x3000:      case 0x3000:
2704      return op_code != OP_HSPACE;      return op_code == OP_NOT_HSPACE;
2705      default:      default:
2706      return op_code == OP_HSPACE;      return op_code != OP_NOT_HSPACE;
2707      }      }
2708    
2709      case OP_ANYNL:
2710    case OP_VSPACE:    case OP_VSPACE:
2711    case OP_NOT_VSPACE:    case OP_NOT_VSPACE:
2712    switch(next)    switch(next)
# Line 2602  if (next >= 0) switch(op_code) Line 2718  if (next >= 0) switch(op_code)
2718      case 0x85:      case 0x85:
2719      case 0x2028:      case 0x2028:
2720      case 0x2029:      case 0x2029:
2721      return op_code != OP_VSPACE;      return op_code == OP_NOT_VSPACE;
2722      default:      default:
2723      return op_code == OP_VSPACE;      return op_code != OP_NOT_VSPACE;
2724      }      }
2725    
2726    #ifdef SUPPORT_UCP
2727      case OP_PROP:
2728      return check_char_prop(next, previous[0], previous[1], FALSE);
2729    
2730      case OP_NOTPROP:
2731      return check_char_prop(next, previous[0], previous[1], TRUE);
2732    #endif
2733    
2734    default:    default:
2735    return FALSE;    return FALSE;
2736    }    }
2737    
2738    
2739  /* Handle the case when the next item is \d, \s, etc. */  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
2740    is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
2741    generated only when PCRE_UCP is *not* set, that is, when only ASCII
2742    characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are
2743    replaced by OP_PROP codes when PCRE_UCP is set. */
2744    
2745  switch(op_code)  switch(op_code)
2746    {    {
2747    case OP_CHAR:    case OP_CHAR:
2748    case OP_CHARNC:    case OP_CHARNC:
2749  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2750    if (utf8 && item > 127) { GETCHAR(item, utf8_char); }    GETCHARTEST(c, previous);
2751    #else
2752      c = *previous;
2753  #endif  #endif
2754    switch(-next)    switch(-next)
2755      {      {
2756      case ESC_d:      case ESC_d:
2757      return item > 127 || (cd->ctypes[item] & ctype_digit) == 0;      return c > 127 || (cd->ctypes[c] & ctype_digit) == 0;
2758    
2759      case ESC_D:      case ESC_D:
2760      return item <= 127 && (cd->ctypes[item] & ctype_digit) != 0;      return c <= 127 && (cd->ctypes[c] & ctype_digit) != 0;
2761    
2762      case ESC_s:      case ESC_s:
2763      return item > 127 || (cd->ctypes[item] & ctype_space) == 0;      return c > 127 || (cd->ctypes[c] & ctype_space) == 0;
2764    
2765      case ESC_S:      case ESC_S:
2766      return item <= 127 && (cd->ctypes[item] & ctype_space) != 0;      return c <= 127 && (cd->ctypes[c] & ctype_space) != 0;
2767    
2768      case ESC_w:      case ESC_w:
2769      return item > 127 || (cd->ctypes[item] & ctype_word) == 0;      return c > 127 || (cd->ctypes[c] & ctype_word) == 0;
2770    
2771      case ESC_W:      case ESC_W:
2772      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;      return c <= 127 && (cd->ctypes[c] & ctype_word) != 0;
2773    
2774      case ESC_h:      case ESC_h:
2775      case ESC_H:      case ESC_H:
2776      switch(item)      switch(c)
2777        {        {
2778        case 0x09:        case 0x09:
2779        case 0x20:        case 0x20:
# Line 2671  switch(op_code) Line 2801  switch(op_code)
2801    
2802      case ESC_v:      case ESC_v:
2803      case ESC_V:      case ESC_V:
2804      switch(item)      switch(c)
2805        {        {
2806        case 0x0a:        case 0x0a:
2807        case 0x0b:        case 0x0b:
# Line 2685  switch(op_code) Line 2815  switch(op_code)
2815        return -next == ESC_v;        return -next == ESC_v;
2816        }        }
2817    
2818        /* When PCRE_UCP is set, these values get generated for \d etc. Find
2819        their substitutions and process them. The result will always be either
2820        -ESC_p or -ESC_P. Then fall through to process those values. */
2821    
2822    #ifdef SUPPORT_UCP
2823        case ESC_du:
2824        case ESC_DU:
2825        case ESC_wu:
2826        case ESC_WU:
2827        case ESC_su:
2828        case ESC_SU:
2829          {
2830          int temperrorcode = 0;
2831          ptr = substitutes[-next - ESC_DU];
2832          next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);
2833          if (temperrorcode != 0) return FALSE;
