/[pcre]/code/trunk/pcre_compile.c
ViewVC logotype

Diff of /code/trunk/pcre_compile.c

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

revision 533 by ph10, Wed Jun 2 19:02:41 2010 UTC revision 578 by ph10, Tue Nov 23 15:34:55 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 1099  top-level call starts at the beginning o Line 1105  top-level call starts at the beginning o
1105  start at a parenthesis. It scans along a pattern's text looking for capturing  start at a parenthesis. It scans along a pattern's text looking for capturing
1106  subpatterns, and counting them. If it finds a named pattern that matches the  subpatterns, and counting them. If it finds a named pattern that matches the
1107  name it is given, it returns its number. Alternatively, if the name is NULL, it  name it is given, it returns its number. Alternatively, if the name is NULL, it
1108  returns when it reaches a given numbered subpattern. We know that if (?P< is  returns when it reaches a given numbered subpattern. Recursion is used to keep
1109  encountered, the name will be terminated by '>' because that is checked in the  track of subpatterns that reset the capturing group numbers - the (?| feature.
1110  first pass. Recursion is used to keep track of subpatterns that reset the  
1111  capturing group numbers - the (?| feature.  This function was originally called only from the second pass, in which we know
1112    that if (?< or (?' or (?P< is encountered, the name will be correctly
1113    terminated because that is checked in the first pass. There is now one call to
1114    this function in the first pass, to check for a recursive back reference by
1115    name (so that we can make the whole group atomic). In this case, we need check
1116    only up to the current position in the pattern, and that is still OK because
1117    and previous occurrences will have been checked. To make this work, the test
1118    for "end of pattern" is a check against cd->end_pattern in the main loop,
1119    instead of looking for a binary zero. This means that the special first-pass
1120    call can adjust cd->end_pattern temporarily. (Checks for binary zero while
1121    processing items within the loop are OK, because afterwards the main loop will
1122    terminate.)
1123    
1124  Arguments:  Arguments:
1125    ptrptr       address of the current character pointer (updated)    ptrptr       address of the current character pointer (updated)
# Line 1110  Arguments: Line 1127  Arguments:
1127    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1128    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1129    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1130      utf8         TRUE if we are in UTF-8 mode
1131    count        pointer to the current capturing subpattern number (updated)    count        pointer to the current capturing subpattern number (updated)
1132    
1133  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 1135  Returns:       the number of the named s
1135    
1136  static int  static int
1137  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,
1138    BOOL xmode, int *count)    BOOL xmode, BOOL utf8, int *count)
1139  {  {
1140  uschar *ptr = *ptrptr;  uschar *ptr = *ptrptr;
1141  int start_count = *count;  int start_count = *count;
# Line 1129  dealing with. The very first call may no Line 1147  dealing with. The very first call may no
1147    
1148  if (ptr[0] == CHAR_LEFT_PARENTHESIS)  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1149    {    {
1150    if (ptr[1] == CHAR_QUESTION_MARK &&    /* Handle specials such as (*SKIP) or (*UTF8) etc. */
1151        ptr[2] == CHAR_VERTICAL_LINE)  
1152      if (ptr[1] == CHAR_ASTERISK) ptr += 2;
1153    
1154      /* Handle a normal, unnamed capturing parenthesis. */
1155    
1156      else if (ptr[1] != CHAR_QUESTION_MARK)
1157        {
1158        *count += 1;
1159        if (name == NULL && *count == lorn) return *count;
1160        ptr++;
1161        }
1162    
1163      /* All cases now have (? at the start. Remember when we are in a group
1164      where the parenthesis numbers are duplicated. */
1165    
1166      else if (ptr[2] == CHAR_VERTICAL_LINE)
1167      {      {
1168      ptr += 3;      ptr += 3;
1169      dup_parens = TRUE;      dup_parens = TRUE;
1170      }      }
1171    
1172    /* Handle a normal, unnamed capturing parenthesis */    /* Handle comments; all characters are allowed until a ket is reached. */
1173    
1174    else if (ptr[1] != CHAR_QUESTION_MARK && ptr[1] != CHAR_ASTERISK)    else if (ptr[2] == CHAR_NUMBER_SIGN)
1175      {      {
1176      *count += 1;      for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
1177      if (name == NULL && *count == lorn) return *count;      goto FAIL_EXIT;
     ptr++;  
1178      }      }
1179    
1180    /* 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
1181    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
