/[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 178 by ph10, Wed Jun 13 08:44:34 2007 UTC revision 216 by ph10, Wed Aug 15 14:35:57 2007 UTC
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  supporting internal functions that are not used by other modules. */  supporting internal functions that are not used by other modules. */
43    
44    
45    #ifdef HAVE_CONFIG_H
46    #include <config.h>
47    #endif
48    
49  #define NLBLOCK cd             /* Block containing newline information */  #define NLBLOCK cd             /* Block containing newline information */
50  #define PSSTART start_pattern  /* Field containing processed string start */  #define PSSTART start_pattern  /* Field containing processed string start */
51  #define PSEND   end_pattern    /* Field containing processed string end */  #define PSEND   end_pattern    /* Field containing processed string end */
52    
   
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    
# Line 62  used by pcretest. DEBUG is not defined w Line 65  used by pcretest. DEBUG is not defined w
65    
66  #define SETBIT(a,b) a[b/8] |= (1 << (b%8))  #define SETBIT(a,b) a[b/8] |= (1 << (b%8))
67    
68    /* Maximum length value to check against when making sure that the integer that
69    holds the compiled pattern length does not overflow. We make it a bit less than
70    INT_MAX to allow for adding in group terminating bytes, so that we don't have
71    to check them every time. */
72    
73    #define OFLOW_MAX (INT_MAX - 20)
74    
75    
76  /*************************************************  /*************************************************
77  *      Code parameters and static tables         *  *      Code parameters and static tables         *
# Line 120  static const short int escapes[] = { Line 130  static const short int escapes[] = {
130  /*  B8 */     0,     0,      0,       0,      0,   ']',    '=',    '-',  /*  B8 */     0,     0,      0,       0,      0,   ']',    '=',    '-',
131  /*  C0 */   '{',-ESC_A, -ESC_B,  -ESC_C, -ESC_D,-ESC_E,      0, -ESC_G,  /*  C0 */   '{',-ESC_A, -ESC_B,  -ESC_C, -ESC_D,-ESC_E,      0, -ESC_G,
132  /*  C8 */-ESC_H,     0,      0,       0,      0,     0,      0,      0,  /*  C8 */-ESC_H,     0,      0,       0,      0,     0,      0,      0,
133  /*  D0 */   '}',     0,      0,       0,      0,     0,      0, -ESC_P,  /*  D0 */   '}',     0, -ESC_K,       0,      0,     0,      0, -ESC_P,
134  /*  D8 */-ESC_Q,-ESC_R,      0,       0,      0,     0,      0,      0,  /*  D8 */-ESC_Q,-ESC_R,      0,       0,      0,     0,      0,      0,
135  /*  E0 */  '\\',     0, -ESC_S,       0,      0,-ESC_V, -ESC_W, -ESC_X,  /*  E0 */  '\\',     0, -ESC_S,       0,      0,-ESC_V, -ESC_W, -ESC_X,
136  /*  E8 */     0,-ESC_Z,      0,       0,      0,     0,      0,      0,  /*  E8 */     0,-ESC_Z,      0,       0,      0,     0,      0,      0,
# Line 130  static const short int escapes[] = { Line 140  static const short int escapes[] = {
140  #endif  #endif
141    
142    
143    /* Table of special "verbs" like (*PRUNE) */
144    
145    typedef struct verbitem {
146      const char *name;
147      int   len;
148      int   op;
149    } verbitem;
150    
151    static verbitem verbs[] = {
152      { "ACCEPT", 6, OP_ACCEPT },
153      { "COMMIT", 6, OP_COMMIT },
154      { "F",      1, OP_FAIL },
155      { "FAIL",   4, OP_FAIL },
156      { "PRUNE",  5, OP_PRUNE },
157      { "SKIP",   4, OP_SKIP  },
158      { "THEN",   4, OP_THEN  }
159    };
160    
161    static int verbcount = sizeof(verbs)/sizeof(verbitem);
162    
163    
164  /* Tables of names of POSIX character classes and their lengths. The list is  /* Tables of names of POSIX character classes and their lengths. The list is
165  terminated by a zero length entry. The first three must be alpha, lower, upper,  terminated by a zero length entry. The first three must be alpha, lower, upper,
166  as this is assumed for handling case independence. */  as this is assumed for handling case independence. */
# Line 203  static const char *error_texts[] = { Line 234  static const char *error_texts[] = {
234    "missing ) after comment",    "missing ) after comment",
235    "parentheses nested too deeply",  /** DEAD **/    "parentheses nested too deeply",  /** DEAD **/
236    /* 20 */    /* 20 */
237    "regular expression too large",    "regular expression is too large",
238    "failed to get memory",    "failed to get memory",
239    "unmatched parentheses",    "unmatched parentheses",
240    "internal error: code overflow",    "internal error: code overflow",
# Line 239  static const char *error_texts[] = { Line 270  static const char *error_texts[] = {
270    "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)",    "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)",
