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

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

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

revision 175 by ph10, Mon Jun 11 13:38:38 2007 UTC revision 206 by ph10, Fri Aug 3 14:53:04 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 58  used by pcretest. DEBUG is not defined w Line 61  used by pcretest. DEBUG is not defined w
61  #endif  #endif
62    
63    
64    /* Macro for setting individual bits in class bitmaps. */
65    
66    #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         *
78  *************************************************/  *************************************************/
# Line 87  static const short int escapes[] = { Line 102  static const short int escapes[] = {
102       0,      0,      0,      0,      0,      0,      0,      0,   /* 0 - 7 */       0,      0,      0,      0,      0,      0,      0,      0,   /* 0 - 7 */
103       0,      0,    ':',    ';',    '<',    '=',    '>',    '?',   /* 8 - ? */       0,      0,    ':',    ';',    '<',    '=',    '>',    '?',   /* 8 - ? */
104     '@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E,      0, -ESC_G,   /* @ - G */     '@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E,      0, -ESC_G,   /* @ - G */
105       0,      0,      0, -ESC_K,      0,      0,      0,      0,   /* H - O */  -ESC_H,      0,      0, -ESC_K,      0,      0,      0,      0,   /* H - O */
106  -ESC_P, -ESC_Q, -ESC_R, -ESC_S,      0,      0,      0, -ESC_W,   /* P - W */  -ESC_P, -ESC_Q, -ESC_R, -ESC_S,      0,      0, -ESC_V, -ESC_W,   /* P - W */
107  -ESC_X,      0, -ESC_Z,    '[',   '\\',    ']',    '^',    '_',   /* X - _ */  -ESC_X,      0, -ESC_Z,    '[',   '\\',    ']',    '^',    '_',   /* X - _ */
108     '`',      7, -ESC_b,      0, -ESC_d,  ESC_e,  ESC_f,      0,   /* ` - g */     '`',      7, -ESC_b,      0, -ESC_d,  ESC_e,  ESC_f,      0,   /* ` - g */
109       0,      0,      0, -ESC_k,      0,      0,  ESC_n,      0,   /* h - o */  -ESC_h,      0,      0, -ESC_k,      0,      0,  ESC_n,      0,   /* h - o */
110  -ESC_p,      0,  ESC_r, -ESC_s,  ESC_tee,    0,      0, -ESC_w,   /* p - w */  -ESC_p,      0,  ESC_r, -ESC_s,  ESC_tee,    0, -ESC_v, -ESC_w,   /* p - w */
111       0,      0, -ESC_z                                            /* x - z */       0,      0, -ESC_z                                            /* x - z */
112  };  };
113    
# Line 106  static const short int escapes[] = { Line 121  static const short int escapes[] = {
121  /*  70 */     0,     0,      0,       0,      0,     0,      0,      0,  /*  70 */     0,     0,      0,       0,      0,     0,      0,      0,
122  /*  78 */     0,   '`',    ':',     '#',    '@',  '\'',    '=',    '"',  /*  78 */     0,   '`',    ':',     '#',    '@',  '\'',    '=',    '"',
123  /*  80 */     0,     7, -ESC_b,       0, -ESC_d, ESC_e,  ESC_f,      0,  /*  80 */     0,     7, -ESC_b,       0, -ESC_d, ESC_e,  ESC_f,      0,
124  /*  88 */     0,     0,      0,     '{',      0,     0,      0,      0,  /*  88 */-ESC_h,     0,      0,     '{',      0,     0,      0,      0,
125  /*  90 */     0,     0, -ESC_k,     'l',      0, ESC_n,      0, -ESC_p,  /*  90 */     0,     0, -ESC_k,     'l',      0, ESC_n,      0, -ESC_p,
126  /*  98 */     0, ESC_r,      0,     '}',      0,     0,      0,      0,  /*  98 */     0, ESC_r,      0,     '}',      0,     0,      0,      0,
127  /*  A0 */     0,   '~', -ESC_s, ESC_tee,      0,     0, -ESC_w,      0,  /*  A0 */     0,   '~', -ESC_s, ESC_tee,      0,-ESC_v, -ESC_w,      0,
128  /*  A8 */     0,-ESC_z,      0,       0,      0,   '[',      0,      0,  /*  A8 */     0,-ESC_z,      0,       0,      0,   '[',      0,      0,
129  /*  B0 */     0,     0,      0,       0,      0,     0,      0,      0,  /*  B0 */     0,     0,      0,       0,      0,     0,      0,      0,
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 */     0,     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,     0, -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,
137  /*  F0 */     0,     0,      0,       0,      0,     0,      0,      0,  /*  F0 */     0,     0,      0,       0,      0,     0,      0,      0,
138  /*  F8 */     0,     0,      0,       0,      0,     0,      0,      0  /*  F8 */     0,     0,      0,       0,      0,     0,      0,      0
# Line 198  static const char *error_texts[] = { Line 213  static const char *error_texts[] = {
213    "missing ) after comment",    "missing ) after comment",