1182    condition (there can't be any nested parens. */    condition (there can't be any nested parens). */
1183    
1184    else if (ptr[2] == CHAR_LEFT_PARENTHESIS)    else if (ptr[2] == CHAR_LEFT_PARENTHESIS)
1185      {      {
# Line 1159  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1191  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1191        }        }
1192      }      }
1193    
1194    /* We have either (? or (* and not a condition */    /* Start with (? but not a condition. */
1195    
1196    else    else
1197      {      {
# Line 1188  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1220  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1220    }    }
1221    
1222  /* Past any initial parenthesis handling, scan for parentheses or vertical  /* Past any initial parenthesis handling, scan for parentheses or vertical
1223  bars. */  bars. Stop if we get to cd->end_pattern. Note that this is important for the
1224    first-pass call when this value is temporarily adjusted to stop at the current
1225    position. So DO NOT change this to a test for binary zero. */
1226    
1227  for (; *ptr != 0; ptr++)  for (; ptr < cd->end_pattern; ptr++)
1228    {    {
1229    /* Skip over backslashed characters and also entire \Q...\E */    /* Skip over backslashed characters and also entire \Q...\E */
1230    
# Line 1264  for (; *ptr != 0; ptr++) Line 1298  for (; *ptr != 0; ptr++)
1298    
1299    if (xmode && *ptr == CHAR_NUMBER_SIGN)    if (xmode && *ptr == CHAR_NUMBER_SIGN)
1300      {      {
1301      while (*(++ptr) != 0 && *ptr != CHAR_NL) {};      ptr++;
1302        while (*ptr != 0)
1303          {
1304          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
1305          ptr++;
1306    #ifdef SUPPORT_UTF8
1307          if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
1308    #endif
1309          }
1310      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == 0) goto FAIL_EXIT;
1311      continue;      continue;
1312      }      }
# Line 1273  for (; *ptr != 0; ptr++) Line 1315  for (; *ptr != 0; ptr++)
1315    
1316    if (*ptr == CHAR_LEFT_PARENTHESIS)    if (*ptr == CHAR_LEFT_PARENTHESIS)
1317      {      {
1318      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, count);      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, count);
1319      if (rc > 0) return rc;      if (rc > 0) return rc;
1320      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == 0) goto FAIL_EXIT;
1321      }      }
# Line 1281  for (; *ptr != 0; ptr++) Line 1323  for (; *ptr != 0; ptr++)
1323    else if (*ptr == CHAR_RIGHT_PARENTHESIS)    else if (*ptr == CHAR_RIGHT_PARENTHESIS)
1324      {      {
1325      if (dup_parens && *count < hwm_count) *count = hwm_count;      if (dup_parens && *count < hwm_count) *count = hwm_count;
1326      *ptrptr = ptr;      goto FAIL_EXIT;
     return -1;  
1327      }      }
1328    
1329    else if (*ptr == CHAR_VERTICAL_LINE && dup_parens)    else if (*ptr == CHAR_VERTICAL_LINE && dup_parens)
# Line 1320  Arguments: Line 1361  Arguments:
1361    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1362    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1363    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1364      utf8         TRUE if we are in UTF-8 mode
1365    
1366  Returns:       the number of the found subpattern, or -1 if not found  Returns:       the number of the found subpattern, or -1 if not found
1367  */  */
1368    
1369  static int  static int
1370  find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode)  find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode,
1371      BOOL utf8)
1372  {  {
1373  uschar *ptr = (uschar *)cd->start_pattern;  uschar *ptr = (uschar *)cd->start_pattern;
1374  int count = 0;  int count = 0;
# Line 1338  matching closing parens. That is why we Line 1381  matching closing parens. That is why we
1381    
1382  for (;;)  for (;;)
1383    {    {
1384    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, &count);    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, &count);
1385    if (rc > 0 || *ptr++ == 0) break;    if (rc > 0 || *ptr++ == 0) break;
1386    }    }
1387    
# Line 1711  for (;;) Line 1754  for (;;)
1754        case OP_MARK:        case OP_MARK:
1755        case OP_PRUNE_ARG:        case OP_PRUNE_ARG:
1756        case OP_SKIP_ARG:        case OP_SKIP_ARG:
       case OP_THEN_ARG:  
1757        code += code[1];        code += code[1];
1758        break;        break;
1759    
1760          case OP_THEN_ARG:
1761          code += code[1+LINK_SIZE];
1762          break;
1763        }        }
1764    
1765      /* Add in the fixed length from the table */      /* Add in the fixed length from the table */
# Line 1814  for (;;) Line 1860  for (;;)
1860        case OP_MARK:        case OP_MARK:
1861        case OP_PRUNE_ARG:        case OP_PRUNE_ARG:
