/[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 514 by ph10, Mon May 3 12:54:22 2010 UTC revision 533 by ph10, Wed Jun 2 19:02:41 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
265    substitutes must be in the order of the names, defined above, and there are
266    both positive and negative cases. NULL means no substitute. */
267    
268    #ifdef SUPPORT_UCP
269    static const uschar *substitutes[] = {
270      (uschar *)"\\P{Nd}",    /* \D */
271      (uschar *)"\\p{Nd}",    /* \d */
272      (uschar *)"\\P{Xsp}",   /* \S */       /* NOTE: Xsp is Perl space */
273      (uschar *)"\\p{Xsp}",   /* \s */
274      (uschar *)"\\P{Xwd}",   /* \W */
275      (uschar *)"\\p{Xwd}"    /* \w */
276    };
277    
278    static const uschar *posix_substitutes[] = {
279      (uschar *)"\\p{L}",     /* alpha */
280      (uschar *)"\\p{Ll}",    /* lower */
281      (uschar *)"\\p{Lu}",    /* upper */
282      (uschar *)"\\p{Xan}",   /* alnum */
283      NULL,                   /* ascii */
284      (uschar *)"\\h",        /* blank */
285      NULL,                   /* cntrl */
286      (uschar *)"\\p{Nd}",    /* digit */
287      NULL,                   /* graph */
288      NULL,                   /* print */
289      NULL,                   /* punct */
290      (uschar *)"\\p{Xps}",   /* space */    /* NOTE: Xps is POSIX space */
291      (uschar *)"\\p{Xwd}",   /* word */
292      NULL,                   /* xdigit */
293      /* Negated cases */
294      (uschar *)"\\P{L}",     /* ^alpha */
295      (uschar *)"\\P{Ll}",    /* ^lower */
296      (uschar *)"\\P{Lu}",    /* ^upper */
297      (uschar *)"\\P{Xan}",   /* ^alnum */
298      NULL,                   /* ^ascii */
299      (uschar *)"\\H",        /* ^blank */
300      NULL,                   /* ^cntrl */
301      (uschar *)"\\P{Nd}",    /* ^digit */
302      NULL,                   /* ^graph */
303      NULL,                   /* ^print */
304      NULL,                   /* ^punct */
305      (uschar *)"\\P{Xps}",   /* ^space */   /* NOTE: Xps is POSIX space */
306      (uschar *)"\\P{Xwd}",   /* ^word */
307      NULL                    /* ^xdigit */
308    };
309    #define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *))
310    #endif
311    
312  #define STRING(a)  # a  #define STRING(a)  # a
313  #define XSTRING(s) STRING(s)  #define XSTRING(s) STRING(s)
# Line 360  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"
411    ;    ;
412    
413  /* 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 829  else Line 877  else
877      break;      break;
878      }      }
879    }    }
880    
881  /* Perl supports \N{name} for character names, as well as plain \N for "not  /* Perl supports \N{name} for character names, as well as plain \N for "not
882  newline". PCRE does not support \N{name}. */  newline". PCRE does not support \N{name}. */
883    
884  if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET)  if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET)
885    *errorcodeptr = ERR37;    *errorcodeptr = ERR37;
886    
887    /* If PCRE_UCP is set, we change the values for \d etc. */
888    
889    if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w)
890      c -= (ESC_DU - ESC_D);
891    
892    /* Set the pointer to the final character before returning. */
893    
894  *ptrptr = ptr;  *ptrptr = ptr;
895  return c;  return c;
# Line 2261  auto_callout(uschar *code, const uschar Line 2316  auto_callout(uschar *code, const uschar
2316  {  {
2317  *code++ = OP_CALLOUT;  *code++ = OP_CALLOUT;
2318  *code++ = 255;  *code++ = 255;
2319  PUT(code, 0, ptr - cd->start_pattern);  /* Pattern offset */  PUT(code, 0, (int)(ptr - cd->start_pattern));  /* Pattern offset */
2320  PUT(code, LINK_SIZE, 0);                /* Default length */  PUT(code, LINK_SIZE, 0);                       /* Default length */
2321  return code + 2*LINK_SIZE;  return code + 2*LINK_SIZE;
2322  }  }
2323    
# Line 2287  Returns: nothing Line 2342  Returns: nothing
2342  static void  static void
2343  complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd)  complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd)
2344  {  {
2345  int length = ptr - cd->start_pattern - GET(previous_callout, 2);  int length = (int)(ptr - cd->start_pattern - GET(previous_callout, 2));
2346  PUT(previous_callout, 2 + LINK_SIZE, length);  PUT(previous_callout, 2 + LINK_SIZE, length);
2347  }  }
2348    
# Line 2337  for (++c; c <= d; c++) Line 2392  for (++c; c <= d; c++)
2392    
2393  return TRUE;  return TRUE;
2394  }  }
2395    
2396    
2397    
2398    /*************************************************
2399    *        Check a character and a property        *
2400    *************************************************/
2401    
2402    /* This function is called by check_auto_possessive() when a property item
2403    is adjacent to a fixed character.
2404    
2405    Arguments:
2406      c            the character
2407      ptype        the property type
2408      pdata        the data for the type
2409      negated      TRUE if it's a negated property (\P or \p{^)
2410    
2411    Returns:       TRUE if auto-possessifying is OK
2412    */
2413    
2414    static BOOL
2415    check_char_prop(int c, int ptype, int pdata, BOOL negated)
2416    {
2417    const ucd_record *prop = GET_UCD(c);
2418    switch(ptype)
2419      {
2420      case PT_LAMP:
2421      return (prop->chartype == ucp_Lu ||
2422              prop->chartype == ucp_Ll ||
2423              prop->chartype == ucp_Lt) == negated;
2424    
2425      case PT_GC:
2426      return (pdata == _pcre_ucp_gentype[prop->chartype]) == negated;
2427    
2428      case PT_PC:
2429      return (pdata == prop->chartype) == negated;
2430    
2431      case PT_SC:
2432      return (pdata == prop->script) == negated;
2433    
2434      /* These are specials */
2435    
2436      case PT_ALNUM:
2437      return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||
2438              _pcre_ucp_gentype[prop->chartype] == ucp_N) == negated;
2439    
2440      case PT_SPACE:    /* Perl space */
2441      return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
2442              c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2443              == negated;
2444    
2445      case PT_PXSPACE:  /* POSIX space */
2446      return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
2447              c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2448              c == CHAR_FF || c == CHAR_CR)
2449              == negated;
2450    
2451      case PT_WORD:
2452      return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||
2453              _pcre_ucp_gentype[prop->chartype] == ucp_N ||
2454              c == CHAR_UNDERSCORE) == negated;
2455      }
2456    return FALSE;
2457    }
2458  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2459    
2460    
# Line 2350  whether the next thing could possibly ma Line 2468  whether the next thing could possibly ma
2468  sense to automatically possessify the repeated item.  sense to automatically possessify the repeated item.
