/[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 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 264 by ph10, Tue Nov 13 11:07:16 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,      0,      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 125  static const short int escapes[] = { Line 140  static const short int escapes[] = {
140  #endif  #endif
141    
142    
143  /* Tables of names of POSIX character classes and their lengths. The list is  /* Table of special "verbs" like (*PRUNE). This is a short table, so it is
144  terminated by a zero length entry. The first three must be alpha, lower, upper,  searched linearly. Put all the names into a single string, in order to reduce
145  as this is assumed for handling case independence. */  the number of relocations when a shared library is dynamically linked. */
146    
147  static const char *const posix_names[] = {  typedef struct verbitem {
148    "alpha", "lower", "upper",    int   len;
149    "alnum", "ascii", "blank", "cntrl", "digit", "graph",    int   op;
150    "print", "punct", "space", "word",  "xdigit" };  } verbitem;
151    
152    static const char verbnames[] =
153      "ACCEPT\0"
154      "COMMIT\0"
155      "F\0"
156      "FAIL\0"
157      "PRUNE\0"
158      "SKIP\0"
159      "THEN";
160    
161    static verbitem verbs[] = {
162      { 6, OP_ACCEPT },
163      { 6, OP_COMMIT },
164      { 1, OP_FAIL },
165      { 4, OP_FAIL },
166      { 5, OP_PRUNE },
167      { 4, OP_SKIP  },
168      { 4, OP_THEN  }
169    };
170    
171    static int verbcount = sizeof(verbs)/sizeof(verbitem);
172    
173    
174    /* Tables of names of POSIX character classes and their lengths. The names are
175    now all in a single string, to reduce the number of relocations when a shared
176    library is dynamically loaded. The list of lengths is terminated by a zero
177    length entry. The first three must be alpha, lower, upper, as this is assumed
178    for handling case independence. */
179    
180    static const char posix_names[] =
181      "alpha\0"  "lower\0"  "upper\0"  "alnum\0"  "ascii\0"  "blank\0"
182      "cntrl\0"  "digit\0"  "graph\0"  "print\0"  "punct\0"  "space\0"
183      "word\0"   "xdigit";
184    
185  static const uschar posix_name_lengths[] = {  static const uschar posix_name_lengths[] = {
186    5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 };    5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 };
# Line 171  static const int posix_class_maps[] = { Line 219  static const int posix_class_maps[] = {
219  /* The texts of compile-time error messages. These are "char *" because they  /* The texts of compile-time error messages. These are "char *" because they
220  are passed to the outside world. Do not ever re-use any error number, because  are passed to the outside world. Do not ever re-use any error number, because
221  they are documented. Always add a new error instead. Messages marked DEAD below  they are documented. Always add a new error instead. Messages marked DEAD below
222  are no longer used. */  are no longer used. This used to be a table of strings, but in order to reduce
223    the number of relocations needed when a shared library is loaded dynamically,
224  static const char *error_texts[] = {  it is now one long string. We cannot use a table of offsets, because the
225    "no error",  lengths of inserts such as XSTRING(MAX_NAME_SIZE) are not known. Instead, we
226    "\\ at end of pattern",  simply count through to the one we want - this isn't a performance issue
227    "\\c at end of pattern",  because these strings are used only when there is a compilation error. */
228    "unrecognized character follows \\",  
229    "numbers out of order in {} quantifier",  static const char error_texts[] =
230      "no error\0"
231      "\\ at end of pattern\0"
232      "\\c at end of pattern\0"
233      "unrecognized character follows \\\0"
234      "numbers out of order in {} quantifier\0"
235    /* 5 */    /* 5 */
236    "number too big in {} quantifier",    "number too big in {} quantifier\0"
237    "missing terminating ] for character class",    "missing terminating ] for character class\0"
238    "invalid escape sequence in character class",    "invalid escape sequence in character class\0"
239    "range out of order in character class",    "range out of order in character class\0"
240    "nothing to repeat",    "nothing to repeat\0"
241    /* 10 */    /* 10 */
242    "operand of unlimited repeat could match the empty string",  /** DEAD **/    "operand of unlimited repeat could match the empty string\0"  /** DEAD **/
243    "internal error: unexpected repeat",    "internal error: unexpected repeat\0"
244    "unrecognized character after (?",    "unrecognized character after (?\0"
245    "POSIX named classes are supported only within a class",    "POSIX named classes are supported only within a class\0"
246    "missing )",    "missing )\0"
247    /* 15 */    /* 15 */
248    "reference to non-existent subpattern",    "reference to non-existent subpattern\0"
249    "erroffset passed as NULL",    "erroffset passed as NULL\0"
250    "unknown option bit(s) set",    "unknown option bit(s) set\0"
251    "missing ) after comment",    "missing ) after comment\0"
252    "parentheses nested too deeply",  /** DEAD **/    "parentheses nested too deeply\0"  /** DEAD **/
253    /* 20 */    /* 20 */
254    "regular expression too large",    "regular expression is too large\0"
255    "failed to get memory",    "failed to get memory\0"
256    "unmatched parentheses",    "unmatched parentheses\0"
257    "internal error: code overflow",    "internal error: code overflow\0"
258    "unrecognized character after (?<",    "unrecognized character after (?<\0"
259    /* 25 */    /* 25 */
260    "lookbehind assertion is not fixed length",    "lookbehind assertion is not fixed length\0"
261    "malformed number or name after (?(",    "malformed number or name after (?(\0"
262    "conditional group contains more than two branches",    "conditional group contains more than two branches\0"
263    "assertion expected after (?(",    "assertion expected after (?(\0"
264    "(?R or (?[+-]digits must be followed by )",    "(?R or (?[+-]digits must be followed by )\0"
265    /* 30 */    /* 30 */