1862        case OP_SKIP_ARG:        case OP_SKIP_ARG:
       case OP_THEN_ARG:  
1863        code += code[1];        code += code[1];
1864        break;        break;
1865    
1866          case OP_THEN_ARG:
1867          code += code[1+LINK_SIZE];
1868          break;
1869        }        }
1870    
1871      /* 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 2141  for (code = first_significant_code(code
2141      case OP_MARK:      case OP_MARK:
2142      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
2143      case OP_SKIP_ARG:      case OP_SKIP_ARG:
     case OP_THEN_ARG:  
2144      code += code[1];      code += code[1];
2145      break;      break;
2146    
2147        case OP_THEN_ARG:
2148        code += code[1+LINK_SIZE];
2149        break;
2150    
2151      /* None of the remaining opcodes are required to match a character. */      /* None of the remaining opcodes are required to match a character. */
2152    
2153      default:      default:
# Line 2407  Arguments: Line 2459  Arguments:
2459    ptype        the property type    ptype        the property type
2460    pdata        the data for the type    pdata        the data for the type
2461    negated      TRUE if it's a negated property (\P or \p{^)    negated      TRUE if it's a negated property (\P or \p{^)
2462    
2463  Returns:       TRUE if auto-possessifying is OK  Returns:       TRUE if auto-possessifying is OK
2464  */  */
2465    
2466  static BOOL  static BOOL
2467  check_char_prop(int c, int ptype, int pdata, BOOL negated)  check_char_prop(int c, int ptype, int pdata, BOOL negated)
# Line 2453  switch(ptype) Line 2505  switch(ptype)
2505            _pcre_ucp_gentype[prop->chartype] == ucp_N ||            _pcre_ucp_gentype[prop->chartype] == ucp_N ||
2506            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
2507    }    }
2508  return FALSE;  return FALSE;
2509  }  }
2510  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2511    
# Line 2478  Returns:        TRUE if possessifying is Line 2530  Returns:        TRUE if possessifying is
2530  */  */
2531    
2532  static BOOL  static BOOL
2533  check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr,  check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr,
2534    int options, compile_data *cd)    int options, compile_data *cd)
2535  {  {
2536  int c, next;  int c, next;
# Line 2493  if ((options & PCRE_EXTENDED) != 0) Line 2545  if ((options & PCRE_EXTENDED) != 0)
2545      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
2546      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
2547        {        {
2548        while (*(++ptr) != 0)        ptr++;
2549          while (*ptr != 0)
2550            {
2551          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2552            ptr++;
2553    #ifdef SUPPORT_UTF8
2554            if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
2555    #endif
2556            }
2557        }        }
2558      else break;      else break;
2559      }      }
# Line 2530  if ((options & PCRE_EXTENDED) != 0) Line 2589  if ((options & PCRE_EXTENDED) != 0)
2589      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
2590      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
2591        {        {
2592        while (*(++ptr) != 0)        ptr++;
2593          while (*ptr != 0)
2594            {
2595          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2596            ptr++;
2597    #ifdef SUPPORT_UTF8
2598            if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
2599    #endif
2600            }
2601        }        }
2602      else break;      else break;
2603      }      }
# Line 2549  the next item is a character. */ Line 2615  the next item is a character. */
2615  if (next >= 0) switch(op_code)  if (next >= 0) switch(op_code)
2616    {    {
2617    case OP_CHAR:    case OP_CHAR:
2618  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2619    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
2620  #else  #else
2621    c = *previous;    c = *previous;
2622  #endif  #endif
2623    return c != next;    return c != next;
2624    
2625    /* 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
2626    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
2627    high-valued characters. */    high-valued characters. */
2628    
2629    case OP_CHARNC:    case OP_CHARNC:
2630  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2631    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
2632  #else  #else
2633    c = *previous;    c = *previous;
2634  #endif  #endif
2635    if (c == next) return FALSE;    if (c == next) return FALSE;
2636  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2637    if (utf8)    if (utf8)
# Line 2603  if (next >= 0) switch(op_code) Line 2669  if (next >= 0) switch(op_code)
2669    else    else
2670  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