214    "parentheses nested too deeply",  /** DEAD **/    "parentheses nested too deeply",  /** DEAD **/
215    /* 20 */    /* 20 */
216    "regular expression too large",    "regular expression is too large",
217    "failed to get memory",    "failed to get memory",
218    "unmatched parentheses",    "unmatched parentheses",
219    "internal error: code overflow",    "internal error: code overflow",
# Line 234  static const char *error_texts[] = { Line 249  static const char *error_texts[] = {
249    "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)",    "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)",
250    "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")",    "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")",
251    /* 50 */    /* 50 */
252    "repeated subpattern is too long",    "repeated subpattern is too long",    /** DEAD **/
253    "octal value is greater than \\377 (not in UTF-8 mode)",    "octal value is greater than \\377 (not in UTF-8 mode)",
254    "internal error: overran compiling workspace",    "internal error: overran compiling workspace",
255    "internal error: previously-checked referenced subpattern not found",    "internal error: previously-checked referenced subpattern not found",
# Line 374  static const unsigned char ebcdic_charta Line 389  static const unsigned char ebcdic_charta
389  /* Definition to allow mutual recursion */  /* Definition to allow mutual recursion */
390    
391  static BOOL  static BOOL
392    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,
393      int *, int *, branch_chain *, compile_data *, int *);      int *, int *, branch_chain *, compile_data *, int *);
394    
395    
# Line 696  if (c == '{') Line 711  if (c == '{')
711      *negptr = TRUE;      *negptr = TRUE;
712      ptr++;      ptr++;
713      }      }
714    for (i = 0; i < sizeof(name) - 1; i++)    for (i = 0; i < (int)sizeof(name) - 1; i++)
715      {      {
716      c = *(++ptr);      c = *(++ptr);
717      if (c == 0) goto ERROR_RETURN;      if (c == 0) goto ERROR_RETURN;
# Line 1394  for (code = first_significant_code(code Line 1409  for (code = first_significant_code(code
1409    
1410    /* For other groups, scan the branches. */    /* For other groups, scan the branches. */
1411    
1412    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE)    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE || c == OP_COND)
1413      {      {
1414      BOOL empty_branch;      BOOL empty_branch;
1415      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */
# Line 1944  if (next >= 0) switch(op_code) Line 1959  if (next >= 0) switch(op_code)
1959    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
1960    return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;    return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;
1961    
1962      case OP_HSPACE:
1963      case OP_NOT_HSPACE:
1964      switch(next)
1965        {
1966        case 0x09:
1967        case 0x20:
1968        case 0xa0:
1969        case 0x1680:
1970        case 0x180e:
1971        case 0x2000:
1972        case 0x2001:
1973        case 0x2002:
1974        case 0x2003:
1975        case 0x2004:
1976        case 0x2005:
1977        case 0x2006:
1978        case 0x2007:
1979        case 0x2008:
1980        case 0x2009:
1981        case 0x200A:
1982        case 0x202f:
1983        case 0x205f:
1984        case 0x3000:
1985        return op_code != OP_HSPACE;
1986        default:
1987        return op_code == OP_HSPACE;
1988        }
1989    
1990      case OP_VSPACE:
1991      case OP_NOT_VSPACE:
1992      switch(next)
1993        {
1994        case 0x0a:
1995        case 0x0b:
1996        case 0x0c:
1997        case 0x0d:
1998        case 0x85:
1999        case 0x2028:
2000        case 0x2029:
2001        return op_code != OP_VSPACE;
2002        default:
2003        return op_code == OP_VSPACE;
2004        }
2005    
2006    default:    default:
2007    return FALSE;    return FALSE;
2008    }    }
# Line 1978  switch(op_code) Line 2037  switch(op_code)
2037      case ESC_W:      case ESC_W:
2038      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;
2039    
2040        case ESC_h:
2041        case ESC_H:
2042        switch(item)
2043          {
2044          case 0x09:
2045          case 0x20:
2046          case 0xa0:
2047          case 0x1680:
2048          case 0x180e:
2049          case 0x2000:
2050          case 0x2001:
2051          case 0x2002:
2052          case 0x2003:
2053          case 0x2004:
2054          case 0x2005:
2055          case 0x2006:
2056          case 0x2007:
2057          case 0x2008:
2058          case 0x2009:
2059          case 0x200A:
2060          case 0x202f:
2061          case 0x205f:
2062          case 0x3000:
2063          return -next != ESC_h;
2064          default:
2065          return -next == ESC_h;
2066          }
2067    
2068        case ESC_v:
2069        case ESC_V:
2070        switch(item)
2071          {
2072          case 0x0a:
2073          case 0x0b:
2074          case 0x0c:
2075          case 0x0d:
2076          case 0x85:
2077          case 0x2028:
2078          case 0x2029:
2079          return -next != ESC_v;
2080          default:
2081          return -next == ESC_v;
2082          }
2083    
2084      default:      default:
2085      return FALSE;      return FALSE;
2086      }      }
2087    
2088    case OP_DIGIT:    case OP_DIGIT:
2089    return next == -ESC_D || next == -ESC_s || next == -ESC_W;    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
2090             next == -ESC_h || next == -ESC_v;
2091    
2092    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
2093    return next == -ESC_d;    return next == -ESC_d;
# Line 1992  switch(op_code) Line 2096  switch(op_code)
2096    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return next == -ESC_S || next == -ESC_d || next == -ESC_w;
2097    
2098    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
2099    return next == -ESC_s;    return next == -ESC_s || next == -ESC_h || next == -ESC_v;
2100    
2101      case OP_HSPACE:
2102      return next == -ESC_S || next == -ESC_H || next == -ESC_d || next == -ESC_w;
2103    
2104      case OP_NOT_HSPACE:
2105      return next == -ESC_h;
2106    
2107      /* Can't have \S in here because VT matches \S (Perl anomaly) */
2108      case OP_VSPACE:
2109      return next == -ESC_V || next == -ESC_d || next == -ESC_w;
2110    
2111      case OP_NOT_VSPACE:
2112      return next == -ESC_v;
2113    
2114    case OP_WORDCHAR:    case OP_WORDCHAR:
2115    return next == -ESC_W || next == -ESC_s;    return next == -ESC_W || next == -ESC_s || next == -ESC_h || next == -ESC_v;
2116    
2117    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2118    return next == -ESC_w || next == -ESC_d;    return next == -ESC_w || next == -ESC_d;
# Line 2110  for (;; ptr++) Line 2227  for (;; ptr++)
2227    BOOL possessive_quantifier;    BOOL possessive_quantifier;
2228    BOOL is_quantifier;    BOOL is_quantifier;
2229    BOOL is_recurse;    BOOL is_recurse;
2230    BOOL reset_bracount;    BOOL reset_bracount;
2231    int class_charcount;    int class_charcount;
2232    int class_lastchar;    int class_lastchar;
2233    int newoptions;    int newoptions;
# Line 2148  for (;; ptr++) Line 2265  for (;; ptr++)
2265      */      */
2266    
2267      if (code < last_code) code = last_code;      if (code < last_code) code = last_code;
2268    
2269        /* Paranoid check for integer overflow */
2270    
2271        if (OFLOW_MAX - *lengthptr < code - last_code)
2272          {
2273          *errorcodeptr = ERR20;
2274          goto FAILED;
2275          }
2276    
2277      *lengthptr += code - last_code;      *lengthptr += code - last_code;
2278      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));
2279    
# Line 2260  for (;; ptr++) Line 2386  for (;; ptr++)
2386      *ptrptr = ptr;      *ptrptr = ptr;
2387      if (lengthptr != NULL)      if (lengthptr != NULL)
2388        {        {
2389          if (OFLOW_MAX - *lengthptr < code - last_code)
2390            {
2391            *errorcodeptr = ERR20;
2392            goto FAILED;
2393            }
2394        *lengthptr += code - last_code;   /* To include callout length */        *lengthptr += code - last_code;   /* To include callout length */
2395        DPRINTF((">> end branch\n"));        DPRINTF((">> end branch\n"));
2396        }        }
# Line 2322  for (;; ptr++) Line 2453  for (;; ptr++)
2453        goto FAILED;        goto FAILED;
2454        }        }
2455    
2456      /* 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,
2457        if the first few characters (either before or after ^) are \Q\E or \E we
2458      if ((c = *(++ptr)) == '^')      skip them too. This makes for compatibility with Perl. */
2459    
2460        negate_class = FALSE;
2461        for (;;)
2462        {        {
       negate_class = TRUE;  
2463        c = *(++ptr);        c = *(++ptr);
2464        }        if (c == '\\')
2465      else          {
2466        {          if (ptr[1] == 'E') ptr++;
