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

Diff of /code/trunk/pcre_compile.c

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

revision 533 by ph10, Wed Jun 2 19:02:41 2010 UTC revision 535 by ph10, Thu Jun 3 19:18:24 2010 UTC
# Line 261  static const int posix_class_maps[] = { Line 261  static const int posix_class_maps[] = {
261    cbit_xdigit,-1,          0              /* xdigit */    cbit_xdigit,-1,          0              /* xdigit */
262  };  };
263    
264  /* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class  /* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class
265  substitutes must be in the order of the names, defined above, and there are  substitutes must be in the order of the names, defined above, and there are
266  both positive and negative cases. NULL means no substitute. */  both positive and negative cases. NULL means no substitute. */
267    
268  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 272  static const uschar *substitutes[] = { Line 272  static const uschar *substitutes[] = {
272    (uschar *)"\\P{Xsp}",   /* \S */       /* NOTE: Xsp is Perl space */    (uschar *)"\\P{Xsp}",   /* \S */       /* NOTE: Xsp is Perl space */
273    (uschar *)"\\p{Xsp}",   /* \s */    (uschar *)"\\p{Xsp}",   /* \s */
274    (uschar *)"\\P{Xwd}",   /* \W */    (uschar *)"\\P{Xwd}",   /* \W */
275    (uschar *)"\\p{Xwd}"    /* \w */    (uschar *)"\\p{Xwd}"    /* \w */
276  };  };
277    
278  static const uschar *posix_substitutes[] = {  static const uschar *posix_substitutes[] = {
279    (uschar *)"\\p{L}",     /* alpha */    (uschar *)"\\p{L}",     /* alpha */
280    (uschar *)"\\p{Ll}",    /* lower */    (uschar *)"\\p{Ll}",    /* lower */
281    (uschar *)"\\p{Lu}",    /* upper */    (uschar *)"\\p{Lu}",    /* upper */
282    (uschar *)"\\p{Xan}",   /* alnum */    (uschar *)"\\p{Xan}",   /* alnum */
283    NULL,                   /* ascii */    NULL,                   /* ascii */
284    (uschar *)"\\h",        /* blank */    (uschar *)"\\h",        /* blank */
285    NULL,                   /* cntrl */    NULL,                   /* cntrl */
# Line 289  static const uschar *posix_substitutes[] Line 289  static const uschar *posix_substitutes[]
289    NULL,                   /* punct */    NULL,                   /* punct */
290    (uschar *)"\\p{Xps}",   /* space */    /* NOTE: Xps is POSIX space */    (uschar *)"\\p{Xps}",   /* space */    /* NOTE: Xps is POSIX space */
291    (uschar *)"\\p{Xwd}",   /* word */    (uschar *)"\\p{Xwd}",   /* word */
292    NULL,                   /* xdigit */    NULL,                   /* xdigit */
293    /* Negated cases */    /* Negated cases */
294    (uschar *)"\\P{L}",     /* ^alpha */    (uschar *)"\\P{L}",     /* ^alpha */
295    (uschar *)"\\P{Ll}",    /* ^lower */    (uschar *)"\\P{Ll}",    /* ^lower */
296    (uschar *)"\\P{Lu}",    /* ^upper */    (uschar *)"\\P{Lu}",    /* ^upper */
297    (uschar *)"\\P{Xan}",   /* ^alnum */    (uschar *)"\\P{Xan}",   /* ^alnum */
298    NULL,                   /* ^ascii */    NULL,                   /* ^ascii */
299    (uschar *)"\\H",        /* ^blank */    (uschar *)"\\H",        /* ^blank */
300    NULL,                   /* ^cntrl */    NULL,                   /* ^cntrl */
# Line 304  static const uschar *posix_substitutes[] Line 304  static const uschar *posix_substitutes[]
304    NULL,                   /* ^punct */    NULL,                   /* ^punct */
305    (uschar *)"\\P{Xps}",   /* ^space */   /* NOTE: Xps is POSIX space */    (uschar *)"\\P{Xps}",   /* ^space */   /* NOTE: Xps is POSIX space */
306    (uschar *)"\\P{Xwd}",   /* ^word */    (uschar *)"\\P{Xwd}",   /* ^word */
307    NULL                    /* ^xdigit */    NULL                    /* ^xdigit */
308  };  };
309  #define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *))  #define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *))
310  #endif  #endif
311    
312  #define STRING(a)  # a  #define STRING(a)  # a
313  #define XSTRING(s) STRING(s)  #define XSTRING(s) STRING(s)
# Line 407  static const char error_texts[] = Line 407  static const char error_texts[] =
407    /* 65 */    /* 65 */
408    "different names for subpatterns of the same number are not allowed\0"    "different names for subpatterns of the same number are not allowed\0"
409    "(*MARK) must have an argument\0"    "(*MARK) must have an argument\0"
410    "this version of PCRE is not compiled with PCRE_UCP support\0"    "this version of PCRE is not compiled with PCRE_UCP support\0"
411    ;    ;
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 2407  Arguments: Line 2407  Arguments:
2407    ptype        the property type    ptype        the property type
2408    pdata        the data for the type    pdata        the data for the type
2409    negated      TRUE if it's a negated property (\P or \p{^)    negated      TRUE if it's a negated property (\P or \p{^)
2410    
2411  Returns:       TRUE if auto-possessifying is OK  Returns:       TRUE if auto-possessifying is OK
2412  */  */
2413    
2414  static BOOL  static BOOL
2415  check_char_prop(int c, int ptype, int pdata, BOOL negated)  check_char_prop(int c, int ptype, int pdata, BOOL negated)
# Line 2453  switch(ptype) Line 2453  switch(ptype)
2453            _pcre_ucp_gentype[prop->chartype] == ucp_N ||            _pcre_ucp_gentype[prop->chartype] == ucp_N ||
2454            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
2455    }    }
2456  return FALSE;  return FALSE;
2457  }  }
2458  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2459    
# Line 2478  Returns: TRUE if possessifying is Line 2478  Returns: TRUE if possessifying is
2478  */  */
2479    
2480  static BOOL  static BOOL