271    "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")",    "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")",
272    /* 50 */    /* 50 */
273    "repeated subpattern is too long",    "repeated subpattern is too long",    /** DEAD **/
274    "octal value is greater than \\377 (not in UTF-8 mode)",    "octal value is greater than \\377 (not in UTF-8 mode)",
275    "internal error: overran compiling workspace",    "internal error: overran compiling workspace",
276    "internal error: previously-checked referenced subpattern not found",    "internal error: previously-checked referenced subpattern not found",
# Line 248  static const char *error_texts[] = { Line 279  static const char *error_texts[] = {
279    "repeating a DEFINE group is not allowed",    "repeating a DEFINE group is not allowed",
280    "inconsistent NEWLINE options",    "inconsistent NEWLINE options",
281    "\\g is not followed by a braced name or an optionally braced non-zero number",    "\\g is not followed by a braced name or an optionally braced non-zero number",
282    "(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number"    "(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number",
283      "(*VERB) with an argument is not supported",
284      /* 60 */
285      "(*VERB) not recognized",
286      "number is too big"
287  };  };
288    
289    
# Line 379  static const unsigned char ebcdic_charta Line 414  static const unsigned char ebcdic_charta
414  /* Definition to allow mutual recursion */  /* Definition to allow mutual recursion */
415    
416  static BOOL  static BOOL
417    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,
418      int *, int *, branch_chain *, compile_data *, int *);      int *, int *, branch_chain *, compile_data *, int *);
419    
420    
# Line 405  Arguments: Line 440  Arguments:
440    
441  Returns:         zero or positive => a data character  Returns:         zero or positive => a data character
442                   negative => a special escape sequence                   negative => a special escape sequence
443                   on error, errorptr is set                   on error, errorcodeptr is set
444  */  */
445    
446  static int  static int
# Line 489  else Line 524  else
524      c = 0;      c = 0;
525      while ((digitab[ptr[1]] & ctype_digit) != 0)      while ((digitab[ptr[1]] & ctype_digit) != 0)
526        c = c * 10 + *(++ptr) - '0';        c = c * 10 + *(++ptr) - '0';
527    
528        if (c < 0)
529          {
530          *errorcodeptr = ERR61;
531          break;
532          }
533    
534      if (c == 0 || (braced && *(++ptr) != '}'))      if (c == 0 || (braced && *(++ptr) != '}'))
535        {        {
536        *errorcodeptr = ERR57;        *errorcodeptr = ERR57;
537        return 0;        break;
538        }        }
539    
540      if (negated)      if (negated)
# Line 501  else Line 542  else
542        if (c > bracount)        if (c > bracount)
543          {          {
544          *errorcodeptr = ERR15;          *errorcodeptr = ERR15;
545          return 0;          break;
546          }          }
547        c = bracount - (c - 1);        c = bracount - (c - 1);
548        }        }
# Line 530  else Line 571  else
571        c -= '0';        c -= '0';
572        while ((digitab[ptr[1]] & ctype_digit) != 0)        while ((digitab[ptr[1]] & ctype_digit) != 0)
573          c = c * 10 + *(++ptr) - '0';          c = c * 10 + *(++ptr) - '0';
574          if (c < 0)
575            {
576            *errorcodeptr = ERR61;
577            break;
578            }
579        if (c < 10 || c <= bracount)        if (c < 10 || c <= bracount)
580          {          {
581          c = -(ESC_REF + c);          c = -(ESC_REF + c);
# Line 625  else Line 671  else
671      if (c == 0)      if (c == 0)
672        {        {
673        *errorcodeptr = ERR2;        *errorcodeptr = ERR2;
674        return 0;        break;
675        }        }
676    
677  #ifndef EBCDIC  /* ASCII coding */  #ifndef EBCDIC  /* ASCII coding */
# Line 701  if (c == '{') Line 747  if (c == '{')
747      *negptr = TRUE;      *negptr = TRUE;
748      ptr++;      ptr++;
749      }      }
750    for (i = 0; i < sizeof(name) - 1; i++)    for (i = 0; i < (int)sizeof(name) - 1; i++)
751      {      {
752      c = *(++ptr);      c = *(++ptr);
753      if (c == 0) goto ERROR_RETURN;      if (c == 0) goto ERROR_RETURN;
# Line 904  for (; *ptr != 0; ptr++) Line 950  for (; *ptr != 0; ptr++)
950      {      {
951      while (*(++ptr) != ']')      while (*(++ptr) != ']')
952        {        {
953          if (*ptr == 0) return -1;