2467        negate_class = FALSE;            else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3;
2468        }              else break;
2469            }
2470          else if (!negate_class && c == '^')
2471            negate_class = TRUE;
2472          else break;
2473          }
2474    
2475      /* 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
2476      of just a single character (as long as it's < 256). However, For higher      of just a single character (as long as it's < 256). However, For higher
# Line 2472  for (;; ptr++) Line 2610  for (;; ptr++)
2610        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
2611        case. Inside a class (and only there) it is treated as backspace.        case. Inside a class (and only there) it is treated as backspace.
2612        Elsewhere it marks a word boundary. Other escapes have preset maps ready        Elsewhere it marks a word boundary. Other escapes have preset maps ready
2613        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
2614        character in them, so set class_charcount bigger than one. */        character in them, so set class_charcount bigger than one. */
2615    
2616        if (c == '\\')        if (c == '\\')
# Line 2540  for (;; ptr++) Line 2678  for (;; ptr++)
2678            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||
2679                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;
2680    
2681              /* We need to deal with \H, \h, \V, and \v in both phases because
2682              they use extra memory. */
2683    
2684              if (-c == ESC_h)
2685                {
2686                SETBIT(classbits, 0x09); /* VT */
2687                SETBIT(classbits, 0x20); /* SPACE */
2688                SETBIT(classbits, 0xa0); /* NSBP */
2689    #ifdef SUPPORT_UTF8
2690                if (utf8)
2691                  {
2692                  class_utf8 = TRUE;
2693                  *class_utf8data++ = XCL_SINGLE;
2694                  class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);
2695                  *class_utf8data++ = XCL_SINGLE;
2696                  class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);
2697                  *class_utf8data++ = XCL_RANGE;
2698                  class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);
2699                  class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);
2700                  *class_utf8data++ = XCL_SINGLE;
2701                  class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);
2702                  *class_utf8data++ = XCL_SINGLE;
2703                  class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);
2704                  *class_utf8data++ = XCL_SINGLE;
2705                  class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);
2706                  }
2707    #endif
2708                continue;
2709                }
2710    
2711              if (-c == ESC_H)
2712                {
2713                for (c = 0; c < 32; c++)
2714                  {
2715                  int x = 0xff;
2716                  switch (c)
2717                    {
2718                    case 0x09/8: x ^= 1 << (0x09%8); break;
2719                    case 0x20/8: x ^= 1 << (0x20%8); break;
2720                    case 0xa0/8: x ^= 1 << (0xa0%8); break;
2721                    default: break;
2722                    }
2723                  classbits[c] |= x;
2724                  }
2725    
2726    #ifdef SUPPORT_UTF8
2727                if (utf8)
2728                  {
2729                  class_utf8 = TRUE;
2730                  *class_utf8data++ = XCL_RANGE;
2731                  class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2732                  class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);
2733                  *class_utf8data++ = XCL_RANGE;
2734                  class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);
2735                  class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);
2736                  *class_utf8data++ = XCL_RANGE;
2737                  class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);
2738                  class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);
2739                  *class_utf8data++ = XCL_RANGE;
2740                  class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);
2741                  class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);
2742                  *class_utf8data++ = XCL_RANGE;