2481  check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr,  check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr,
2482    int options, compile_data *cd)    int options, compile_data *cd)
2483  {  {
2484  int c, next;  int c, next;
# Line 2549  the next item is a character. */ Line 2549  the next item is a character. */
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    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
2554  #else  #else
2555    c = *previous;    c = *previous;
2556  #endif  #endif
2557    return c != 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    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
2566  #else  #else
2567    c = *previous;    c = *previous;
2568  #endif  #endif
2569    if (c == next) return FALSE;    if (c == next) return FALSE;
2570  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2571    if (utf8)    if (utf8)
# Line 2603  if (next >= 0) switch(op_code) Line 2603  if (next >= 0) switch(op_code)
2603    else    else
2604  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
2605    return (c == 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.    /* 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. */    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 2673  if (next >= 0) switch(op_code) Line 2673  if (next >= 0) switch(op_code)
2673  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2674    case OP_PROP:    case OP_PROP:
2675    return check_char_prop(next, previous[0], previous[1], FALSE);    return check_char_prop(next, previous[0], previous[1], FALSE);
2676    
2677    case OP_NOTPROP:    case OP_NOTPROP:
2678    return check_char_prop(next, previous[0], previous[1], TRUE);    return check_char_prop(next, previous[0], previous[1], TRUE);
2679  #endif  #endif
# Line 2683  if (next >= 0) switch(op_code) Line 2683  if (next >= 0) switch(op_code)
2683    }    }
2684    
2685    
2686  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
2687  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
2688  generated only when PCRE_UCP is *not* set, that is, when only ASCII  generated only when PCRE_UCP is *not* set, that is, when only ASCII
2689  characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are  characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are
2690  replaced by OP_PROP codes when PCRE_UCP is set. */  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    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
2698  #else  #else
2699    c = *previous;    c = *previous;
2700  #endif  #endif
2701    switch(-next)    switch(-next)
2702      {      {
2703      case ESC_d:      case ESC_d:
# Line 2761  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      /* 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      their substitutions and process them. The result will always be either
2767      -ESC_p or -ESC_P. Then fall through to process those values. */      -ESC_p or -ESC_P. Then fall through to process those values. */
2768    
2769  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2770      case ESC_du:      case ESC_du:
2771      case ESC_DU:      case ESC_DU:
# Line 2780  switch(op_code) Line 2780  switch(op_code)
2780        if (temperrorcode != 0) return FALSE;        if (temperrorcode != 0) return FALSE;
2781        ptr++;    /* For compatibility */        ptr++;    /* For compatibility */
2782        }        }
2783      /* Fall through */      /* Fall through */
2784    
2785      case ESC_p:      case ESC_p:
2786      case ESC_P:      case ESC_P:
2787        {        {
2788        int ptype, pdata, errorcodeptr;        int ptype, pdata, errorcodeptr;
2789        BOOL negated;        BOOL negated;
2790    
2791        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
2792        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);
2793        if (ptype < 0) return FALSE;        if (ptype < 0) return FALSE;
2794        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
2795    
2796        /* If the property item is optional, we have to give up. (When generated        /* If the property item is optional, we have to give up. (When generated
2797        from \d etc by PCRE_UCP, this test will have been applied much earlier,        from \d etc by PCRE_UCP, this test will have been applied much earlier,
2798        to the original \d etc. At this point, ptr will point to a zero byte. */        to the original \d etc. At this point, ptr will point to a zero byte. */
2799    
2800        if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||        if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
2801          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)
2802            return FALSE;            return FALSE;
2803    
2804        /* Do the property check. */        /* Do the property check. */
2805    
2806        return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);        return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);
2807        }        }
2808  #endif  #endif
2809    
2810      default:      default:
2811      return FALSE;      return FALSE;
2812      }      }
2813    
2814    /* In principle, support for Unicode properties should be integrated here as    /* 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    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    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,    combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set,
2818    these op-codes are never generated.) */    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 ||
# Line 2831  switch(op_code) Line 2831  switch(op_code)
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 ||    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||