2834          ptr++;    /* For compatibility */
2835          }
2836        /* Fall through */
2837    
2838        case ESC_p:
2839        case ESC_P:
2840          {
2841          int ptype, pdata, errorcodeptr;
2842          BOOL negated;
2843    
2844          ptr--;      /* Make ptr point at the p or P */
2845          ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);
2846          if (ptype < 0) return FALSE;
2847          ptr++;      /* Point past the final curly ket */
2848    
2849          /* If the property item is optional, we have to give up. (When generated
2850          from \d etc by PCRE_UCP, this test will have been applied much earlier,
2851          to the original \d etc. At this point, ptr will point to a zero byte. */
2852    
2853          if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
2854            strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
2855              return FALSE;
2856    
2857          /* Do the property check. */
2858    
2859          return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);
2860          }
2861    #endif
2862    
2863      default:      default:
2864      return FALSE;      return FALSE;
2865      }      }
2866    
2867      /* In principle, support for Unicode properties should be integrated here as
2868      well. It means re-organizing the above code so as to get hold of the property
2869      values before switching on the op-code. However, I wonder how many patterns
2870      combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set,
2871      these op-codes are never generated.) */
2872    
2873    case OP_DIGIT:    case OP_DIGIT:
2874    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
2875           next == -ESC_h || next == -ESC_v;           next == -ESC_h || next == -ESC_v || next == -ESC_R;
2876    
2877    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
2878    return next == -ESC_d;    return next == -ESC_d;
2879    
2880    case OP_WHITESPACE:    case OP_WHITESPACE:
2881    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return next == -ESC_S || next == -ESC_d || next == -ESC_w || next == -ESC_R;
2882    
2883    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
2884    return next == -ESC_s || next == -ESC_h || next == -ESC_v;    return next == -ESC_s || next == -ESC_h || next == -ESC_v;
2885    
2886    case OP_HSPACE:    case OP_HSPACE:
2887    return next == -ESC_S || next == -ESC_H || next == -ESC_d || next == -ESC_w;    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||
2888             next == -ESC_w || next == -ESC_v || next == -ESC_R;
2889    
2890    case OP_NOT_HSPACE:    case OP_NOT_HSPACE:
2891    return next == -ESC_h;    return next == -ESC_h;
2892    
2893    /* Can't have \S in here because VT matches \S (Perl anomaly) */    /* Can't have \S in here because VT matches \S (Perl anomaly) */
2894      case OP_ANYNL:
2895    case OP_VSPACE:    case OP_VSPACE:
2896    return next == -ESC_V || next == -ESC_d || next == -ESC_w;    return next == -ESC_V || next == -ESC_d || next == -ESC_w;
2897    
2898    case OP_NOT_VSPACE:    case OP_NOT_VSPACE:
2899    return next == -ESC_v;    return next == -ESC_v || next == -ESC_R;
2900    
2901    case OP_WORDCHAR:    case OP_WORDCHAR:
2902    return next == -ESC_W || next == -ESC_s || next == -ESC_h || next == -ESC_v;    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||
2903             next == -ESC_v || next == -ESC_R;
2904    
2905    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2906    return next == -ESC_w || next == -ESC_d;    return next == -ESC_w || next == -ESC_d;
# Line 2851  for (;; ptr++) Line 3035  for (;; ptr++)
3035    
3036    c = *ptr;    c = *ptr;
3037    
3038    /* If we are at the end of a nested substitution, revert to the outer level    /* If we are at the end of a nested substitution, revert to the outer level
3039    string. Nesting only happens one level deep. */    string. Nesting only happens one level deep. */
3040    
3041    if (c == 0 && nestptr != NULL)    if (c == 0 && nestptr != NULL)
# Line 2891  for (;; ptr++) Line 3075  for (;; ptr++)
3075        goto FAILED;        goto FAILED;
3076        }        }
3077    
3078      *lengthptr += code - last_code;      *lengthptr += (int)(code - last_code);