2671    return (c == cd->fcc[next]);  /* Non-UTF-8 mode */    return (c == cd->fcc[next]);  /* Non-UTF-8 mode */
2672    
2673    /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.    /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
2674    When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */    When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
2675    
2676    case OP_DIGIT:    case OP_DIGIT:
2677    return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;    return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;
# Line 2673  if (next >= 0) switch(op_code) Line 2739  if (next >= 0) switch(op_code)
2739  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2740    case OP_PROP:    case OP_PROP:
2741    return check_char_prop(next, previous[0], previous[1], FALSE);    return check_char_prop(next, previous[0], previous[1], FALSE);
2742    
2743    case OP_NOTPROP:    case OP_NOTPROP:
2744    return check_char_prop(next, previous[0], previous[1], TRUE);    return check_char_prop(next, previous[0], previous[1], TRUE);
2745  #endif  #endif
# Line 2683  if (next >= 0) switch(op_code) Line 2749  if (next >= 0) switch(op_code)
2749    }    }
2750    
2751    
2752  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
2753  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
2754  generated only when PCRE_UCP is *not* set, that is, when only ASCII  generated only when PCRE_UCP is *not* set, that is, when only ASCII
2755  characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are  characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are
2756  replaced by OP_PROP codes when PCRE_UCP is set. */  replaced by OP_PROP codes when PCRE_UCP is set. */
2757    
2758  switch(op_code)  switch(op_code)
2759    {    {
2760    case OP_CHAR:    case OP_CHAR:
2761    case OP_CHARNC:    case OP_CHARNC:
2762  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2763    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
2764  #else  #else
2765    c = *previous;    c = *previous;
2766  #endif  #endif
2767    switch(-next)    switch(-next)
2768      {      {
2769      case ESC_d:      case ESC_d:
# Line 2761  switch(op_code) Line 2827  switch(op_code)
2827        default:        default:
2828        return -next == ESC_v;        return -next == ESC_v;
2829        }        }
2830    
2831      /* When PCRE_UCP is set, these values get generated for \d etc. Find      /* When PCRE_UCP is set, these values get generated for \d etc. Find
2832      their substitutions and process them. The result will always be either      their substitutions and process them. The result will always be either
2833      -ESC_p or -ESC_P. Then fall through to process those values. */      -ESC_p or -ESC_P. Then fall through to process those values. */
2834    
2835  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2836      case ESC_du:      case ESC_du:
2837      case ESC_DU:      case ESC_DU:
# Line 2780  switch(op_code) Line 2846  switch(op_code)
2846        if (temperrorcode != 0) return FALSE;        if (temperrorcode != 0) return FALSE;
2847        ptr++;    /* For compatibility */        ptr++;    /* For compatibility */
2848        }        }
2849      /* Fall through */      /* Fall through */
2850    
2851      case ESC_p:      case ESC_p:
2852      case ESC_P:      case ESC_P:
2853        {        {
2854        int ptype, pdata, errorcodeptr;        int ptype, pdata, errorcodeptr;
2855        BOOL negated;        BOOL negated;
2856    
2857        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
2858        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);
2859        if (ptype < 0) return FALSE;        if (ptype < 0) return FALSE;
2860        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
2861    
2862        /* If the property item is optional, we have to give up. (When generated        /* If the property item is optional, we have to give up. (When generated
2863        from \d etc by PCRE_UCP, this test will have been applied much earlier,        from \d etc by PCRE_UCP, this test will have been applied much earlier,
2864        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. */
2865    
2866        if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||        if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
2867          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)
2868            return FALSE;            return FALSE;
2869    
2870        /* Do the property check. */        /* Do the property check. */
2871    
2872        return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);        return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);
2873        }        }
2874  #endif  #endif
2875    
2876      default:      default:
2877      return FALSE;      return FALSE;
2878      }      }
2879    
2880    /* In principle, support for Unicode properties should be integrated here as    /* In principle, support for Unicode properties should be integrated here as