2835           next == -ESC_w || next == -ESC_v || next == -ESC_R;           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:    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    
# Line 2846  switch(op_code) Line 2846  switch(op_code)
2846    return next == -ESC_v || next == -ESC_R;    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 ||    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||
2850           next == -ESC_v || next == -ESC_R;           next == -ESC_v || next == -ESC_R;
2851    
2852    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
# Line 2982  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    /* If we are at the end of a nested substitution, revert to the outer level
2986    string. Nesting only happens one level deep. */    string. Nesting only happens one level deep. */
2987    
2988    if (c == 0 && nestptr != NULL)    if (c == 0 && nestptr != NULL)
# Line 3289  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 3358  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          /* When PCRE_UCP is set, some of the POSIX classes are converted to          /* When PCRE_UCP is set, some of the POSIX classes are converted to
3363          different escape sequences that use Unicode properties. */          different escape sequences that use Unicode properties. */
3364    
3365  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3366          if ((options & PCRE_UCP) != 0)          if ((options & PCRE_UCP) != 0)
3367            {            {
3368            int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);            int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);
3369            if (posix_substitutes[pc] != NULL)            if (posix_substitutes[pc] != NULL)
3370              {              {
3371              nestptr = tempptr + 1;              nestptr = tempptr + 1;
3372              ptr = posix_substitutes[pc] - 1;              ptr = posix_substitutes[pc] - 1;
3373              continue;              continue;
3374              }              }
3375            }            }
3376  #endif  #endif
3377          /* In the non-UCP case, we build the bit map for the POSIX class in a          /* In the non-UCP case, we build the bit map for the POSIX class in a
3378          chunk of local store because we may be adding and subtracting from it,          chunk of local store because we may be adding and subtracting from it,
3379          and we don't want to subtract bits that may be in the main map already.          and we don't want to subtract bits that may be in the main map already.
# Line 3460  for (;; ptr++) Line 3460  for (;; ptr++)
3460              case ESC_SU:              case ESC_SU:
3461              nestptr = ptr;              nestptr = ptr;
3462              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
3463              class_charcount -= 2;                /* Undo! */              class_charcount -= 2;                /* Undo! */
3464              continue;              continue;
3465  #endif  #endif
3466              case ESC_d:              case ESC_d:
# Line 3911  for (;; ptr++) Line 3911  for (;; ptr++)
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 3991  for (;; ptr++) Line 3991  for (;; ptr++)
3991        }        }
3992  #endif  #endif
3993    
3994      /* If there are no characters > 255, or they are all to be included or      /* If there are no characters > 255, or they are all to be included or
3995      excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the      excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the
3996      whole class was negated and whether there were negative specials such as \S      whole class was negated and whether there were negative specials such as \S
3997      (non-UCP) in the class. Then copy the 32-byte map into the code vector,      (non-UCP) in the class. Then copy the 32-byte map into the code vector,
# Line 5795  for (;; ptr++) Line 5795  for (;; ptr++)
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 in the      are arranged to be the negation of the corresponding OP_values in the
5799      default case when PCRE_UCP is not set. For the back references, the values      default case when PCRE_UCP is not set. For the back references, the values
5800      are ESC_REF plus the reference number. Only back references and those types      are ESC_REF plus the reference number. Only back references and those types
5801      that consume a character may be repeated. We can test for values between      that consume a character may be repeated. We can test for values between
# Line 5973  for (;; ptr++) Line 5973  for (;; ptr++)
5973            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
5974            }            }
5975          else          else
5976  #endif  #endif
5977            {            {
5978            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
5979            *code++ = -c;            *code++ = -c;
5980            }            }
5981          }          }
5982        continue;        continue;
5983        }        }
# Line 6809  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;  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. */
# Line 6835  if (utf8) Line 6835  if (utf8)
6835  if ((options & PCRE_UCP) != 0)  if ((options & PCRE_UCP) != 0)
6836    {    {
6837    errorcode = ERR67;    errorcode = ERR67;
6838    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
6839    }    }
6840  #endif  #endif
6841    
6842  /* Check validity of \R options. */  /* Check validity of \R options. */

Legend:
Removed from v.533  
changed lines
  Added in v.535

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12