2469    
2470  Arguments:  Arguments:
2471    op_code       the repeated op code    previous      pointer to the repeated opcode
   this          data for this item, depends on the opcode  
2472    utf8          TRUE in UTF-8 mode    utf8          TRUE in UTF-8 mode
   utf8_char     used for utf8 character bytes, NULL if not relevant  
2473    ptr           next character in pattern    ptr           next character in pattern
2474    options       options bits    options       options bits
2475    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
# Line 2362  Returns: TRUE if possessifying is Line 2478  Returns: TRUE if possessifying is
2478  */  */
2479    
2480  static BOOL  static BOOL
2481  check_auto_possessive(int op_code, int item, BOOL utf8, uschar *utf8_char,  check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr,
2482    const uschar *ptr, int options, compile_data *cd)    int options, compile_data *cd)
2483  {  {
2484  int next;  int c, next;
2485    int op_code = *previous++;
2486    
2487  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
2488    
# Line 2426  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 2543  if (*ptr == CHAR_ASTERISK || *ptr == CHA
2543    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)
2544      return FALSE;      return FALSE;
2545    
2546  /* Now compare the next item with the previous opcode. If the previous is a  /* Now compare the next item with the previous opcode. First, handle cases when
2547  positive single character match, "item" either contains the character or, if  the next item is a character. */
 "item" is greater than 127 in utf8 mode, the character's bytes are in  
 utf8_char. */  
   
   
 /* Handle cases when the next item is a character. */  
2548    
2549  if (next >= 0) switch(op_code)  if (next >= 0) switch(op_code)
2550    {    {
2551    case OP_CHAR:    case OP_CHAR:
2552  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2553    if (utf8 && item > 127) { GETCHAR(item, utf8_char); }    GETCHARTEST(c, previous);
2554  #else  #else
2555    (void)(utf8_char);  /* Keep compiler happy by referencing function argument */    c = *previous;
2556  #endif  #endif
2557    return item != next;    return c != next;
2558    
2559    /* 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
2560    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
2561    high-valued characters. */    high-valued characters. */
2562    
2563    case OP_CHARNC:    case OP_CHARNC:
2564  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2565    if (utf8 && item > 127) { GETCHAR(item, utf8_char); }    GETCHARTEST(c, previous);
2566  #endif  #else
2567    if (item == next) return FALSE;    c = *previous;
2568    #endif
2569      if (c == next) return FALSE;
2570  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2571    if (utf8)    if (utf8)
2572      {      {
# Line 2463  if (next >= 0) switch(op_code) Line 2577  if (next >= 0) switch(op_code)
2577  #else  #else
2578      othercase = NOTACHAR;      othercase = NOTACHAR;
2579  #endif  #endif
2580      return (unsigned int)item != othercase;      return (unsigned int)c != othercase;
2581      }      }
2582    else    else
2583  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
2584    return (item != cd->fcc[next]);  /* Non-UTF-8 mode */    return (c != cd->fcc[next]);  /* Non-UTF-8 mode */
2585    
2586    /* For OP_NOT, "item" must be a single-byte character. */    /* For OP_NOT, its data is always a single-byte character. */
2587    
2588    case OP_NOT:    case OP_NOT:
2589    if (item == next) return TRUE;    if ((c = *previous) == next) return TRUE;
2590    if ((options & PCRE_CASELESS) == 0) return FALSE;    if ((options & PCRE_CASELESS) == 0) return FALSE;
2591  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2592    if (utf8)    if (utf8)
# Line 2484  if (next >= 0) switch(op_code) Line 2598  if (next >= 0) switch(op_code)
2598  #else  #else
2599      othercase = NOTACHAR;      othercase = NOTACHAR;
2600  #endif  #endif
2601      return (unsigned int)item == othercase;      return (unsigned int)c == othercase;
2602      }      }
2603    else    else
2604  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
2605    return (item == cd->fcc[next]);  /* Non-UTF-8 mode */    return (c == cd->fcc[next]);  /* Non-UTF-8 mode */
2606    
2607      /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
2608      When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
2609    
2610    case OP_DIGIT:    case OP_DIGIT:
2611    return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;    return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;
# Line 2531  if (next >= 0) switch(op_code) Line 2648  if (next >= 0) switch(op_code)
2648      case 0x202f:      case 0x202f:
2649      case 0x205f:      case 0x205f:
2650      case 0x3000:      case 0x3000:
2651      return op_code != OP_HSPACE;      return op_code == OP_NOT_HSPACE;
2652      default:      default:
2653      return op_code == OP_HSPACE;      return op_code != OP_NOT_HSPACE;
2654      }      }
2655    
2656      case OP_ANYNL:
2657    case OP_VSPACE:    case OP_VSPACE:
2658    case OP_NOT_VSPACE:    case OP_NOT_VSPACE:
2659    switch(next)    switch(next)
# Line 2547  if (next >= 0) switch(op_code) Line 2665  if (next >= 0) switch(op_code)
2665      case 0x85:      case 0x85:
2666      case 0x2028:      case 0x2028:
2667      case 0x2029:      case 0x2029:
2668      return op_code != OP_VSPACE;      return op_code == OP_NOT_VSPACE;
2669      default:      default:
2670      return op_code == OP_VSPACE;      return op_code != OP_NOT_VSPACE;
2671      }      }
2672    
2673    #ifdef SUPPORT_UCP
2674      case OP_PROP:
2675      return check_char_prop(next, previous[0], previous[1], FALSE);