2881    well. It means re-organizing the above code so as to get hold of the property    well. It means re-organizing the above code so as to get hold of the property
2882    values before switching on the op-code. However, I wonder how many patterns    values before switching on the op-code. However, I wonder how many patterns
2883    combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set,    combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set,
2884    these op-codes are never generated.) */    these op-codes are never generated.) */
2885    
2886    case OP_DIGIT:    case OP_DIGIT:
2887    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
# Line 2831  switch(op_code) Line 2897  switch(op_code)
2897    return next == -ESC_s || next == -ESC_h || next == -ESC_v;    return next == -ESC_s || next == -ESC_h || next == -ESC_v;
2898    
2899    case OP_HSPACE:    case OP_HSPACE:
2900    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||
2901           next == -ESC_w || next == -ESC_v || next == -ESC_R;           next == -ESC_w || next == -ESC_v || next == -ESC_R;
2902    
2903    case OP_NOT_HSPACE:    case OP_NOT_HSPACE:
2904    return next == -ESC_h;    return next == -ESC_h;
2905    
2906    /* Can't have \S in here because VT matches \S (Perl anomaly) */    /* Can't have \S in here because VT matches \S (Perl anomaly) */
2907    case OP_ANYNL:    case OP_ANYNL:
2908    case OP_VSPACE:    case OP_VSPACE:
2909    return next == -ESC_V || next == -ESC_d || next == -ESC_w;    return next == -ESC_V || next == -ESC_d || next == -ESC_w;
2910    
# Line 2846  switch(op_code) Line 2912  switch(op_code)
2912    return next == -ESC_v || next == -ESC_R;    return next == -ESC_v || next == -ESC_R;
2913    
2914    case OP_WORDCHAR:    case OP_WORDCHAR:
2915    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||
2916           next == -ESC_v || next == -ESC_R;           next == -ESC_v || next == -ESC_R;
2917    
2918    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
# Line 2982  for (;; ptr++) Line 3048  for (;; ptr++)
3048    
3049    c = *ptr;    c = *ptr;
3050    
3051    /* 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
3052    string. Nesting only happens one level deep. */    string. Nesting only happens one level deep. */
3053    
3054    if (c == 0 && nestptr != NULL)    if (c == 0 && nestptr != NULL)
# Line 3104  for (;; ptr++) Line 3170  for (;; ptr++)
3170      if ((cd->ctypes[c] & ctype_space) != 0) continue;      if ((cd->ctypes[c] & ctype_space) != 0) continue;
3171      if (c == CHAR_NUMBER_SIGN)      if (c == CHAR_NUMBER_SIGN)
3172        {        {
3173        while (*(++ptr) != 0)        ptr++;
3174          while (*ptr != 0)
3175          {          {
3176          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
3177            ptr++;
3178    #ifdef SUPPORT_UTF8
3179            if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
3180    #endif
3181          }          }
3182        if (*ptr != 0) continue;        if (*ptr != 0) continue;
3183    
# Line 3289  for (;; ptr++) Line 3360  for (;; ptr++)
3360          {                           /* Braces are required because the */          {                           /* Braces are required because the */
3361          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
3362          }          }
3363    
3364        /* 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
3365        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
3366        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 3358  for (;; ptr++) Line 3429  for (;; ptr++)
3429    
3430          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
3431            posix_class = 0;            posix_class = 0;
3432    
3433          /* 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
3434          different escape sequences that use Unicode properties. */          different escape sequences that use Unicode properties. */
3435    
3436  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3437          if ((options & PCRE_UCP) != 0)          if ((options & PCRE_UCP) != 0)
3438            {            {
3439            int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);            int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);
3440            if (posix_substitutes[pc] != NULL)            if (posix_substitutes[pc] != NULL)
3441              {              {
3442              nestptr = tempptr + 1;              nestptr = tempptr + 1;
3443              ptr = posix_substitutes[pc] - 1;              ptr = posix_substitutes[pc] - 1;
3444              continue;              continue;
3445              }              }
3446            }            }
3447  #endif  #endif
3448          /* 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
3449          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,