954        if (*ptr == '\\')        if (*ptr == '\\')
955          {          {
956          if (*(++ptr) == 0) return -1;          if (*(++ptr) == 0) return -1;
# Line 931  for (; *ptr != 0; ptr++) Line 978  for (; *ptr != 0; ptr++)
978    /* An opening parens must now be a real metacharacter */    /* An opening parens must now be a real metacharacter */
979    
980    if (*ptr != '(') continue;    if (*ptr != '(') continue;
981    if (ptr[1] != '?')    if (ptr[1] != '?' && ptr[1] != '*')
982      {      {
983      count++;      count++;
984      if (name == NULL && count == lorn) return count;      if (name == NULL && count == lorn) return count;
# Line 1399  for (code = first_significant_code(code Line 1446  for (code = first_significant_code(code
1446    
1447    /* For other groups, scan the branches. */    /* For other groups, scan the branches. */
1448    
1449    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE)    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE || c == OP_COND)
1450      {      {
1451      BOOL empty_branch;      BOOL empty_branch;
1452      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */
# Line 1423  for (code = first_significant_code(code Line 1470  for (code = first_significant_code(code
1470    
1471    switch (c)    switch (c)
1472      {      {
1473      /* Check for quantifiers after a class */      /* Check for quantifiers after a class. XCLASS is used for classes that
1474        cannot be represented just by a bit map. This includes negated single
1475        high-valued characters. The length in _pcre_OP_lengths[] is zero; the
1476        actual length is stored in the compiled code, so we must update "code"
1477        here. */
1478    
1479  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1480      case OP_XCLASS:      case OP_XCLASS:
1481      ccode = code + GET(code, 1);      ccode = code += GET(code, 1);
1482      goto CHECK_CLASS_REPEAT;      goto CHECK_CLASS_REPEAT;
1483  #endif  #endif
1484    
# Line 1949  if (next >= 0) switch(op_code) Line 2000  if (next >= 0) switch(op_code)
2000    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2001    return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;    return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;
2002    
2003      case OP_HSPACE:
2004      case OP_NOT_HSPACE:
2005      switch(next)
2006        {
2007        case 0x09:
2008        case 0x20:
2009        case 0xa0:
2010        case 0x1680:
2011        case 0x180e:
2012        case 0x2000:
2013        case 0x2001:
2014        case 0x2002:
2015        case 0x2003:
2016        case 0x2004:
2017        case 0x2005:
2018        case 0x2006:
2019        case 0x2007:
2020        case 0x2008:
2021        case 0x2009:
2022        case 0x200A:
2023        case 0x202f:
2024        case 0x205f:
2025        case 0x3000:
2026        return op_code != OP_HSPACE;
2027        default:
2028        return op_code == OP_HSPACE;
2029        }
2030    
2031      case OP_VSPACE:
2032      case OP_NOT_VSPACE:
2033      switch(next)
2034        {
2035        case 0x0a:
2036        case 0x0b:
2037        case 0x0c:
2038        case 0x0d:
2039        case 0x85:
2040        case 0x2028:
2041        case 0x2029:
2042        return op_code != OP_VSPACE;
2043        default:
2044        return op_code == OP_VSPACE;
2045        }
2046    
2047    default:    default:
2048    return FALSE;    return FALSE;
2049    }    }
# Line 1983  switch(op_code) Line 2078  switch(op_code)
2078      case ESC_W:      case ESC_W:
2079      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;
2080    
2081        case ESC_h:
2082        case ESC_H:
2083        switch(item)
2084          {
2085          case 0x09:
2086          case 0x20:
2087          case 0xa0:
2088          case 0x1680:
2089          case 0x180e:
2090          case 0x2000:
2091          case 0x2001:
2092          case 0x2002:
2093          case 0x2003:
2094          case 0x2004:
2095          case 0x2005:
2096          case 0x2006:
2097          case 0x2007:
2098          case 0x2008:
2099          case 0x2009:
2100          case 0x200A:
2101          case 0x202f:
2102          case 0x205f:
2103          case 0x3000:
2104          return -next != ESC_h;
2105          default:
2106          return -next == ESC_h;
2107          }
2108    
2109        case ESC_v:
2110        case ESC_V:
2111        switch(item)
2112          {
2113          case 0x0a:
2114          case 0x0b:
2115          case 0x0c:
2116          case 0x0d:
2117          case 0x85:
2118          case 0x2028:
2119          case 0x2029:
2120          return -next != ESC_v;
2121          default:
2122          return -next == ESC_v;
2123          }
2124    
2125      default:      default:
2126      return FALSE;      return FALSE;
2127      }      }
2128    
2129    case OP_DIGIT:    case OP_DIGIT:
2130    return next == -ESC_D || next == -ESC_s || next == -ESC_W;    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
2131             next == -ESC_h || next == -ESC_v;
2132    
2133    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
2134    return next == -ESC_d;    return next == -ESC_d;
# Line 1997  switch(op_code) Line 2137  switch(op_code)
2137    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return next == -ESC_S || next == -ESC_d || next == -ESC_w;
2138    
2139    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
2140    return next == -ESC_s;    return next == -ESC_s || next == -ESC_h || next == -ESC_v;
2141    
2142      case OP_HSPACE:
2143      return next == -ESC_S || next == -ESC_H || next == -ESC_d || next == -ESC_w;
2144    
2145      case OP_NOT_HSPACE:
2146      return next == -ESC_h;
2147    
2148      /* Can't have \S in here because VT matches \S (Perl anomaly) */
2149      case OP_VSPACE:
2150      return next == -ESC_V || next == -ESC_d || next == -ESC_w;
2151    
2152      case OP_NOT_VSPACE:
2153      return next == -ESC_v;
2154    
2155    case OP_WORDCHAR:    case OP_WORDCHAR:
2156    return next == -ESC_W || next == -ESC_s;    return next == -ESC_W || next == -ESC_s || next == -ESC_h || next == -ESC_v;
2157    
2158    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2159    return next == -ESC_w || next == -ESC_d;    return next == -ESC_w || next == -ESC_d;
# Line 2115  for (;; ptr++) Line 2268  for (;; ptr++)
2268    BOOL possessive_quantifier;    BOOL possessive_quantifier;
2269    BOOL is_quantifier;    BOOL is_quantifier;
2270    BOOL is_recurse;    BOOL is_recurse;
2271    BOOL reset_bracount;    BOOL reset_bracount;
2272    int class_charcount;    int class_charcount;
2273    int class_lastchar;    int class_lastchar;
2274    int newoptions;    int newoptions;
# Line 2153  for (;; ptr++) Line 2306  for (;; ptr++)
2306      */      */
2307    
2308      if (code < last_code) code = last_code;      if (code < last_code) code = last_code;
2309    
2310        /* Paranoid check for integer overflow */
2311    
2312        if (OFLOW_MAX - *lengthptr < code - last_code)
2313          {
2314          *errorcodeptr = ERR20;
2315          goto FAILED;
2316          }
2317    
2318      *lengthptr += code - last_code;      *lengthptr += code - last_code;
2319      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));
2320    
# Line 2265  for (;; ptr++) Line 2427  for (;; ptr++)
2427      *ptrptr = ptr;      *ptrptr = ptr;
2428      if (lengthptr != NULL)      if (lengthptr != NULL)
2429        {        {
2430          if (OFLOW_MAX - *lengthptr < code - last_code)
2431            {
2432            *errorcodeptr = ERR20;
2433            goto FAILED;
2434            }
2435        *lengthptr += code - last_code;   /* To include callout length */        *lengthptr += code - last_code;   /* To include callout length */
2436        DPRINTF((">> end branch\n"));        DPRINTF((">> end branch\n"));
2437        }        }
# Line 2327  for (;; ptr++) Line 2494  for (;; ptr++)
2494        goto FAILED;        goto FAILED;
2495        }        }
2496    
2497      /* If the first character is '^', set the negation flag and skip it. */      /* If the first character is '^', set the negation flag and skip it. Also,
2498        if the first few characters (either before or after ^) are \Q\E or \E we
2499        skip them too. This makes for compatibility with Perl. */
2500    
2501      if ((c = *(++ptr)) == '^')      negate_class = FALSE;
2502        for (;;)
2503        {        {
       negate_class = TRUE;  
2504        c = *(++ptr);        c = *(++ptr);
2505        }        if (c == '\\')
2506      else          {
2507        {          if (ptr[1] == 'E') ptr++;
2508        negate_class = FALSE;            else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3;
2509                else break;
2510            }
2511          else if (!negate_class && c == '^')
2512            negate_class = TRUE;
2513          else break;
2514        }        }
2515    
2516      /* Keep a count of chars with values < 256 so that we can optimize the case      /* Keep a count of chars with values < 256 so that we can optimize the case
# Line 2477  for (;; ptr++) Line 2651  for (;; ptr++)
2651        of the specials, which just set a flag. The sequence \b is a special        of the specials, which just set a flag. The sequence \b is a special
2652        case. Inside a class (and only there) it is treated as backspace.        case. Inside a class (and only there) it is treated as backspace.