2676    
2677      case OP_NOTPROP:
2678      return check_char_prop(next, previous[0], previous[1], TRUE);
2679    #endif
2680    
2681    default:    default:
2682    return FALSE;    return FALSE;
2683    }    }
2684    
2685    
2686  /* Handle the case when the next item is \d, \s, etc. */  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
2687    is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
2688    generated only when PCRE_UCP is *not* set, that is, when only ASCII
2689    characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are
2690    replaced by OP_PROP codes when PCRE_UCP is set. */
2691    
2692  switch(op_code)  switch(op_code)
2693    {    {
2694    case OP_CHAR:    case OP_CHAR:
2695    case OP_CHARNC:    case OP_CHARNC:
2696  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2697    if (utf8 && item > 127) { GETCHAR(item, utf8_char); }    GETCHARTEST(c, previous);
2698  #endif  #else
2699      c = *previous;
2700    #endif
2701    switch(-next)    switch(-next)
2702      {      {
2703      case ESC_d:      case ESC_d:
2704      return item > 127 || (cd->ctypes[item] & ctype_digit) == 0;      return c > 127 || (cd->ctypes[c] & ctype_digit) == 0;
2705    
2706      case ESC_D:      case ESC_D:
2707      return item <= 127 && (cd->ctypes[item] & ctype_digit) != 0;      return c <= 127 && (cd->ctypes[c] & ctype_digit) != 0;
2708    
2709      case ESC_s:      case ESC_s:
2710      return item > 127 || (cd->ctypes[item] & ctype_space) == 0;      return c > 127 || (cd->ctypes[c] & ctype_space) == 0;
2711    
2712      case ESC_S:      case ESC_S:
2713      return item <= 127 && (cd->ctypes[item] & ctype_space) != 0;      return c <= 127 && (cd->ctypes[c] & ctype_space) != 0;
2714    
2715      case ESC_w:      case ESC_w:
2716      return item > 127 || (cd->ctypes[item] & ctype_word) == 0;      return c > 127 || (cd->ctypes[c] & ctype_word) == 0;
2717    
2718      case ESC_W:      case ESC_W:
2719      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;      return c <= 127 && (cd->ctypes[c] & ctype_word) != 0;
2720    
2721      case ESC_h:      case ESC_h:
2722      case ESC_H:      case ESC_H:
2723      switch(item)      switch(c)
2724        {        {
2725        case 0x09:        case 0x09:
2726        case 0x20:        case 0x20:
# Line 2616  switch(op_code) Line 2748  switch(op_code)
2748    
2749      case ESC_v:      case ESC_v:
2750      case ESC_V:      case ESC_V:
2751      switch(item)      switch(c)
2752        {        {
2753        case 0x0a:        case 0x0a:
2754        case 0x0b:        case 0x0b:
# Line 2629  switch(op_code) Line 2761  switch(op_code)
2761        default:        default:
2762        return -next == ESC_v;        return -next == ESC_v;
2763        }        }
2764    
2765        /* When PCRE_UCP is set, these values get generated for \d etc. Find
2766        their substitutions and process them. The result will always be either
2767        -ESC_p or -ESC_P. Then fall through to process those values. */
2768    
2769    #ifdef SUPPORT_UCP
2770        case ESC_du:
2771        case ESC_DU:
2772        case ESC_wu:
2773        case ESC_WU:
2774        case ESC_su:
2775        case ESC_SU:
2776          {
2777          int temperrorcode = 0;
2778          ptr = substitutes[-next - ESC_DU];
2779          next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);
2780          if (temperrorcode != 0) return FALSE;
2781          ptr++;    /* For compatibility */
2782          }
2783        /* Fall through */
2784    
2785        case ESC_p:
2786        case ESC_P:
2787          {
2788          int ptype, pdata, errorcodeptr;
2789          BOOL negated;
2790    
2791          ptr--;      /* Make ptr point at the p or P */
2792          ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);
2793          if (ptype < 0) return FALSE;
2794          ptr++;      /* Point past the final curly ket */
2795    
2796          /* If the property item is optional, we have to give up. (When generated
2797          from \d etc by PCRE_UCP, this test will have been applied much earlier,
2798          to the original \d etc. At this point, ptr will point to a zero byte. */
2799    
2800          if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
2801            strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
2802              return FALSE;
2803    
2804          /* Do the property check. */
2805    
2806          return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);
2807          }
2808    #endif
2809    
2810      default:      default:
2811      return FALSE;      return FALSE;
2812      }      }
2813    
2814      /* In principle, support for Unicode properties should be integrated here as
2815      well. It means re-organizing the above code so as to get hold of the property
2816      values before switching on the op-code. However, I wonder how many patterns
2817      combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set,
2818      these op-codes are never generated.) */
2819    
2820    case OP_DIGIT:    case OP_DIGIT:
2821    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
2822           next == -ESC_h || next == -ESC_v;           next == -ESC_h || next == -ESC_v || next == -ESC_R;
2823    
2824    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
2825    return next == -ESC_d;    return next == -ESC_d;
2826    
2827    case OP_WHITESPACE:    case OP_WHITESPACE:
2828    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return next == -ESC_S || next == -ESC_d || next == -ESC_w || next == -ESC_R;
2829    
2830    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