3450          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 3460  for (;; ptr++) Line 3531  for (;; ptr++)
3531              case ESC_SU:              case ESC_SU:
3532              nestptr = ptr;              nestptr = ptr;
3533              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
3534              class_charcount -= 2;                /* Undo! */              class_charcount -= 2;                /* Undo! */
3535              continue;              continue;
3536  #endif  #endif
3537              case ESC_d:              case ESC_d:
# Line 3481  for (;; ptr++) Line 3552  for (;; ptr++)
3552              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];
3553              continue;              continue;
3554    
3555                /* Perl 5.004 onwards omits VT from \s, but we must preserve it
3556                if it was previously set by something earlier in the character
3557                class. */
3558    
3559              case ESC_s:              case ESC_s:
3560              for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space];              classbits[0] |= cbits[cbit_space];
3561              classbits[1] &= ~0x08;   /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= cbits[cbit_space+1] & ~0x08;
3562                for (c = 2; c < 32; c++) classbits[c] |= cbits[c+cbit_space];
3563              continue;              continue;
3564    
3565              case ESC_S:              case ESC_S:
# Line 3911  for (;; ptr++) Line 3987  for (;; ptr++)
3987      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
3988      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
3989      reqbyte, save the previous value for reinstating. */      reqbyte, save the previous value for reinstating. */
3990    
3991  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3992      if (class_charcount == 1 && !class_utf8 &&      if (class_charcount == 1 && !class_utf8 &&
3993        (!utf8 || !negate_class || class_lastchar < 128))        (!utf8 || !negate_class || class_lastchar < 128))
# Line 3991  for (;; ptr++) Line 4067  for (;; ptr++)
4067        }        }
4068  #endif  #endif
4069    
4070      /* 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
4071      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
4072      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
4073      (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 4795  for (;; ptr++) Line 4871  for (;; ptr++)
4871                *errorcodeptr = ERR66;                *errorcodeptr = ERR66;
4872                goto FAILED;                goto FAILED;
4873                }                }
4874              *code++ = verbs[i].op;              *code = verbs[i].op;
4875                if (*code++ == OP_THEN)
4876                  {
4877                  PUT(code, 0, code - bcptr->current_branch - 1);
4878                  code += LINK_SIZE;
4879                  }
4880              }              }
4881    
4882            else            else
# Line 4805  for (;; ptr++) Line 4886  for (;; ptr++)
4886                *errorcodeptr = ERR59;                *errorcodeptr = ERR59;
4887                goto FAILED;                goto FAILED;
4888                }                }
4889              *code++ = verbs[i].op_arg;              *code = verbs[i].op_arg;
4890                if (*code++ == OP_THEN_ARG)
4891                  {
4892                  PUT(code, 0, code - bcptr->current_branch - 1);
4893                  code += LINK_SIZE;
4894                  }
4895              *code++ = arglen;              *code++ = arglen;
4896              memcpy(code, arg, arglen);              memcpy(code, arg, arglen);
4897              code += arglen;              code += arglen;
# Line 4999  for (;; ptr++) Line 5085  for (;; ptr++)
5085          /* Search the pattern for a forward reference */          /* Search the pattern for a forward reference */
5086    
5087          else if ((i = find_parens(cd, name, namelen,          else if ((i = find_parens(cd, name, namelen,
5088                          (options & PCRE_EXTENDED) != 0)) > 0)                          (options & PCRE_EXTENDED) != 0, utf8)) > 0)
5089            {            {
5090            PUT2(code, 2+LINK_SIZE, i);            PUT2(code, 2+LINK_SIZE, i);
5091            code[1+LINK_SIZE]++;            code[1+LINK_SIZE]++;
# Line 5300  for (;; ptr++) Line 5386  for (;; ptr++)
5386          while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;          while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
5387          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5388    
5389          /* In the pre-compile phase, do a syntax check and set a dummy          /* In the pre-compile phase, do a syntax check. We used to just set
5390          reference number. */          a dummy reference number, because it was not used in the first pass.