3079      DPRINTF(("length=%d added %d c=%c\n", *lengthptr, code - last_code, c));      DPRINTF(("length=%d added %d c=%c\n", *lengthptr, code - last_code, c));
3080    
3081      /* 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
# Line 2973  for (;; ptr++) Line 3157  for (;; ptr++)
3157      if ((cd->ctypes[c] & ctype_space) != 0) continue;      if ((cd->ctypes[c] & ctype_space) != 0) continue;
3158      if (c == CHAR_NUMBER_SIGN)      if (c == CHAR_NUMBER_SIGN)
3159        {        {
3160        while (*(++ptr) != 0)        ptr++;
3161          while (*ptr != 0)
3162          {          {
3163          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
3164            ptr++;
3165    #ifdef SUPPORT_UTF8
3166            if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
3167    #endif
3168          }          }
3169        if (*ptr != 0) continue;        if (*ptr != 0) continue;
3170    
# Line 3009  for (;; ptr++) Line 3198  for (;; ptr++)
3198          *errorcodeptr = ERR20;          *errorcodeptr = ERR20;
3199          goto FAILED;          goto FAILED;
3200          }          }
3201        *lengthptr += code - last_code;   /* To include callout length */        *lengthptr += (int)(code - last_code);   /* To include callout length */
3202        DPRINTF((">> end branch\n"));        DPRINTF((">> end branch\n"));
3203        }        }
3204      return TRUE;      return TRUE;
# Line 3158  for (;; ptr++) Line 3347  for (;; ptr++)
3347          {                           /* Braces are required because the */          {                           /* Braces are required because the */
3348          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
3349          }          }
3350    
3351        /* In the pre-compile phase, accumulate the length of any UTF-8 extra        /* In the pre-compile phase, accumulate the length of any UTF-8 extra
3352        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
3353        contain a zillion UTF-8 characters no longer overwrite the work space        contain a zillion UTF-8 characters no longer overwrite the work space
# Line 3214  for (;; ptr++) Line 3403  for (;; ptr++)
3403            ptr++;            ptr++;
3404            }            }
3405    
3406          posix_class = check_posix_name(ptr, tempptr - ptr);          posix_class = check_posix_name(ptr, (int)(tempptr - ptr));
3407          if (posix_class < 0)          if (posix_class < 0)
3408            {            {
3409            *errorcodeptr = ERR30;            *errorcodeptr = ERR30;
# Line 3227  for (;; ptr++) Line 3416  for (;; ptr++)
3416    
3417          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
3418            posix_class = 0;            posix_class = 0;
3419    
3420          /* When PCRE_UCP is set, some of the POSIX classes are converted to          /* When PCRE_UCP is set, some of the POSIX classes are converted to
3421          different escape sequences that use Unicode properties. */          different escape sequences that use Unicode properties. */
3422    
3423  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3424          if ((options & PCRE_UCP) != 0)          if ((options & PCRE_UCP) != 0)
3425            {            {
3426            int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);            int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);
3427            if (posix_substitutes[pc] != NULL)            if (posix_substitutes[pc] != NULL)
3428              {              {
3429              nestptr = tempptr + 1;              nestptr = tempptr + 1;
3430              ptr = posix_substitutes[pc] - 1;              ptr = posix_substitutes[pc] - 1;
3431              continue;              continue;
3432              }              }
3433            }            }
3434  #endif  #endif
3435          /* In the non-UCP case, we build the bit map for the POSIX class in a          /* In the non-UCP case, we build the bit map for the POSIX class in a
3436          chunk of local store because we may be adding and subtracting from it,          chunk of local store because we may be adding and subtracting from it,
3437          and we don't want to subtract bits that may be in the main map already.          and we don't want to subtract bits that may be in the main map already.