2743                  class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);
2744                  class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);
2745                  *class_utf8data++ = XCL_RANGE;
2746                  class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);
2747                  class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);
2748                  *class_utf8data++ = XCL_RANGE;
2749                  class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);
2750                  class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2751                  }
2752    #endif
2753                continue;
2754                }
2755    
2756              if (-c == ESC_v)
2757                {
2758                SETBIT(classbits, 0x0a); /* LF */
2759                SETBIT(classbits, 0x0b); /* VT */
2760                SETBIT(classbits, 0x0c); /* FF */
2761                SETBIT(classbits, 0x0d); /* CR */
2762                SETBIT(classbits, 0x85); /* NEL */
2763    #ifdef SUPPORT_UTF8
2764                if (utf8)
2765                  {
2766                  class_utf8 = TRUE;
2767                  *class_utf8data++ = XCL_RANGE;
2768                  class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);
2769                  class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2770                  }
2771    #endif
2772                continue;
2773                }
2774    
2775              if (-c == ESC_V)
2776                {
2777                for (c = 0; c < 32; c++)
2778                  {
2779                  int x = 0xff;
2780                  switch (c)
2781                    {
2782                    case 0x0a/8: x ^= 1 << (0x0a%8);
2783                                 x ^= 1 << (0x0b%8);
2784                                 x ^= 1 << (0x0c%8);
2785                                 x ^= 1 << (0x0d%8);
2786                                 break;
2787                    case 0x85/8: x ^= 1 << (0x85%8); break;
2788                    default: break;
2789                    }
2790                  classbits[c] |= x;
2791                  }
2792    
2793    #ifdef SUPPORT_UTF8
2794                if (utf8)
2795                  {
2796                  class_utf8 = TRUE;
2797                  *class_utf8data++ = XCL_RANGE;
2798                  class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2799                  class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);
2800                  *class_utf8data++ = XCL_RANGE;
2801                  class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2802                  class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2803                  }
2804    #endif
2805                continue;
2806                }
2807    
2808            /* We need to deal with \P and \p in both phases. */            /* We need to deal with \P and \p in both phases. */
2809    
2810  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2680  for (;; ptr++) Line 2945  for (;; ptr++)
2945              unsigned int origd = d;              unsigned int origd = d;
2946              while (get_othercase_range(&cc, origd, &occ, &ocd))              while (get_othercase_range(&cc, origd, &occ, &ocd))
2947                {                {
2948                if (occ >= c && ocd <= d) continue;  /* Skip embedded ranges */                if (occ >= (unsigned int)c &&
2949                      ocd <= (unsigned int)d)
2950                    continue;                          /* Skip embedded ranges */
2951    
2952                if (occ < c  && ocd >= c - 1)        /* Extend the basic range */                if (occ < (unsigned int)c  &&
2953                      ocd >= (unsigned int)c - 1)      /* Extend the basic range */
2954                  {                                  /* if there is overlap,   */                  {                                  /* if there is overlap,   */
2955                  c = occ;                           /* noting that if occ < c */                  c = occ;                           /* noting that if occ < c */
2956                  continue;                          /* we can't have ocd > d  */                  continue;                          /* we can't have ocd > d  */
2957                  }                                  /* because a subrange is  */                  }                                  /* because a subrange is  */