2653        Elsewhere it marks a word boundary. Other escapes have preset maps ready        Elsewhere it marks a word boundary. Other escapes have preset maps ready
2654        to or into the one we are building. We assume they have more than one        to 'or' into the one we are building. We assume they have more than one
2655        character in them, so set class_charcount bigger than one. */        character in them, so set class_charcount bigger than one. */
2656    
2657        if (c == '\\')        if (c == '\\')
# Line 2535  for (;; ptr++) Line 2709  for (;; ptr++)
2709    
2710              case ESC_E: /* Perl ignores an orphan \E */              case ESC_E: /* Perl ignores an orphan \E */
2711              continue;              continue;
2712    
2713              default:    /* Not recognized; fall through */              default:    /* Not recognized; fall through */
2714              break;      /* Need "default" setting to stop compiler warning. */              break;      /* Need "default" setting to stop compiler warning. */
2715              }              }
# Line 2544  for (;; ptr++) Line 2718  for (;; ptr++)
2718    
2719            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||
2720                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;
2721    
2722            /* We need to deal with \H, \h, \V, and \v in both phases because            /* We need to deal with \H, \h, \V, and \v in both phases because
2723            they use extra memory. */            they use extra memory. */
2724    
2725            if (-c == ESC_h)            if (-c == ESC_h)
2726              {              {
2727              SETBIT(classbits, 0x09); /* VT */              SETBIT(classbits, 0x09); /* VT */
2728              SETBIT(classbits, 0x20); /* SPACE */              SETBIT(classbits, 0x20); /* SPACE */
2729              SETBIT(classbits, 0xa0); /* NSBP */              SETBIT(classbits, 0xa0); /* NSBP */
2730  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2731              if (utf8)              if (utf8)
2732                {                {
2733                class_utf8 = TRUE;                class_utf8 = TRUE;
2734                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2735                class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);
2736                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2737                class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);
2738                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2739                class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);
2740                class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);
2741                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2742                class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);
2743                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2744                class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);
2745                *class_utf8data++ = XCL_SINGLE;                *class_utf8data++ = XCL_SINGLE;
2746                class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);
2747                }                }
2748  #endif  #endif
2749              continue;              continue;
2750              }              }
2751    
2752            if (-c == ESC_H)            if (-c == ESC_H)
2753              {              {
# Line 2581  for (;; ptr++) Line 2755  for (;; ptr++)
2755                {                {
2756                int x = 0xff;                int x = 0xff;
2757                switch (c)                switch (c)
2758                  {                  {
2759                  case 0x09/8: x ^= 1 << (0x09%8); break;                  case 0x09/8: x ^= 1 << (0x09%8); break;
2760                  case 0x20/8: x ^= 1 << (0x20%8); break;                  case 0x20/8: x ^= 1 << (0x20%8); break;
2761                  case 0xa0/8: x ^= 1 << (0xa0%8); break;                  case 0xa0/8: x ^= 1 << (0xa0%8); break;
2762                  default: break;                  default: break;
2763                  }                  }
2764                classbits[c] |= x;                classbits[c] |= x;
2765                }                }
2766    
2767  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2768              if (utf8)              if (utf8)
2769                {                {
2770                class_utf8 = TRUE;                class_utf8 = TRUE;
2771                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2772                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2773                class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);
2774                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2775                class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);
2776                class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);
2777                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2778                class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);
2779                class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);
2780                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2781                class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);
2782                class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);
2783                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2784                class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);
2785                class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);
2786                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2787                class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);
2788                class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);
2789                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2790                class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);
2791                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2792                }                }
2793  #endif  #endif
2794              continue;              continue;
2795              }              }
2796    
2797            if (-c == ESC_v)            if (-c == ESC_v)
2798              {              {
2799              SETBIT(classbits, 0x0a); /* LF */              SETBIT(classbits, 0x0a); /* LF */
2800              SETBIT(classbits, 0x0b); /* VT */              SETBIT(classbits, 0x0b); /* VT */