# Line 3329  for (;; ptr++) Line 3518  for (;; ptr++)
3518              case ESC_SU:              case ESC_SU:
3519              nestptr = ptr;              nestptr = ptr;
3520              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
3521              class_charcount -= 2;                /* Undo! */              class_charcount -= 2;                /* Undo! */
3522              continue;              continue;
3523  #endif  #endif
3524              case ESC_d:              case ESC_d:
# Line 3350  for (;; ptr++) Line 3539  for (;; ptr++)
3539              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];
3540              continue;              continue;
3541    
3542                /* Perl 5.004 onwards omits VT from \s, but we must preserve it
3543                if it was previously set by something earlier in the character
3544                class. */
3545    
3546              case ESC_s:              case ESC_s:
3547              for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space];              classbits[0] |= cbits[cbit_space];
3548              classbits[1] &= ~0x08;   /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= cbits[cbit_space+1] & ~0x08;
3549                for (c = 2; c < 32; c++) classbits[c] |= cbits[c+cbit_space];
3550              continue;              continue;
3551    
3552              case ESC_S:              case ESC_S:
# Line 3780  for (;; ptr++) Line 3974  for (;; ptr++)
3974      can cause firstbyte to be set. Otherwise, there can be no first char if      can cause firstbyte to be set. Otherwise, there can be no first char if
3975      this item is first, whatever repeat count may follow. In the case of      this item is first, whatever repeat count may follow. In the case of
3976      reqbyte, save the previous value for reinstating. */      reqbyte, save the previous value for reinstating. */
3977    
3978  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3979      if (class_charcount == 1 && !class_utf8 &&      if (class_charcount == 1 && !class_utf8 &&
3980        (!utf8 || !negate_class || class_lastchar < 128))        (!utf8 || !negate_class || class_lastchar < 128))
# Line 3860  for (;; ptr++) Line 4054  for (;; ptr++)
4054        }        }
4055  #endif  #endif
4056    
4057      /* If there are no characters > 255, or they are all to be included or      /* If there are no characters > 255, or they are all to be included or
4058      excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the      excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the
4059      whole class was negated and whether there were negative specials such as \S      whole class was negated and whether there were negative specials such as \S
4060      (non-UCP) in the class. Then copy the 32-byte map into the code vector,      (non-UCP) in the class. Then copy the 32-byte map into the code vector,
# Line 3988  for (;; ptr++) Line 4182  for (;; ptr++)
4182    
4183        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4184            repeat_max < 0 &&            repeat_max < 0 &&
4185            check_auto_possessive(*previous, c, utf8, utf8_char, ptr + 1,            check_auto_possessive(previous, utf8, ptr + 1, options, cd))
             options, cd))  
4186          {          {
4187          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4188          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4010  for (;; ptr++) Line 4203  for (;; ptr++)
4203        c = previous[1];        c = previous[1];
4204        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4205            repeat_max < 0 &&            repeat_max < 0 &&
4206            check_auto_possessive(OP_NOT, c, utf8, NULL, ptr + 1, options, cd))            check_auto_possessive(previous, utf8, ptr + 1, options, cd))
4207          {          {
4208          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4209          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4034  for (;; ptr++) Line 4227  for (;; ptr++)
4227    
4228        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4229            repeat_max < 0 &&            repeat_max < 0 &&
4230            check_auto_possessive(c, 0, utf8, NULL, ptr + 1, options, cd))            check_auto_possessive(previous, utf8, ptr + 1, options, cd))