2831    return next == -ESC_s || next == -ESC_h || next == -ESC_v;    return next == -ESC_s || next == -ESC_h || next == -ESC_v;
2832    
2833    case OP_HSPACE:    case OP_HSPACE:
2834    return next == -ESC_S || next == -ESC_H || next == -ESC_d || next == -ESC_w;    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||
2835             next == -ESC_w || next == -ESC_v || next == -ESC_R;
2836    
2837    case OP_NOT_HSPACE:    case OP_NOT_HSPACE:
2838    return next == -ESC_h;    return next == -ESC_h;
2839    
2840    /* Can't have \S in here because VT matches \S (Perl anomaly) */    /* Can't have \S in here because VT matches \S (Perl anomaly) */
2841      case OP_ANYNL:
2842    case OP_VSPACE:    case OP_VSPACE:
2843    return next == -ESC_V || next == -ESC_d || next == -ESC_w;    return next == -ESC_V || next == -ESC_d || next == -ESC_w;
2844    
2845    case OP_NOT_VSPACE:    case OP_NOT_VSPACE:
2846    return next == -ESC_v;    return next == -ESC_v || next == -ESC_R;
2847    
2848    case OP_WORDCHAR:    case OP_WORDCHAR:
2849    return next == -ESC_W || next == -ESC_s || next == -ESC_h || next == -ESC_v;    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||
2850             next == -ESC_v || next == -ESC_R;
2851    
2852    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2853    return next == -ESC_w || next == -ESC_d;    return next == -ESC_w || next == -ESC_d;
# Line 2725  BOOL inescq = FALSE; Line 2911  BOOL inescq = FALSE;
2911  BOOL groupsetfirstbyte = FALSE;  BOOL groupsetfirstbyte = FALSE;
2912  const uschar *ptr = *ptrptr;  const uschar *ptr = *ptrptr;
2913  const uschar *tempptr;  const uschar *tempptr;
2914    const uschar *nestptr = NULL;
2915  uschar *previous = NULL;  uschar *previous = NULL;
2916  uschar *previous_callout = NULL;  uschar *previous_callout = NULL;
2917  uschar *save_hwm = NULL;  uschar *save_hwm = NULL;
# Line 2795  for (;; ptr++) Line 2982  for (;; ptr++)
2982    
2983    c = *ptr;    c = *ptr;
2984    
2985      /* If we are at the end of a nested substitution, revert to the outer level
2986      string. Nesting only happens one level deep. */
2987    
2988      if (c == 0 && nestptr != NULL)
2989        {
2990        ptr = nestptr;
2991        nestptr = NULL;
2992        c = *ptr;
2993        }
2994    
2995    /* If we are in the pre-compile phase, accumulate the length used for the    /* If we are in the pre-compile phase, accumulate the length used for the
2996    previous cycle of this loop. */    previous cycle of this loop. */
2997    
# Line 2825  for (;; ptr++) Line 3022  for (;; ptr++)
3022        goto FAILED;        goto FAILED;
3023        }        }
3024    
3025      *lengthptr += code - last_code;      *lengthptr += (int)(code - last_code);
3026      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));
3027    
3028      /* If "previous" is set and it is not at the start of the work space, move      /* If "previous" is set and it is not at the start of the work space, move
# Line 2943  for (;; ptr++) Line 3140  for (;; ptr++)
3140          *errorcodeptr = ERR20;          *errorcodeptr = ERR20;
3141          goto FAILED;          goto FAILED;
3142          }          }
3143        *lengthptr += code - last_code;   /* To include callout length */        *lengthptr += (int)(code - last_code);   /* To include callout length */
3144        DPRINTF((">> end branch\n"));        DPRINTF((">> end branch\n"));
3145        }        }
3146      return TRUE;      return TRUE;
# Line 3092  for (;; ptr++) Line 3289  for (;; ptr++)
3289          {                           /* Braces are required because the */          {                           /* Braces are required because the */
3290          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
3291          }          }
3292    
3293        /* 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
3294        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
3295        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 3148  for (;; ptr++) Line 3345  for (;; ptr++)
3345            ptr++;            ptr++;
3346            }            }
3347    
3348          posix_class = check_posix_name(ptr, tempptr - ptr);          posix_class = check_posix_name(ptr, (int)(tempptr - ptr));
3349          if (posix_class < 0)          if (posix_class < 0)
3350            {            {
3351            *errorcodeptr = ERR30;            *errorcodeptr = ERR30;
# Line 3161  for (;; ptr++) Line 3358  for (;; ptr++)
3358    
3359          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
3360            posix_class = 0;            posix_class = 0;
3361    
3362          /* We build the bit map for the POSIX class in a chunk of local store          /* When PCRE_UCP is set, some of the POSIX classes are converted to
3363          because we may be adding and subtracting from it, and we don't want to          different escape sequences that use Unicode properties. */
3364          subtract bits that may be in the main map already. At the end we or the  
3365          result into the bit map that is being built. */  #ifdef SUPPORT_UCP
3366            if ((options & PCRE_UCP) != 0)
3367              {
3368              int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);
3369              if (posix_substitutes[pc] != NULL)
3370                {
3371                nestptr = tempptr + 1;
3372                ptr = posix_substitutes[pc] - 1;
3373                continue;
3374                }
3375              }
3376    #endif
3377            /* In the non-UCP case, we build the bit map for the POSIX class in a
3378            chunk of local store because we may be adding and subtracting from it,
3379            and we don't want to subtract bits that may be in the main map already.