5391            However, with the change of recursive back references to be atomic,
5392            we have to look for the number so that this state can be identified, as
5393            otherwise the incorrect length is computed. If it's not a backwards
5394            reference, the dummy number will do. */
5395    
5396          if (lengthptr != NULL)          if (lengthptr != NULL)
5397            {            {
5398              const uschar *temp;
5399    
5400            if (namelen == 0)            if (namelen == 0)
5401              {              {
5402              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
# Line 5320  for (;; ptr++) Line 5412  for (;; ptr++)
5412              *errorcodeptr = ERR48;              *errorcodeptr = ERR48;
5413              goto FAILED;              goto FAILED;
5414              }              }
5415            recno = 0;  
5416              /* The name table does not exist in the first pass, so we cannot
5417              do a simple search as in the code below. Instead, we have to scan the
5418              pattern to find the number. It is important that we scan it only as
5419              far as we have got because the syntax of named subpatterns has not
5420              been checked for the rest of the pattern, and find_parens() assumes
5421              correct syntax. In any case, it's a waste of resources to scan
5422              further. We stop the scan at the current point by temporarily
5423              adjusting the value of cd->endpattern. */
5424    
5425              temp = cd->end_pattern;
5426              cd->end_pattern = ptr;
5427              recno = find_parens(cd, name, namelen,
5428                (options & PCRE_EXTENDED) != 0, utf8);
5429              cd->end_pattern = temp;
5430              if (recno < 0) recno = 0;    /* Forward ref; set dummy number */
5431            }            }
5432    
5433          /* In the real compile, seek the name in the table. We check the name          /* In the real compile, seek the name in the table. We check the name
# Line 5345  for (;; ptr++) Line 5452  for (;; ptr++)
5452              }              }
5453            else if ((recno =                /* Forward back reference */            else if ((recno =                /* Forward back reference */
5454                      find_parens(cd, name, namelen,                      find_parens(cd, name, namelen,
5455                        (options & PCRE_EXTENDED) != 0)) <= 0)                        (options & PCRE_EXTENDED) != 0, utf8)) <= 0)
5456              {              {
5457              *errorcodeptr = ERR15;              *errorcodeptr = ERR15;
5458              goto FAILED;              goto FAILED;
# Line 5456  for (;; ptr++) Line 5563  for (;; ptr++)
5563              if (called == NULL)              if (called == NULL)
5564                {                {
5565                if (find_parens(cd, NULL, recno,                if (find_parens(cd, NULL, recno,
5566                      (options & PCRE_EXTENDED) != 0) < 0)                      (options & PCRE_EXTENDED) != 0, utf8) < 0)
5567                  {                  {
5568                  *errorcodeptr = ERR15;                  *errorcodeptr = ERR15;
5569                  goto FAILED;                  goto FAILED;
# Line 5795  for (;; ptr++) Line 5902  for (;; ptr++)
5902    
5903      /* ===================================================================*/      /* ===================================================================*/
5904      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
5905      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
5906      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
5907      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
5908      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 5973  for (;; ptr++) Line 6080  for (;; ptr++)
6080            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
6081            }            }
6082          else          else
6083  #endif  #endif
6084            {            {
6085            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
6086            *code++ = -c;            *code++ = -c;
6087            }            }
6088          }          }
6089        continue;        continue;
6090        }        }
# Line 6786  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 6893  while (ptr[skipatstart] == CHAR_LEFT_PAR
6893      { skipatstart += 7; options |= PCRE_UTF8; continue; }      { skipatstart += 7; options |= PCRE_UTF8; continue; }
6894    else if (strncmp((char *)(ptr+skipatstart+2), STRING_UCP_RIGHTPAR, 4) == 0)    else if (strncmp((char *)(ptr+skipatstart+2), STRING_UCP_RIGHTPAR, 4) == 0)
6895      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
6896      else if (strncmp((char *)(ptr+skipatstart+2), STRING_NO_START_OPT_RIGHTPAR, 13) == 0)
6897        { skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; }
6898    
6899    if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0)    if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0)
6900      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }
# Line 6809  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 6918  while (ptr[skipatstart] == CHAR_LEFT_PAR
6918      options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr;      options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr;
6919    else break;    else break;
6920    }    }
6921    
6922  utf8 = (options & PCRE_UTF8) != 0;  utf8 = (options & PCRE_UTF8) != 0;
6923    
6924  /* 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 6835  if (utf8) Line 6944  if (utf8)
6944  if ((options & PCRE_UCP) != 0)  if ((options & PCRE_UCP) != 0)
6945    {    {
6946    errorcode = ERR67;    errorcode = ERR67;
6947    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
6948    }    }
6949  #endif  #endif
6950    
6951  /* Check validity of \R options. */  /* Check validity of \R options. */

Legend:
Removed from v.533  
changed lines
  Added in v.578

  ViewVC Help
Powered by ViewVC 1.1.5