4231          {          {
4232          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4233          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4244  for (;; ptr++) Line 4437  for (;; ptr++)
4437        {        {
4438        register int i;        register int i;
4439        int ketoffset = 0;        int ketoffset = 0;
4440        int len = code - previous;        int len = (int)(code - previous);
4441        uschar *bralink = NULL;        uschar *bralink = NULL;
4442    
4443        /* Repeating a DEFINE group is pointless */        /* Repeating a DEFINE group is pointless */
# Line 4265  for (;; ptr++) Line 4458  for (;; ptr++)
4458          {          {
4459          register uschar *ket = previous;          register uschar *ket = previous;
4460          do ket += GET(ket, 1); while (*ket != OP_KET);          do ket += GET(ket, 1); while (*ket != OP_KET);
4461          ketoffset = code - ket;          ketoffset = (int)(code - ket);
4462          }          }
4463    
4464        /* The case of a zero minimum is special because of the need to stick        /* The case of a zero minimum is special because of the need to stick
# Line 4333  for (;; ptr++) Line 4526  for (;; ptr++)
4526            /* We chain together the bracket offset fields that have to be            /* We chain together the bracket offset fields that have to be
4527            filled in later when the ends of the brackets are reached. */            filled in later when the ends of the brackets are reached. */
4528    
4529            offset = (bralink == NULL)? 0 : previous - bralink;            offset = (bralink == NULL)? 0 : (int)(previous - bralink);
4530            bralink = previous;            bralink = previous;
4531            PUTINC(previous, 0, offset);            PUTINC(previous, 0, offset);
4532            }            }
# Line 4442  for (;; ptr++) Line 4635  for (;; ptr++)
4635              {              {
4636              int offset;              int offset;
4637              *code++ = OP_BRA;              *code++ = OP_BRA;
4638              offset = (bralink == NULL)? 0 : code - bralink;              offset = (bralink == NULL)? 0 : (int)(code - bralink);
4639              bralink = code;              bralink = code;
4640              PUTINC(code, 0, offset);              PUTINC(code, 0, offset);
4641              }              }
# Line 4463  for (;; ptr++) Line 4656  for (;; ptr++)
4656          while (bralink != NULL)          while (bralink != NULL)
4657            {            {
4658            int oldlinkoffset;            int oldlinkoffset;
4659            int offset = code - bralink + 1;            int offset = (int)(code - bralink + 1);
4660            uschar *bra = code - offset;            uschar *bra = code - offset;
4661            oldlinkoffset = GET(bra, 1);            oldlinkoffset = GET(bra, 1);
4662            bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset;            bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset;
# Line 4551  for (;; ptr++) Line 4744  for (;; ptr++)
4744  #endif  #endif
4745          }          }
4746    
4747        len = code - tempcode;        len = (int)(code - tempcode);
4748        if (len > 0) switch (*tempcode)        if (len > 0) switch (*tempcode)
4749          {          {
4750          case OP_STAR:  *tempcode = OP_POSSTAR; break;          case OP_STAR:  *tempcode = OP_POSSTAR; break;
# Line 4620  for (;; ptr++) Line 4813  for (;; ptr++)
4813        const uschar *arg = NULL;        const uschar *arg = NULL;
4814        previous = NULL;        previous = NULL;
4815        while ((cd->ctypes[*++ptr] & ctype_letter) != 0) {};        while ((cd->ctypes[*++ptr] & ctype_letter) != 0) {};
4816        namelen = ptr - name;        namelen = (int)(ptr - name);
4817    
4818        if (*ptr == CHAR_COLON)        if (*ptr == CHAR_COLON)
4819          {          {
4820          arg = ++ptr;          arg = ++ptr;
4821          while ((cd->ctypes[*ptr] & (ctype_letter|ctype_digit)) != 0          while ((cd->ctypes[*ptr] & (ctype_letter|ctype_digit)) != 0
4822            || *ptr == '_') ptr++;            || *ptr == '_') ptr++;
4823          arglen = ptr - arg;          arglen = (int)(ptr - arg);