2958                if (ocd > d && occ <= d + 1)         /* always shorter than    */                if (ocd > (unsigned int)d &&
2959                      occ <= (unsigned int)d + 1)      /* always shorter than    */
2960                  {                                  /* the basic range.       */                  {                                  /* the basic range.       */
2961                  d = ocd;                  d = ocd;
2962                  continue;                  continue;
# Line 2806  for (;; ptr++) Line 3075  for (;; ptr++)
3075        *errorcodeptr = ERR6;        *errorcodeptr = ERR6;
3076        goto FAILED;        goto FAILED;
3077        }        }
3078    
3079      /* If class_charcount is 1, we saw precisely one character whose value is      /* If class_charcount is 1, we saw precisely one character whose value is
3080      less than 256. In non-UTF-8 mode we can always optimize. In UTF-8 mode, we      less than 256. In non-UTF-8 mode we can always optimize. In UTF-8 mode, we
3081      can optimize the negative case only if there were no characters >= 128      can optimize the negative case only if there were no characters >= 128
# Line 3283  for (;; ptr++) Line 3552  for (;; ptr++)
3552          goto FAILED;          goto FAILED;
3553          }          }
3554    
       /* This is a paranoid check to stop integer overflow later on */  
   
       if (len > MAX_DUPLENGTH)  
         {  
         *errorcodeptr = ERR50;  
         goto FAILED;  
         }  
   
3555        /* 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
3556        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
3557        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 3379  for (;; ptr++) Line 3640  for (;; ptr++)
3640          if (repeat_min > 1)          if (repeat_min > 1)
3641            {            {
3642            /* 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
3643            just adjust the length as if we had. */            just adjust the length as if we had. Do some paranoid checks for
3644              potential integer overflow. */
3645    
3646            if (lengthptr != NULL)            if (lengthptr != NULL)
3647              *lengthptr += (repeat_min - 1)*length_prevgroup;              {
3648                int delta = (repeat_min - 1)*length_prevgroup;
3649                if ((double)(repeat_min - 1)*(double)length_prevgroup >
3650                                                                (double)INT_MAX ||
3651                    OFLOW_MAX - *lengthptr < delta)
3652                  {
3653                  *errorcodeptr = ERR20;
3654                  goto FAILED;
3655                  }
3656                *lengthptr += delta;
3657                }
3658    
3659            /* This is compiling for real */            /* This is compiling for real */
3660    
# Line 3420  for (;; ptr++) Line 3692  for (;; ptr++)
3692          /* 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
3693          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
3694          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
3695          add 2 + 2*LINKSIZE to allow for the nesting that occurs. */          add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some
3696            paranoid checks to avoid integer overflow. */
3697    
3698          if (lengthptr != NULL && repeat_max > 0)          if (lengthptr != NULL && repeat_max > 0)
3699            *lengthptr += repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -            {
3700              2 - 2*LINK_SIZE;  /* Last one doesn't nest */            int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -
3701                          2 - 2*LINK_SIZE;   /* Last one doesn't nest */
3702              if ((double)repeat_max *
3703                    (double)(length_prevgroup + 1 + 2 + 2*LINK_SIZE)
3704                      > (double)INT_MAX ||
3705                  OFLOW_MAX - *lengthptr < delta)
3706                {
3707                *errorcodeptr = ERR20;
3708                goto FAILED;
3709                }
3710              *lengthptr += delta;
3711              }
3712    
3713          /* This is compiling for real */          /* This is compiling for real */
3714    
# Line 3585  for (;; ptr++) Line 3869  for (;; ptr++)
3869      skipbytes = 0;      skipbytes = 0;
3870      bravalue = OP_CBRA;      bravalue = OP_CBRA;
3871      save_hwm = cd->hwm;      save_hwm = cd->hwm;
3872      reset_bracount = FALSE;      reset_bracount = FALSE;
3873    
3874      if (*(++ptr) == '?')      if (*(++ptr) == '?')