2801              SETBIT(classbits, 0x0c); /* FF */              SETBIT(classbits, 0x0c); /* FF */
2802              SETBIT(classbits, 0x0d); /* CR */              SETBIT(classbits, 0x0d); /* CR */
2803              SETBIT(classbits, 0x85); /* NEL */              SETBIT(classbits, 0x85); /* NEL */
2804  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2805              if (utf8)              if (utf8)
2806                {                {
2807                class_utf8 = TRUE;                class_utf8 = TRUE;
2808                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2809                class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);
2810                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2811                }                }
2812  #endif  #endif
2813              continue;              continue;
2814              }              }
2815    
2816            if (-c == ESC_V)            if (-c == ESC_V)
2817              {              {
# Line 2645  for (;; ptr++) Line 2819  for (;; ptr++)
2819                {                {
2820                int x = 0xff;                int x = 0xff;
2821                switch (c)                switch (c)
2822                  {                  {
2823                  case 0x0a/8: x ^= 1 << (0x0a%8);                  case 0x0a/8: x ^= 1 << (0x0a%8);
2824                               x ^= 1 << (0x0b%8);                               x ^= 1 << (0x0b%8);
2825                               x ^= 1 << (0x0c%8);                               x ^= 1 << (0x0c%8);
2826                               x ^= 1 << (0x0d%8);                               x ^= 1 << (0x0d%8);
2827                               break;                               break;
2828                  case 0x85/8: x ^= 1 << (0x85%8); break;                  case 0x85/8: x ^= 1 << (0x85%8); break;
2829                  default: break;                  default: break;
2830                  }                  }
2831                classbits[c] |= x;                classbits[c] |= x;
2832                }                }
2833    
2834  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2835              if (utf8)              if (utf8)
2836                {                {
2837                class_utf8 = TRUE;                class_utf8 = TRUE;
2838                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2839                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2840                class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);
2841                *class_utf8data++ = XCL_RANGE;                *class_utf8data++ = XCL_RANGE;
2842                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2843                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);                class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2844                }                }
2845  #endif  #endif
2846              continue;              continue;
2847              }              }
2848    
2849            /* We need to deal with \P and \p in both phases. */            /* We need to deal with \P and \p in both phases. */
2850    
# Line 2812  for (;; ptr++) Line 2986  for (;; ptr++)
2986              unsigned int origd = d;              unsigned int origd = d;
2987              while (get_othercase_range(&cc, origd, &occ, &ocd))              while (get_othercase_range(&cc, origd, &occ, &ocd))
2988                {                {
2989                if (occ >= (unsigned int)c &&                if (occ >= (unsigned int)c &&
2990                    ocd <= (unsigned int)d)                    ocd <= (unsigned int)d)
2991                  continue;                          /* Skip embedded ranges */                  continue;                          /* Skip embedded ranges */
2992    
2993                if (occ < (unsigned int)c  &&                if (occ < (unsigned int)c  &&
2994                    ocd >= (unsigned int)c - 1)      /* Extend the basic range */                    ocd >= (unsigned int)c - 1)      /* Extend the basic range */
2995                  {                                  /* if there is overlap,   */                  {                                  /* if there is overlap,   */
2996                  c = occ;                           /* noting that if occ < c */                  c = occ;                           /* noting that if occ < c */
2997                  continue;                          /* we can't have ocd > d  */                  continue;                          /* we can't have ocd > d  */
2998                  }                                  /* because a subrange is  */                  }                                  /* because a subrange is  */
2999                if (ocd > (unsigned int)d &&                if (ocd > (unsigned int)d &&
3000                    occ <= (unsigned int)d + 1)      /* always shorter than    */                    occ <= (unsigned int)d + 1)      /* always shorter than    */
3001                  {                                  /* the basic range.       */                  {                                  /* the basic range.       */
3002                  d = ocd;                  d = ocd;
# Line 3419  for (;; ptr++) Line 3593  for (;; ptr++)
3593          goto FAILED;          goto FAILED;
3594          }          }
3595    
       /* This is a paranoid check to stop integer overflow later on */  
   
       if (len > MAX_DUPLENGTH)  
         {  
         *errorcodeptr = ERR50;  
         goto FAILED;  
         }  
   
3596        /* If the maximum repeat count is unlimited, find the end of the bracket        /* If the maximum repeat count is unlimited, find the end of the bracket
3597        by scanning through from the start, and compute the offset back to it        by scanning through from the start, and compute the offset back to it
3598        from the current code pointer. There may be an OP_OPT setting following        from the current code pointer. There may be an OP_OPT setting following
# Line 3515  for (;; ptr++) Line 3681  for (;; ptr++)
3681          if (repeat_min > 1)          if (repeat_min > 1)
3682            {            {
3683            /* In the pre-compile phase, we don't actually do the replication. We            /* In the pre-compile phase, we don't actually do the replication. We
3684            just adjust the length as if we had. */            just adjust the length as if we had. Do some paranoid checks for
3685              potential integer overflow. */