4824          }          }
4825    
4826        if (*ptr != CHAR_RIGHT_PARENTHESIS)        if (*ptr != CHAR_RIGHT_PARENTHESIS)
# Line 4665  for (;; ptr++) Line 4858  for (;; ptr++)
4858                *errorcodeptr = ERR66;                *errorcodeptr = ERR66;
4859                goto FAILED;                goto FAILED;
4860                }                }
4861              *code++ = verbs[i].op;              *code = verbs[i].op;
4862                if (*code++ == OP_THEN)
4863                  {
4864                  PUT(code, 0, code - bcptr->current_branch - 1);
4865                  code += LINK_SIZE;
4866                  }
4867              }              }
4868    
4869            else            else
# Line 4675  for (;; ptr++) Line 4873  for (;; ptr++)
4873                *errorcodeptr = ERR59;                *errorcodeptr = ERR59;
4874                goto FAILED;                goto FAILED;
4875                }                }
4876              *code++ = verbs[i].op_arg;              *code = verbs[i].op_arg;
4877                if (*code++ == OP_THEN_ARG)
4878                  {
4879                  PUT(code, 0, code - bcptr->current_branch - 1);
4880                  code += LINK_SIZE;
4881                  }
4882              *code++ = arglen;              *code++ = arglen;
4883              memcpy(code, arg, arglen);              memcpy(code, arg, arglen);
4884              code += arglen;              code += arglen;
# Line 4808  for (;; ptr++) Line 5011  for (;; ptr++)
5011                recno * 10 + *ptr - CHAR_0 : -1;                recno * 10 + *ptr - CHAR_0 : -1;
5012            ptr++;            ptr++;
5013            }            }
5014          namelen = ptr - name;          namelen = (int)(ptr - name);
5015    
5016          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != terminator) ||
5017              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
# Line 4869  for (;; ptr++) Line 5072  for (;; ptr++)
5072          /* Search the pattern for a forward reference */          /* Search the pattern for a forward reference */
5073    
5074          else if ((i = find_parens(cd, name, namelen,          else if ((i = find_parens(cd, name, namelen,
5075                          (options & PCRE_EXTENDED) != 0)) > 0)                          (options & PCRE_EXTENDED) != 0, utf8)) > 0)
5076            {            {
5077            PUT2(code, 2+LINK_SIZE, i);            PUT2(code, 2+LINK_SIZE, i);
5078            code[1+LINK_SIZE]++;            code[1+LINK_SIZE]++;
# Line 5004  for (;; ptr++) Line 5207  for (;; ptr++)
5207              goto FAILED;              goto FAILED;
5208              }              }
5209            *code++ = n;            *code++ = n;
5210            PUT(code, 0, ptr - cd->start_pattern + 1);  /* Pattern offset */            PUT(code, 0, (int)(ptr - cd->start_pattern + 1)); /* Pattern offset */
5211            PUT(code, LINK_SIZE, 0);                    /* Default length */            PUT(code, LINK_SIZE, 0);                          /* Default length */
5212            code += 2 * LINK_SIZE;            code += 2 * LINK_SIZE;
5213            }            }
5214          previous = NULL;          previous = NULL;
# Line 5038  for (;; ptr++) Line 5241  for (;; ptr++)
5241            name = ++ptr;            name = ++ptr;
5242    
5243            while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;            while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
5244            namelen = ptr - name;            namelen = (int)(ptr - name);
5245    
5246            /* In the pre-compile phase, just do a syntax check. */            /* In the pre-compile phase, just do a syntax check. */
5247    
# Line 5168  for (;; ptr++) Line 5371  for (;; ptr++)
5371          NAMED_REF_OR_RECURSE:          NAMED_REF_OR_RECURSE:
5372          name = ++ptr;          name = ++ptr;
5373          while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;          while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
5374          namelen = ptr - name;          namelen = (int)(ptr - name);
5375    
5376          /* In the pre-compile phase, do a syntax check and set a dummy          /* In the pre-compile phase, do a syntax check and set a dummy