3380            At the end we or the result into the bit map that is being built. */
3381    
3382          posix_class *= 3;          posix_class *= 3;
3383    
# Line 3237  for (;; ptr++) Line 3449  for (;; ptr++)
3449            register const uschar *cbits = cd->cbits;            register const uschar *cbits = cd->cbits;
3450            class_charcount += 2;     /* Greater than 1 is what matters */            class_charcount += 2;     /* Greater than 1 is what matters */
3451    
3452            /* Save time by not doing this in the pre-compile phase. */            switch (-c)
   
           if (lengthptr == NULL) switch (-c)  
3453              {              {
3454    #ifdef SUPPORT_UCP
3455                case ESC_du:     /* These are the values given for \d etc */
3456                case ESC_DU:     /* when PCRE_UCP is set. We replace the */
3457                case ESC_wu:     /* escape sequence with an appropriate \p */
3458                case ESC_WU:     /* or \P to test Unicode properties instead */
3459                case ESC_su:     /* of the default ASCII testing. */
3460                case ESC_SU:
3461                nestptr = ptr;
3462                ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
3463                class_charcount -= 2;                /* Undo! */
3464                continue;
3465    #endif
3466              case ESC_d:              case ESC_d:
3467              for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit];              for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit];
3468              continue;              continue;
# Line 3270  for (;; ptr++) Line 3492  for (;; ptr++)
3492              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */
3493              continue;              continue;
3494    
3495              default:    /* Not recognized; fall through */              case ESC_h:
             break;      /* Need "default" setting to stop compiler warning. */  
             }  
   
           /* In the pre-compile phase, just do the recognition. */  
   
           else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||  
                    c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;  
   
           /* We need to deal with \H, \h, \V, and \v in both phases because  
           they use extra memory. */  
   
           if (-c == ESC_h)  
             {  
3496              SETBIT(classbits, 0x09); /* VT */              SETBIT(classbits, 0x09); /* VT */
3497              SETBIT(classbits, 0x20); /* SPACE */              SETBIT(classbits, 0x20); /* SPACE */
3498              SETBIT(classbits, 0xa0); /* NSBP */              SETBIT(classbits, 0xa0); /* NSBP */
# Line 3307  for (;; ptr++) Line 3516  for (;; ptr++)
3516                }                }
3517  #endif  #endif
3518              continue;              continue;
             }  
3519    
3520            if (-c == ESC_H)              case ESC_H:
             {  
3521              for (c = 0; c < 32; c++)              for (c = 0; c < 32; c++)
3522                {                {
3523                int x = 0xff;                int x = 0xff;
# Line 3352  for (;; ptr++) Line 3559  for (;; ptr++)
3559                }                }
3560  #endif  #endif
3561              continue;              continue;
             }  
3562    
3563            if (-c == ESC_v)              case ESC_v:
             {  
3564              SETBIT(classbits, 0x0a); /* LF */              SETBIT(classbits, 0x0a); /* LF */
3565              SETBIT(classbits, 0x0b); /* VT */              SETBIT(classbits, 0x0b); /* VT */
3566              SETBIT(classbits, 0x0c); /* FF */              SETBIT(classbits, 0x0c); /* FF */
# Line 3371  for (;; ptr++) Line 3576  for (;; ptr++)
3576                }                }
3577  #endif  #endif
3578              continue;              continue;
             }  
3579    
3580            if (-c == ESC_V)              case ESC_V:
             {  
3581              for (c = 0; c < 32; c++)              for (c = 0; c < 32; c++)
3582                {                {
3583                int x = 0xff;                int x = 0xff;
# Line 3404  for (;; ptr++) Line 3607  for (;; ptr++)
3607                }                }
3608  #endif  #endif
3609              continue;              continue;
             }  
   
           /* We need to deal with \P and \p in both phases. */  
3610    
3611  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3612            if (-c == ESC_p || -c == ESC_P)              case ESC_p:
3613              {              case ESC_P:
3614              BOOL negated;                {
3615              int pdata;                BOOL negated;
3616              int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                int pdata;
3617              if (ptype < 0) goto FAILED;                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
3618              class_utf8 = TRUE;                if (ptype < 0) goto FAILED;
3619              *class_utf8data++ = ((-c == ESC_p) != negated)?                class_utf8 = TRUE;
3620                XCL_PROP : XCL_NOTPROP;                *class_utf8data++ = ((-c == ESC_p) != negated)?
3621              *class_utf8data++ = ptype;                  XCL_PROP : XCL_NOTPROP;
3622              *class_utf8data++ = pdata;                *class_utf8data++ = ptype;
3623              class_charcount -= 2;   /* Not a < 256 character */                *class_utf8data++ = pdata;
3624              continue;                class_charcount -= 2;   /* Not a < 256 character */
3625              }                continue;
3626                  }
3627  #endif  #endif
3628            /* Unrecognized escapes are faulted if PCRE is running in its              /* Unrecognized escapes are faulted if PCRE is running in its
3629            strict mode. By default, for compatibility with Perl, they are              strict mode. By default, for compatibility with Perl, they are
3630            treated as literals. */              treated as literals. */
3631    
3632            if ((options & PCRE_EXTRA) != 0)              default:
3633              {              if ((options & PCRE_EXTRA) != 0)
3634              *errorcodeptr = ERR7;                {
3635              goto FAILED;                *errorcodeptr = ERR7;
3636                  goto FAILED;
3637                  }
3638                class_charcount -= 2;  /* Undo the default count from above */
3639                c = *ptr;              /* Get the final character and fall through */
3640                break;
3641              }              }
   
           class_charcount -= 2;  /* Undo the default count from above */  
           c = *ptr;              /* Get the final character and fall through */  
3642            }            }
3643    
3644          /* Fall through if we have a single character (c >= 0). This may be          /* Fall through if we have a single character (c >= 0). This may be
# Line 3675  for (;; ptr++) Line 3878  for (;; ptr++)
3878          }          }
3879        }        }
3880    
3881      /* Loop until ']' reached. This "while" is the end of the "do" above. */      /* Loop until ']' reached. This "while" is the end of the "do" far above.
3882        If we are at the end of an internal nested string, revert to the outer
3883        string. */
3884    
3885        while (((c = *(++ptr)) != 0 ||
3886               (nestptr != NULL &&
3887                 (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != 0)) &&
3888               (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));