3686    
3687            if (lengthptr != NULL)            if (lengthptr != NULL)
3688              *lengthptr += (repeat_min - 1)*length_prevgroup;              {
3689                int delta = (repeat_min - 1)*length_prevgroup;
3690                if ((double)(repeat_min - 1)*(double)length_prevgroup >
3691                                                                (double)INT_MAX ||
3692                    OFLOW_MAX - *lengthptr < delta)
3693                  {
3694                  *errorcodeptr = ERR20;
3695                  goto FAILED;
3696                  }
3697                *lengthptr += delta;
3698                }
3699    
3700            /* This is compiling for real */            /* This is compiling for real */
3701    
# Line 3556  for (;; ptr++) Line 3733  for (;; ptr++)
3733          /* In the pre-compile phase, we don't actually do the replication. We          /* In the pre-compile phase, we don't actually do the replication. We
3734          just adjust the length as if we had. For each repetition we must add 1          just adjust the length as if we had. For each repetition we must add 1
3735          to the length for BRAZERO and for all but the last repetition we must          to the length for BRAZERO and for all but the last repetition we must
3736          add 2 + 2*LINKSIZE to allow for the nesting that occurs. */          add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some
3737            paranoid checks to avoid integer overflow. */
3738    
3739          if (lengthptr != NULL && repeat_max > 0)          if (lengthptr != NULL && repeat_max > 0)
3740            *lengthptr += repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -            {
3741              2 - 2*LINK_SIZE;  /* Last one doesn't nest */            int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -
3742                          2 - 2*LINK_SIZE;   /* Last one doesn't nest */
3743              if ((double)repeat_max *
3744                    (double)(length_prevgroup + 1 + 2 + 2*LINK_SIZE)
3745                      > (double)INT_MAX ||
3746                  OFLOW_MAX - *lengthptr < delta)
3747                {
3748                *errorcodeptr = ERR20;
3749                goto FAILED;
3750                }
3751              *lengthptr += delta;
3752              }
3753    
3754          /* This is compiling for real */          /* This is compiling for real */
3755    
# Line 3712  for (;; ptr++) Line 3901  for (;; ptr++)
3901      /* ===================================================================*/      /* ===================================================================*/
3902      /* Start of nested parenthesized sub-expression, or comment or lookahead or      /* Start of nested parenthesized sub-expression, or comment or lookahead or
3903      lookbehind or option setting or condition or all the other extended      lookbehind or option setting or condition or all the other extended
3904      parenthesis forms. First deal with the specials; all are introduced by ?,      parenthesis forms.  */
     and the appearance of any of them means that this is not a capturing  
     group. */  
3905    
3906      case '(':      case '(':
3907      newoptions = options;      newoptions = options;
3908      skipbytes = 0;      skipbytes = 0;
3909      bravalue = OP_CBRA;      bravalue = OP_CBRA;
3910      save_hwm = cd->hwm;      save_hwm = cd->hwm;
3911      reset_bracount = FALSE;      reset_bracount = FALSE;
3912    
3913        /* First deal with various "verbs" that can be introduced by '*'. */
3914    
3915      if (*(++ptr) == '?')      if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0)
3916          {
3917          int i, namelen;
3918          const uschar *name = ++ptr;
3919          previous = NULL;
3920          while ((cd->ctypes[*++ptr] & ctype_letter) != 0);
3921          if (*ptr == ':')
3922            {
3923            *errorcodeptr = ERR59;   /* Not supported */
3924            goto FAILED;
3925            }
3926          if (*ptr != ')')
3927            {
3928            *errorcodeptr = ERR60;
3929            goto FAILED;
3930            }
3931          namelen = ptr - name;
3932          for (i = 0; i < verbcount; i++)
3933            {
3934            if (namelen == verbs[i].len &&
3935                strncmp((char *)name, verbs[i].name, namelen) == 0)
3936              {
3937              *code = verbs[i].op;
3938              if (*code++ == OP_ACCEPT) cd->had_accept = TRUE;
3939              break;
3940              }
3941            }
3942          if (i < verbcount) continue;
3943          *errorcodeptr = ERR60;
3944          goto FAILED;
3945          }
3946    
3947        /* Deal with the extended parentheses; all are introduced by '?', and the
3948        appearance of any of them means that this is not a capturing group. */
3949    
3950        else if (*ptr == '?')
3951        {        {
3952        int i, set, unset, namelen;        int i, set, unset, namelen;
3953        int *optset;        int *optset;
# Line 3746  for (;; ptr++) Line 3970  for (;; ptr++)
3970          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
3971          case '|':                 /* Reset capture count for each branch */          case '|':                 /* Reset capture count for each branch */
3972          reset_bracount = TRUE;          reset_bracount = TRUE;
3973          /* Fall through */          /* Fall through */
3974    
3975          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
3976          case ':':                 /* Non-capturing bracket */          case ':':                 /* Non-capturing bracket */
# Line 3965  for (;; ptr++) Line 4189  for (;; ptr++)
4189    
4190          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4191          case '!':                 /* Negative lookahead */          case '!':                 /* Negative lookahead */
         bravalue = OP_ASSERT_NOT;  
4192          ptr++;          ptr++;
4193            if (*ptr == ')')          /* Optimize (?!) */
4194              {
4195              *code++ = OP_FAIL;
4196              previous = NULL;
4197              continue;
4198              }
4199            bravalue = OP_ASSERT_NOT;
4200          break;          break;
4201    
4202    
# Line 4447  for (;; ptr++) Line 4677  for (;; ptr++)
4677           errorcodeptr,                 /* Where to put an error message */           errorcodeptr,                 /* Where to put an error message */
4678           (bravalue == OP_ASSERTBACK ||           (bravalue == OP_ASSERTBACK ||