5377          reference number. */          reference number. */
# Line 5215  for (;; ptr++) Line 5418  for (;; ptr++)
5418              }              }
5419            else if ((recno =                /* Forward back reference */            else if ((recno =                /* Forward back reference */
5420                      find_parens(cd, name, namelen,                      find_parens(cd, name, namelen,
5421                        (options & PCRE_EXTENDED) != 0)) <= 0)                        (options & PCRE_EXTENDED) != 0, utf8)) <= 0)
5422              {              {
5423              *errorcodeptr = ERR15;              *errorcodeptr = ERR15;
5424              goto FAILED;              goto FAILED;
# Line 5326  for (;; ptr++) Line 5529  for (;; ptr++)
5529              if (called == NULL)              if (called == NULL)
5530                {                {
5531                if (find_parens(cd, NULL, recno,                if (find_parens(cd, NULL, recno,
5532                      (options & PCRE_EXTENDED) != 0) < 0)                      (options & PCRE_EXTENDED) != 0, utf8) < 0)
5533                  {                  {
5534                  *errorcodeptr = ERR15;                  *errorcodeptr = ERR15;
5535                  goto FAILED;                  goto FAILED;
# Line 5337  for (;; ptr++) Line 5540  for (;; ptr++)
5540                of the group. */                of the group. */
5541    
5542                called = cd->start_code + recno;                called = cd->start_code + recno;
5543                PUTINC(cd->hwm, 0, code + 2 + LINK_SIZE - cd->start_code);                PUTINC(cd->hwm, 0, (int)(code + 2 + LINK_SIZE - cd->start_code));
5544                }                }
5545    
5546              /* If not a forward reference, and the subpattern is still open,              /* If not a forward reference, and the subpattern is still open,
# Line 5361  for (;; ptr++) Line 5564  for (;; ptr++)
5564            code += 1 + LINK_SIZE;            code += 1 + LINK_SIZE;
5565    
5566            *code = OP_RECURSE;            *code = OP_RECURSE;
5567            PUT(code, 1, called - cd->start_code);            PUT(code, 1, (int)(called - cd->start_code));
5568            code += 1 + LINK_SIZE;            code += 1 + LINK_SIZE;
5569    
5570            *code = OP_KET;            *code = OP_KET;
# Line 5665  for (;; ptr++) Line 5868  for (;; ptr++)
5868    
5869      /* ===================================================================*/      /* ===================================================================*/
5870      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
5871      are arranged to be the negation of the corresponding OP_values in the      are arranged to be the negation of the corresponding OP_values in the
5872      default case when PCRE_UCP is not set. For the back references, the values      default case when PCRE_UCP is not set. For the back references, the values
5873      are ESC_REF plus the reference number. Only back references and those types      are ESC_REF plus the reference number. Only back references and those types
5874      that consume a character may be repeated. We can test for values between      that consume a character may be repeated. We can test for values between
# Line 5843  for (;; ptr++) Line 6046  for (;; ptr++)
6046            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
6047            }            }
6048          else          else
6049  #endif  #endif
6050            {            {
6051            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
6052            *code++ = -c;            *code++ = -c;
6053            }            }
6054          }          }
6055        continue;        continue;
6056        }        }
# Line 6179  for (;;) Line 6382  for (;;)
6382      {      {
6383      if (lengthptr == NULL)      if (lengthptr == NULL)
6384        {        {
6385        int branch_length = code - last_branch;        int branch_length = (int)(code - last_branch);
6386        do        do
6387          {          {
6388          int prev_length = GET(last_branch, 1);          int prev_length = GET(last_branch, 1);
# Line 6193  for (;;) Line 6396  for (;;)
6396      /* Fill in the ket */      /* Fill in the ket */