3875        {        {
# Line 3610  for (;; ptr++) Line 3894  for (;; ptr++)
3894          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
3895          case '|':                 /* Reset capture count for each branch */          case '|':                 /* Reset capture count for each branch */
3896          reset_bracount = TRUE;          reset_bracount = TRUE;
3897          /* Fall through */          /* Fall through */
3898    
3899          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
3900          case ':':                 /* Non-capturing bracket */          case ':':                 /* Non-capturing bracket */
# Line 4311  for (;; ptr++) Line 4595  for (;; ptr++)
4595           errorcodeptr,                 /* Where to put an error message */           errorcodeptr,                 /* Where to put an error message */
4596           (bravalue == OP_ASSERTBACK ||           (bravalue == OP_ASSERTBACK ||
4597            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
4598           reset_bracount,               /* True if (?| group */           reset_bracount,               /* True if (?| group */
4599           skipbytes,                    /* Skip over bracket number */           skipbytes,                    /* Skip over bracket number */
4600           &subfirstbyte,                /* For possible first char */           &subfirstbyte,                /* For possible first char */
4601           &subreqbyte,                  /* For possible last char */           &subreqbyte,                  /* For possible last char */
# Line 4386  for (;; ptr++) Line 4670  for (;; ptr++)
4670    
4671      if (lengthptr != NULL)      if (lengthptr != NULL)
4672        {        {
4673          if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE)
4674            {
4675            *errorcodeptr = ERR20;
4676            goto FAILED;
4677            }
4678        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;
4679        code++;        code++;
4680        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
# Line 4671  Arguments: Line 4960  Arguments:
4960    ptrptr         -> the address of the current pattern pointer    ptrptr         -> the address of the current pattern pointer
4961    errorcodeptr   -> pointer to error code variable    errorcodeptr   -> pointer to error code variable
4962    lookbehind     TRUE if this is a lookbehind assertion    lookbehind     TRUE if this is a lookbehind assertion
4963    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
4964    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
4965    firstbyteptr   place to put the first required character, or a negative number    firstbyteptr   place to put the first required character, or a negative number
4966    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 4685  Returns: TRUE on success Line 4974  Returns: TRUE on success
4974    
4975  static BOOL  static BOOL
4976  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,
4977    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
4978    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,
4979    int *lengthptr)    int *lengthptr)
4980  {  {
4981  const uschar *ptr = *ptrptr;  const uschar *ptr = *ptrptr;
# Line 4731  orig_bracount = max_bracount = cd->braco Line 5020  orig_bracount = max_bracount = cd->braco
5020  for (;;)  for (;;)
5021    {    {
5022    /* For a (?| group, reset the capturing bracket count so that each branch    /* For a (?| group, reset the capturing bracket count so that each branch
5023    uses the same numbers. */    uses the same numbers. */
5024    
5025    if (reset_bracount) cd->bracount = orig_bracount;    if (reset_bracount) cd->bracount = orig_bracount;
5026    
5027    /* Handle a change of ims options at the start of the branch */    /* Handle a change of ims options at the start of the branch */
5028    
5029    if ((options & PCRE_IMS) != oldims)    if ((options & PCRE_IMS) != oldims)
# Line 4763  for (;;) Line 5052  for (;;)
5052      *ptrptr = ptr;      *ptrptr = ptr;
5053      return FALSE;      return FALSE;
5054      }      }
5055    
5056    /* Keep the highest bracket count in case (?| was used and some branch    /* Keep the highest bracket count in case (?| was used and some branch
5057    has fewer than the rest. */    has fewer than the rest. */
5058    
5059    if (cd->bracount > max_bracount) max_bracount = cd->bracount;    if (cd->bracount > max_bracount) max_bracount = cd->bracount;
5060    
5061    /* 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 4870  for (;;) Line 5159  for (;;)
5159        *code++ = oldims;        *code++ = oldims;
5160        length += 2;        length += 2;
5161        }        }
5162    
5163      /* Retain the highest bracket number, in case resetting was used. */      /* Retain the highest bracket number, in case resetting was used. */
5164    
5165      cd->bracount = max_bracount;      cd->bracount = max_bracount;
5166    
5167      /* Set values to pass back */      /* Set values to pass back */
# Line 4881  for (;;) Line 5170  for (;;)
5170      *ptrptr = ptr;      *ptrptr = ptr;
5171      *firstbyteptr = firstbyte;      *firstbyteptr = firstbyte;
5172      *reqbyteptr = reqbyte;      *reqbyteptr = reqbyte;
5173      if (lengthptr != NULL) *lengthptr += length;      if (lengthptr != NULL)
5174          {
5175          if (OFLOW_MAX - *lengthptr < length)
5176            {
5177            *errorcodeptr = ERR20;
5178            return FALSE;
5179            }
5180          *lengthptr += length;
5181          }
5182      return TRUE;      return TRUE;
5183      }      }
5184    
# Line 5349  outside can help speed up starting point Line 5646  outside can help speed up starting point
5646  code = cworkspace;  code = cworkspace;
5647  *code = OP_BRA;  *code = OP_BRA;
5648  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,
5649    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,
5650    &length);    &length);
5651  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
5652    
Legend:
Removed from v.175  
changed lines
  Added in v.206
webmaster@exim.org
 ViewVC Help
Powered by ViewVC 1.1.12