4679            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
4680           reset_bracount,               /* True if (?| group */           reset_bracount,               /* True if (?| group */
4681           skipbytes,                    /* Skip over bracket number */           skipbytes,                    /* Skip over bracket number */
4682           &subfirstbyte,                /* For possible first char */           &subfirstbyte,                /* For possible first char */
4683           &subreqbyte,                  /* For possible last char */           &subreqbyte,                  /* For possible last char */
# Line 4522  for (;; ptr++) Line 4752  for (;; ptr++)
4752    
4753      if (lengthptr != NULL)      if (lengthptr != NULL)
4754        {        {
4755          if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE)
4756            {
4757            *errorcodeptr = ERR20;
4758            goto FAILED;
4759            }
4760        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;
4761        code++;        code++;
4762        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
# Line 4807  Arguments: Line 5042  Arguments:
5042    ptrptr         -> the address of the current pattern pointer    ptrptr         -> the address of the current pattern pointer
5043    errorcodeptr   -> pointer to error code variable    errorcodeptr   -> pointer to error code variable
5044    lookbehind     TRUE if this is a lookbehind assertion    lookbehind     TRUE if this is a lookbehind assertion
5045    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
5046    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
5047    firstbyteptr   place to put the first required character, or a negative number    firstbyteptr   place to put the first required character, or a negative number
5048    reqbyteptr     place to put the last required character, or a negative number    reqbyteptr     place to put the last required character, or a negative number
# Line 4821  Returns:         TRUE on success Line 5056  Returns:         TRUE on success
5056    
5057  static BOOL  static BOOL
5058  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,
5059    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
5060    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,
5061    int *lengthptr)    int *lengthptr)
5062  {  {
5063  const uschar *ptr = *ptrptr;  const uschar *ptr = *ptrptr;
# Line 4867  orig_bracount = max_bracount = cd->braco Line 5102  orig_bracount = max_bracount = cd->braco
5102  for (;;)  for (;;)
5103    {    {
5104    /* For a (?| group, reset the capturing bracket count so that each branch    /* For a (?| group, reset the capturing bracket count so that each branch
5105    uses the same numbers. */    uses the same numbers. */
5106    
5107    if (reset_bracount) cd->bracount = orig_bracount;    if (reset_bracount) cd->bracount = orig_bracount;
5108    
5109    /* Handle a change of ims options at the start of the branch */    /* Handle a change of ims options at the start of the branch */
5110    
5111    if ((options & PCRE_IMS) != oldims)    if ((options & PCRE_IMS) != oldims)
# Line 4899  for (;;) Line 5134  for (;;)
5134      *ptrptr = ptr;      *ptrptr = ptr;
5135      return FALSE;      return FALSE;
5136      }      }
5137    
5138    /* Keep the highest bracket count in case (?| was used and some branch    /* Keep the highest bracket count in case (?| was used and some branch
5139    has fewer than the rest. */    has fewer than the rest. */
5140    
5141    if (cd->bracount > max_bracount) max_bracount = cd->bracount;    if (cd->bracount > max_bracount) max_bracount = cd->bracount;
5142    
5143    /* In the real compile phase, there is some post-processing to be done. */    /* In the real compile phase, there is some post-processing to be done. */
# Line 5006  for (;;) Line 5241  for (;;)
5241        *code++ = oldims;        *code++ = oldims;
5242        length += 2;        length += 2;
5243        }        }
5244    
5245      /* Retain the highest bracket number, in case resetting was used. */      /* Retain the highest bracket number, in case resetting was used. */
5246    
5247      cd->bracount = max_bracount;      cd->bracount = max_bracount;
5248    
5249      /* Set values to pass back */      /* Set values to pass back */
# Line 5017  for (;;) Line 5252  for (;;)
5252      *ptrptr = ptr;      *ptrptr = ptr;
5253      *firstbyteptr = firstbyte;      *firstbyteptr = firstbyte;
5254      *reqbyteptr = reqbyte;      *reqbyteptr = reqbyte;
5255      if (lengthptr != NULL) *lengthptr += length;      if (lengthptr != NULL)
5256          {
5257          if (OFLOW_MAX - *lengthptr < length)
5258            {
5259            *errorcodeptr = ERR20;
5260            return FALSE;
5261            }
5262          *lengthptr += length;
5263          }
5264      return TRUE;      return TRUE;
5265      }      }
5266    
# Line 5485  outside can help speed up starting point Line 5728  outside can help speed up starting point
5728  code = cworkspace;  code = cworkspace;
5729  *code = OP_BRA;  *code = OP_BRA;
5730  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,
5731    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,
5732    &length);    &length);
5733  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
5734    
# Line 5545  cd->start_code = codestart; Line 5788  cd->start_code = codestart;
5788  cd->hwm = cworkspace;  cd->hwm = cworkspace;
5789  cd->req_varyopt = 0;  cd->req_varyopt = 0;
5790  cd->nopartial = FALSE;  cd->nopartial = FALSE;
5791    cd->had_accept = FALSE;
5792    
5793  /* Set up a starting, non-extracting bracket, then compile the expression. On  /* Set up a starting, non-extracting bracket, then compile the expression. On
5794  error, errorcode will be set non-zero, so we don't need to look at the result  error, errorcode will be set non-zero, so we don't need to look at the result
# Line 5559  re->top_bracket = cd->bracount; Line 5803  re->top_bracket = cd->bracount;
5803  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
5804    
5805  if (cd->nopartial) re->options |= PCRE_NOPARTIAL;  if (cd->nopartial) re->options |= PCRE_NOPARTIAL;
5806    if (cd->had_accept) reqbyte = -1;   /* Must disable after (*ACCEPT) */
5807    
5808  /* If not reached end of pattern on success, there's an excess bracket. */  /* If not reached end of pattern on success, there's an excess bracket. */
5809    

Legend:
Removed from v.178  
changed lines
  Added in v.216

  ViewVC Help
Powered by ViewVC 1.1.5