6397    
6398      *code = OP_KET;      *code = OP_KET;
6399      PUT(code, 1, code - start_bracket);      PUT(code, 1, (int)(code - start_bracket));
6400      code += 1 + LINK_SIZE;      code += 1 + LINK_SIZE;
6401    
6402      /* If it was a capturing subpattern, check to see if it contained any      /* If it was a capturing subpattern, check to see if it contained any
# Line 6208  for (;;) Line 6411  for (;;)
6411            code - start_bracket);            code - start_bracket);
6412          *start_bracket = OP_ONCE;          *start_bracket = OP_ONCE;
6413          code += 1 + LINK_SIZE;          code += 1 + LINK_SIZE;
6414          PUT(start_bracket, 1, code - start_bracket);          PUT(start_bracket, 1, (int)(code - start_bracket));
6415          *code = OP_KET;          *code = OP_KET;
6416          PUT(code, 1, code - start_bracket);          PUT(code, 1, (int)(code - start_bracket));
6417          code += 1 + LINK_SIZE;          code += 1 + LINK_SIZE;
6418          length += 2 + 2*LINK_SIZE;          length += 2 + 2*LINK_SIZE;
6419          }          }
# Line 6265  for (;;) Line 6468  for (;;)
6468    else    else
6469      {      {
6470      *code = OP_ALT;      *code = OP_ALT;
6471      PUT(code, 1, code - last_branch);      PUT(code, 1, (int)(code - last_branch));
6472      bc.current_branch = last_branch = code;      bc.current_branch = last_branch = code;
6473      code += 1 + LINK_SIZE;      code += 1 + LINK_SIZE;
6474      }      }
# Line 6679  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 6882  while (ptr[skipatstart] == CHAR_LEFT_PAR
6882      options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr;      options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr;
6883    else break;    else break;
6884    }    }
6885    
6886  utf8 = (options & PCRE_UTF8) != 0;  utf8 = (options & PCRE_UTF8) != 0;
6887    
6888  /* Can't support UTF8 unless PCRE has been compiled to include the code. */  /* Can't support UTF8 unless PCRE has been compiled to include the code. */
# Line 6705  if (utf8) Line 6908  if (utf8)
6908  if ((options & PCRE_UCP) != 0)  if ((options & PCRE_UCP) != 0)
6909    {    {
6910    errorcode = ERR67;    errorcode = ERR67;
6911    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
6912    }    }
6913  #endif  #endif
6914    
6915  /* Check validity of \R options. */  /* Check validity of \R options. */
# Line 6837  regex compiled on a system with 4-byte p Line 7040  regex compiled on a system with 4-byte p
7040  pointers. */  pointers. */
7041    
7042  re->magic_number = MAGIC_NUMBER;  re->magic_number = MAGIC_NUMBER;
7043  re->size = size;  re->size = (int)size;
7044  re->options = cd->external_options;  re->options = cd->external_options;
7045  re->flags = cd->external_flags;  re->flags = cd->external_flags;
7046  re->dummy1 = 0;  re->dummy1 = 0;
# Line 6908  while (errorcode == 0 && cd->hwm > cwork Line 7111  while (errorcode == 0 && cd->hwm > cwork
7111    recno = GET(codestart, offset);    recno = GET(codestart, offset);
7112    groupptr = _pcre_find_bracket(codestart, utf8, recno);    groupptr = _pcre_find_bracket(codestart, utf8, recno);
7113    if (groupptr == NULL) errorcode = ERR53;    if (groupptr == NULL) errorcode = ERR53;
7114      else PUT(((uschar *)codestart), offset, groupptr - codestart);      else PUT(((uschar *)codestart), offset, (int)(groupptr - codestart));
7115    }    }
7116    
7117  /* Give an error if there's back reference to a non-existent capturing  /* Give an error if there's back reference to a non-existent capturing
# Line 6963  if (errorcode != 0) Line 7166  if (errorcode != 0)
7166    {    {
7167    (pcre_free)(re);    (pcre_free)(re);
7168    PCRE_EARLY_ERROR_RETURN:    PCRE_EARLY_ERROR_RETURN:
7169    *erroroffset = ptr - (const uschar *)pattern;    *erroroffset = (int)(ptr - (const uschar *)pattern);
7170    PCRE_EARLY_ERROR_RETURN2:    PCRE_EARLY_ERROR_RETURN2:
7171    *errorptr = find_error_text(errorcode);    *errorptr = find_error_text(errorcode);
7172    if (errorcodeptr != NULL) *errorcodeptr = errorcode;    if (errorcodeptr != NULL) *errorcodeptr = errorcode;

Legend:
Removed from v.518  
changed lines
  Added in v.574

  ViewVC Help
Powered by ViewVC 1.1.5