3889    
3890      while ((c = *(++ptr)) != 0 && (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));      /* Check for missing terminating ']' */
3891    
3892      if (c == 0)                          /* Missing terminating ']' */      if (c == 0)
3893        {        {
3894        *errorcodeptr = ERR6;        *errorcodeptr = ERR6;
3895        goto FAILED;        goto FAILED;
3896        }        }
3897    
   
 /* This code has been disabled because it would mean that \s counts as  
 an explicit \r or \n reference, and that's not really what is wanted. Now  
 we set the flag only if there is a literal "\r" or "\n" in the class. */  
   
 #if 0  
     /* Remember whether \r or \n are in this class */  
   
     if (negate_class)  
       {  
       if ((classbits[1] & 0x24) != 0x24) cd->external_flags |= PCRE_HASCRORLF;  
       }  
     else  
       {  
       if ((classbits[1] & 0x24) != 0) cd->external_flags |= PCRE_HASCRORLF;  
       }  
 #endif  
   
   
3898      /* 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
3899      less than 256. As long as there were no characters >= 128 and there was no      less than 256. As long as there were no characters >= 128 and there was no
3900      use of \p or \P, in other words, no use of any XCLASS features, we can      use of \p or \P, in other words, no use of any XCLASS features, we can
# Line 3720  we set the flag only if there is a liter Line 3911  we set the flag only if there is a liter
3911      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
3912      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
3913      reqbyte, save the previous value for reinstating. */      reqbyte, save the previous value for reinstating. */
3914    
3915  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3916      if (class_charcount == 1 && !class_utf8 &&      if (class_charcount == 1 && !class_utf8 &&
3917        (!utf8 || !negate_class || class_lastchar < 128))        (!utf8 || !negate_class || class_lastchar < 128))
# Line 3767  we set the flag only if there is a liter Line 3958  we set the flag only if there is a liter
3958    
3959      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
3960      extended class, with its own opcode, unless there was a negated special      extended class, with its own opcode, unless there was a negated special
3961      such as \S in the class, because in that case all characters > 255 are in      such as \S in the class, and PCRE_UCP is not set, because in that case all
3962      the class, so any that were explicitly given as well can be ignored. If      characters > 255 are in the class, so any that were explicitly given as
3963      (when there are explicit characters > 255 that must be listed) there are no      well can be ignored. If (when there are explicit characters > 255 that must
3964      characters < 256, we can omit the bitmap in the actual compiled code. */      be listed) there are no characters < 256, we can omit the bitmap in the
3965        actual compiled code. */
3966    
3967  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3968      if (class_utf8 && !should_flip_negation)      if (class_utf8 && (!should_flip_negation || (options & PCRE_UCP) != 0))
3969        {        {
3970        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */
3971        *code++ = OP_XCLASS;        *code++ = OP_XCLASS;
# Line 3799  we set the flag only if there is a liter Line 3991  we set the flag only if there is a liter
3991        }        }
3992  #endif  #endif
3993    
3994      /* If there are no characters > 255, set the opcode to OP_CLASS or      /* If there are no characters > 255, or they are all to be included or
3995      OP_NCLASS, depending on whether the whole class was negated and whether      excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the
3996      there were negative specials such as \S in the class. Then copy the 32-byte      whole class was negated and whether there were negative specials such as \S
3997      map into the code vector, negating it if necessary. */      (non-UCP) in the class. Then copy the 32-byte map into the code vector,
3998        negating it if necessary. */
3999    
4000      *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;      *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;
4001      if (negate_class)      if (negate_class)
# Line 3926  we set the flag only if there is a liter Line 4119  we set the flag only if there is a liter
4119    
4120        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4121            repeat_max < 0 &&            repeat_max < 0 &&
4122            check_auto_possessive(*previous, c, utf8, utf8_char, ptr + 1,            check_auto_possessive(previous, utf8, ptr + 1, options, cd))
             options, cd))  
4123          {          {
4124          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4125          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 3948  we set the flag only if there is a liter Line 4140  we set the flag only if there is a liter
4140        c = previous[1];        c = previous[1];
4141        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4142            repeat_max < 0 &&            repeat_max < 0 &&
4143            check_auto_possessive(OP_NOT, c, utf8, NULL, ptr + 1, options, cd))            check_auto_possessive(previous, utf8, ptr + 1, options, cd))
4144          {          {
4145          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4146          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 3972  we set the flag only if there is a liter Line 4164  we set the flag only if there is a liter
4164    
4165        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4166            repeat_max < 0 &&            repeat_max < 0 &&
4167            check_auto_possessive(c, 0, utf8, NULL, ptr + 1, options, cd))            check_auto_possessive(previous, utf8, ptr + 1, options, cd))
4168          {          {
4169          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4170          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4182  we set the flag only if there is a liter Line 4374  we set the flag only if there is a liter
4374        {        {
4375        register int i;        register int i;
4376        int ketoffset = 0;        int ketoffset = 0;
4377        int len = code - previous;        int len = (int)(code - previous);
4378        uschar *bralink = NULL;        uschar *bralink = NULL;
4379    
4380        /* Repeating a DEFINE group is pointless */        /* Repeating a DEFINE group is pointless */
# Line 4203  we set the flag only if there is a liter Line 4395  we set the flag only if there is a liter
4395          {          {
4396          register uschar *ket = previous;          register uschar *ket = previous;
4397          do ket += GET(ket, 1); while (*ket != OP_KET);          do ket += GET(ket, 1); while (*ket != OP_KET);
4398          ketoffset = code - ket;          ketoffset = (int)(code - ket);
4399          }          }
4400    
4401        /* The case of a zero minimum is special because of the need to stick        /* The case of a zero minimum is special because of the need to stick
# Line 4271  we set the flag only if there is a liter Line 4463  we set the flag only if there is a liter
4463            /* We chain together the bracket offset fields that have to be            /* We chain together the bracket offset fields that have to be
4464            filled in later when the ends of the brackets are reached. */            filled in later when the ends of the brackets are reached. */
4465    
4466            offset = (bralink == NULL)? 0 : previous - bralink;            offset = (bralink == NULL)? 0 : (int)(previous - bralink);
4467            bralink = previous;            bralink = previous;
4468            PUTINC(previous, 0, offset);            PUTINC(previous, 0, offset);
4469            }            }
# Line 4380  we set the flag only if there is a liter Line 4572  we set the flag only if there is a liter
4572              {              {
4573              int offset;              int offset;
4574              *code++ = OP_BRA;              *code++ = OP_BRA;
4575              offset = (bralink == NULL)? 0 : code - bralink;              offset = (bralink == NULL)? 0 : (int)(code - bralink);
4576              bralink = code;              bralink = code;
4577              PUTINC(code, 0, offset);              PUTINC(code, 0, offset);
4578              }              }
# Line 4401  we set the flag only if there is a liter Line 4593  we set the flag only if there is a liter
4593          while (bralink != NULL)          while (bralink != NULL)
4594            {            {
4595            int oldlinkoffset;            int oldlinkoffset;
4596            int offset = code - bralink + 1;            int offset = (int)(code - bralink + 1);
4597            uschar *bra = code - offset;            uschar *bra = code - offset;
4598            oldlinkoffset = GET(bra, 1);            oldlinkoffset = GET(bra, 1);
4599            bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset;            bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset;
# Line 4489  we set the flag only if there is a liter Line 4681  we set the flag only if there is a liter
4681  #endif  #endif
4682          }          }
4683    
4684        len = code - tempcode;        len = (int)(code - tempcode);
4685        if (len > 0) switch (*tempcode)        if (len > 0) switch (*tempcode)
4686          {          {
4687          case OP_STAR:  *tempcode = OP_POSSTAR; break;          case OP_STAR:  *tempcode = OP_POSSTAR; break;
# Line 4558  we set the flag only if there is a liter Line 4750  we set the flag only if there is a liter
4750        const uschar *arg = NULL;        const uschar *arg = NULL;
4751        previous = NULL;        previous = NULL;
4752        while ((cd->ctypes[*++ptr] & ctype_letter) != 0) {};        while ((cd->ctypes[*++ptr] & ctype_letter) != 0) {};
4753        namelen = ptr - name;        namelen = (int)(ptr - name);
4754    
4755        if (*ptr == CHAR_COLON)        if (*ptr == CHAR_COLON)
4756          {          {
4757          arg = ++ptr;          arg = ++ptr;
4758          while ((cd->ctypes[*ptr] & (ctype_letter|ctype_digit)) != 0          while ((cd->ctypes[*ptr] & (ctype_letter|ctype_digit)) != 0
4759            || *ptr == '_') ptr++;            || *ptr == '_') ptr++;
4760          arglen = ptr - arg;          arglen = (int)(ptr - arg);
4761          }          }
4762    
4763        if (*ptr != CHAR_RIGHT_PARENTHESIS)        if (*ptr != CHAR_RIGHT_PARENTHESIS)
# Line 4746  we set the flag only if there is a liter Line 4938  we set the flag only if there is a liter
4938                recno * 10 + *ptr - CHAR_0 : -1;                recno * 10 + *ptr - CHAR_0 : -1;
4939            ptr++;            ptr++;
4940            }            }
4941          namelen = ptr - name;          namelen = (int)(ptr - name);
4942    
4943          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != terminator) ||
4944              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
# Line 4942  we set the flag only if there is a liter Line 5134  we set the flag only if there is a liter
5134              goto FAILED;              goto FAILED;
5135              }              }
5136            *code++ = n;            *code++ = n;
5137            PUT(code, 0, ptr - cd->start_pattern + 1);  /* Pattern offset */            PUT(code, 0, (int)(ptr - cd->start_pattern + 1)); /* Pattern offset */
5138            PUT(code, LINK_SIZE, 0);                    /* Default length */            PUT(code, LINK_SIZE, 0);                          /* Default length */
5139            code += 2 * LINK_SIZE;            code += 2 * LINK_SIZE;
5140            }            }
5141          previous = NULL;          previous = NULL;
# Line 4976  we set the flag only if there is a liter Line 5168  we set the flag only if there is a liter
5168            name = ++ptr;            name = ++ptr;
5169    
5170            while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;            while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
5171            namelen = ptr - name;            namelen = (int)(ptr - name);
5172    
5173            /* In the pre-compile phase, just do a syntax check. */            /* In the pre-compile phase, just do a syntax check. */
5174    
# Line 5106  we set the flag only if there is a liter Line 5298  we set the flag only if there is a liter
5298          NAMED_REF_OR_RECURSE:          NAMED_REF_OR_RECURSE:
5299          name = ++ptr;          name = ++ptr;
5300          while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;          while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
5301          namelen = ptr - name;          namelen = (int)(ptr - name);
5302    
5303          /* In the pre-compile phase, do a syntax check and set a dummy          /* In the pre-compile phase, do a syntax check and set a dummy
5304          reference number. */          reference number. */
# Line 5275  we set the flag only if there is a liter Line 5467  we set the flag only if there is a liter
5467                of the group. */                of the group. */
5468    
5469                called = cd->start_code + recno;                called = cd->start_code + recno;
5470                PUTINC(cd->hwm, 0, code + 2 + LINK_SIZE - cd->start_code);                PUTINC(cd->hwm, 0, (int)(code + 2 + LINK_SIZE - cd->start_code));
5471                }                }
5472    
5473              /* If not a forward reference, and the subpattern is still open,              /* If not a forward reference, and the subpattern is still open,
# Line 5299  we set the flag only if there is a liter Line 5491  we set the flag only if there is a liter
5491            code += 1 + LINK_SIZE;            code += 1 + LINK_SIZE;
5492    
5493            *code = OP_RECURSE;            *code = OP_RECURSE;
5494            PUT(code, 1, called - cd->start_code);            PUT(code, 1, (int)(called - cd->start_code));
5495            code += 1 + LINK_SIZE;            code += 1 + LINK_SIZE;
5496    
5497            *code = OP_KET;            *code = OP_KET;
# Line 5603  we set the flag only if there is a liter Line 5795  we set the flag only if there is a liter
5795    
5796      /* ===================================================================*/      /* ===================================================================*/
5797      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
5798      are arranged to be the negation of the corresponding OP_values. For the      are arranged to be the negation of the corresponding OP_values in the
5799      back references, the values are ESC_REF plus the reference number. Only      default case when PCRE_UCP is not set. For the back references, the values
5800      back references and those types that consume a character may be repeated.      are ESC_REF plus the reference number. Only back references and those types
5801      We can test for values between ESC_b and ESC_Z for the latter; this may      that consume a character may be repeated. We can test for values between
5802      have to change if any new ones are ever created. */      ESC_b and ESC_Z for the latter; this may have to change if any new ones are
5803        ever created. */
5804    
5805      case CHAR_BACKSLASH:      case CHAR_BACKSLASH:
5806      tempptr = ptr;      tempptr = ptr;
# Line 5767  we set the flag only if there is a liter Line 5960  we set the flag only if there is a liter
5960  #endif  #endif
5961    
5962        /* For the rest (including \X when Unicode properties are supported), we        /* For the rest (including \X when Unicode properties are supported), we
5963        can obtain the OP value by negating the escape value. */        can obtain the OP value by negating the escape value in the default
5964          situation when PCRE_UCP is not set. When it *is* set, we substitute
5965          Unicode property tests. */
5966    
5967        else        else
5968          {          {
5969          previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;  #ifdef SUPPORT_UCP
5970          *code++ = -c;          if (-c >= ESC_DU && -c <= ESC_wu)
5971              {
5972              nestptr = ptr + 1;                   /* Where to resume */
5973              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
5974              }
5975            else
5976    #endif
5977              {
5978              previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
5979              *code++ = -c;
5980              }
5981          }          }
5982        continue;        continue;
5983        }        }
# Line 6104  for (;;) Line 6309  for (;;)
6309      {      {
6310      if (lengthptr == NULL)      if (lengthptr == NULL)
6311        {        {
6312        int branch_length = code - last_branch;        int branch_length = (int)(code - last_branch);
6313        do        do
6314          {          {
6315          int prev_length = GET(last_branch, 1);          int prev_length = GET(last_branch, 1);
# Line 6118  for (;;) Line 6323  for (;;)
6323      /* Fill in the ket */      /* Fill in the ket */
6324    
6325      *code = OP_KET;      *code = OP_KET;
6326      PUT(code, 1, code - start_bracket);      PUT(code, 1, (int)(code - start_bracket));
6327      code += 1 + LINK_SIZE;      code += 1 + LINK_SIZE;
6328    
6329      /* If it was a capturing subpattern, check to see if it contained any      /* If it was a capturing subpattern, check to see if it contained any
# Line 6133  for (;;) Line 6338  for (;;)
6338            code - start_bracket);            code - start_bracket);
6339          *start_bracket = OP_ONCE;          *start_bracket = OP_ONCE;
6340          code += 1 + LINK_SIZE;          code += 1 + LINK_SIZE;
6341          PUT(start_bracket, 1, code - start_bracket);          PUT(start_bracket, 1, (int)(code - start_bracket));
6342          *code = OP_KET;          *code = OP_KET;
6343          PUT(code, 1, code - start_bracket);          PUT(code, 1, (int)(code - start_bracket));
6344          code += 1 + LINK_SIZE;          code += 1 + LINK_SIZE;
6345          length += 2 + 2*LINK_SIZE;          length += 2 + 2*LINK_SIZE;
6346          }          }
# Line 6190  for (;;) Line 6395  for (;;)
6395    else    else
6396      {      {
6397      *code = OP_ALT;      *code = OP_ALT;
6398      PUT(code, 1, code - last_branch);      PUT(code, 1, (int)(code - last_branch));
6399      bc.current_branch = last_branch = code;      bc.current_branch = last_branch = code;
6400      code += 1 + LINK_SIZE;      code += 1 + LINK_SIZE;
6401      }      }
# Line 6509  int length = 1; /* For final END opcode Line 6714  int length = 1; /* For final END opcode
6714  int firstbyte, reqbyte, newline;  int firstbyte, reqbyte, newline;
6715  int errorcode = 0;  int errorcode = 0;
6716  int skipatstart = 0;  int skipatstart = 0;
6717  BOOL utf8 = (options & PCRE_UTF8) != 0;  BOOL utf8;
6718  size_t size;  size_t size;
6719  uschar *code;  uschar *code;
6720  const uschar *codestart;  const uschar *codestart;
# Line 6579  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 6784  while (ptr[skipatstart] == CHAR_LEFT_PAR
6784    
6785    if (strncmp((char *)(ptr+skipatstart+2), STRING_UTF8_RIGHTPAR, 5) == 0)    if (strncmp((char *)(ptr+skipatstart+2), STRING_UTF8_RIGHTPAR, 5) == 0)
6786      { skipatstart += 7; options |= PCRE_UTF8; continue; }      { skipatstart += 7; options |= PCRE_UTF8; continue; }
6787      else if (strncmp((char *)(ptr+skipatstart+2), STRING_UCP_RIGHTPAR, 4) == 0)
6788        { skipatstart += 6; options |= PCRE_UCP; continue; }
6789    
6790    if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0)    if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0)
6791      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }
# Line 6602  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 6809  while (ptr[skipatstart] == CHAR_LEFT_PAR
6809      options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr;      options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr;
6810    else break;    else break;
6811    }    }
6812    
6813    utf8 = (options & PCRE_UTF8) != 0;
6814    
6815  /* 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. */
6816    
# Line 6620  if (utf8) Line 6829  if (utf8)
6829    }    }
6830  #endif  #endif
6831    
6832    /* Can't support UCP unless PCRE has been compiled to include the code. */
6833    
6834    #ifndef SUPPORT_UCP
6835    if ((options & PCRE_UCP) != 0)
6836      {
6837      errorcode = ERR67;
6838      goto PCRE_EARLY_ERROR_RETURN;
6839      }
6840    #endif
6841    
6842  /* Check validity of \R options. */  /* Check validity of \R options. */
6843    
6844  switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))  switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
# Line 6748  regex compiled on a system with 4-byte p Line 6967  regex compiled on a system with 4-byte p
6967  pointers. */  pointers. */
6968    
6969  re->magic_number = MAGIC_NUMBER;  re->magic_number = MAGIC_NUMBER;
6970  re->size = size;  re->size = (int)size;
6971  re->options = cd->external_options;  re->options = cd->external_options;
6972  re->flags = cd->external_flags;  re->flags = cd->external_flags;
6973  re->dummy1 = 0;  re->dummy1 = 0;
# Line 6819  while (errorcode == 0 && cd->hwm > cwork Line 7038  while (errorcode == 0 && cd->hwm > cwork
7038    recno = GET(codestart, offset);    recno = GET(codestart, offset);
7039    groupptr = _pcre_find_bracket(codestart, utf8, recno);    groupptr = _pcre_find_bracket(codestart, utf8, recno);
7040    if (groupptr == NULL) errorcode = ERR53;    if (groupptr == NULL) errorcode = ERR53;
7041      else PUT(((uschar *)codestart), offset, groupptr - codestart);      else PUT(((uschar *)codestart), offset, (int)(groupptr - codestart));
7042    }    }
7043    
7044  /* Give an error if there's back reference to a non-existent capturing  /* Give an error if there's back reference to a non-existent capturing
# Line 6874  if (errorcode != 0) Line 7093  if (errorcode != 0)
7093    {    {
7094    (pcre_free)(re);    (pcre_free)(re);
7095    PCRE_EARLY_ERROR_RETURN:    PCRE_EARLY_ERROR_RETURN:
7096    *erroroffset = ptr - (const uschar *)pattern;    *erroroffset = (int)(ptr - (const uschar *)pattern);
7097    PCRE_EARLY_ERROR_RETURN2:    PCRE_EARLY_ERROR_RETURN2:
7098    *errorptr = find_error_text(errorcode);    *errorptr = find_error_text(errorcode);
7099    if (errorcodeptr != NULL) *errorcodeptr = errorcode;    if (errorcodeptr != NULL) *errorcodeptr = errorcode;

Legend:
Removed from v.514  
changed lines
  Added in v.533

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12