266    "unknown POSIX class name",    "unknown POSIX class name\0"
267    "POSIX collating elements are not supported",    "POSIX collating elements are not supported\0"
268    "this version of PCRE is not compiled with PCRE_UTF8 support",    "this version of PCRE is not compiled with PCRE_UTF8 support\0"
269    "spare error",  /** DEAD **/    "spare error\0"  /** DEAD **/
270    "character value in \\x{...} sequence is too large",    "character value in \\x{...} sequence is too large\0"
271    /* 35 */    /* 35 */
272    "invalid condition (?(0)",    "invalid condition (?(0)\0"
273    "\\C not allowed in lookbehind assertion",    "\\C not allowed in lookbehind assertion\0"
274    "PCRE does not support \\L, \\l, \\N, \\U, or \\u",    "PCRE does not support \\L, \\l, \\N, \\U, or \\u\0"
275    "number after (?C is > 255",    "number after (?C is > 255\0"
276    "closing ) for (?C expected",    "closing ) for (?C expected\0"
277    /* 40 */    /* 40 */
278    "recursive call could loop indefinitely",    "recursive call could loop indefinitely\0"
279    "unrecognized character after (?P",    "unrecognized character after (?P\0"
280    "syntax error in subpattern name (missing terminator)",    "syntax error in subpattern name (missing terminator)\0"
281    "two named subpatterns have the same name",    "two named subpatterns have the same name\0"
282    "invalid UTF-8 string",    "invalid UTF-8 string\0"
283    /* 45 */    /* 45 */
284    "support for \\P, \\p, and \\X has not been compiled",    "support for \\P, \\p, and \\X has not been compiled\0"
285    "malformed \\P or \\p sequence",    "malformed \\P or \\p sequence\0"
286    "unknown property name after \\P or \\p",    "unknown property name after \\P or \\p\0"
287    "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)",    "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)\0"
288    "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")",    "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")\0"
289    /* 50 */    /* 50 */
290    "repeated subpattern is too long",    "repeated subpattern is too long\0"    /** DEAD **/
291    "octal value is greater than \\377 (not in UTF-8 mode)",    "octal value is greater than \\377 (not in UTF-8 mode)\0"
292    "internal error: overran compiling workspace",    "internal error: overran compiling workspace\0"
293    "internal error: previously-checked referenced subpattern not found",    "internal error: previously-checked referenced subpattern not found\0"
294    "DEFINE group contains more than one branch",    "DEFINE group contains more than one branch\0"
295    /* 55 */    /* 55 */
296    "repeating a DEFINE group is not allowed",    "repeating a DEFINE group is not allowed\0"
297    "inconsistent NEWLINE options",    "inconsistent NEWLINE options\0"
298    "\\g is not followed by an (optionally braced) non-zero number",    "\\g is not followed by a braced name or an optionally braced non-zero number\0"
299    "(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number"    "(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number\0"
300  };    "(*VERB) with an argument is not supported\0"
301      /* 60 */
302      "(*VERB) not recognized\0"
303      "number is too big";
304    
305    
306  /* 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 374  static const unsigned char ebcdic_charta Line 430  static const unsigned char ebcdic_charta
430  /* Definition to allow mutual recursion */  /* Definition to allow mutual recursion */
431    
432  static BOOL  static BOOL
433    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, int, int *,    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,
434      int *, branch_chain *, compile_data *, int *);      int *, int *, branch_chain *, compile_data *, int *);
435    
436    
437    
438  /*************************************************  /*************************************************
439    *            Find an error text                  *
440    *************************************************/
441    
442    /* The error texts are now all in one long string, to save on relocations. As
443    some of the text is of unknown length, we can't use a table of offsets.
444    Instead, just count through the strings. This is not a performance issue
445    because it happens only when there has been a compilation error.
446    
447    Argument:   the error number
448    Returns:    pointer to the error string
449    */
450    
451    static const char *
452    find_error_text(int n)
453    {
454    const char *s = error_texts;
455    for (; n > 0; n--) while (*s++ != 0);
456    return s;
457    }
458    
459    
460    /*************************************************
461  *            Handle escapes                      *  *            Handle escapes                      *
462  *************************************************/  *************************************************/
463    
# Line 400  Arguments: Line 478  Arguments:
478    
479  Returns:         zero or positive => a data character  Returns:         zero or positive => a data character
480                   negative => a special escape sequence                   negative => a special escape sequence
481                   on error, errorptr is set                   on error, errorcodeptr is set
482  */  */
483    
484  static int  static int
# Line 453  else Line 531  else
531    
532      /* \g must be followed by a number, either plain or braced. If positive, it      /* \g must be followed by a number, either plain or braced. If positive, it
533      is an absolute backreference. If negative, it is a relative backreference.      is an absolute backreference. If negative, it is a relative backreference.
534      This is a Perl 5.10 feature. */      This is a Perl 5.10 feature. Perl 5.10 also supports \g{name} as a
535        reference to a named group. This is part of Perl's movement towards a
536        unified syntax for back references. As this is synonymous with \k{name}, we
537        fudge it up by pretending it really was \k. */
538    
539      case 'g':      case 'g':
540      if (ptr[1] == '{')      if (ptr[1] == '{')
541        {        {
542          const uschar *p;
543          for (p = ptr+2; *p != 0 && *p != '}'; p++)
544            if (*p != '-' && (digitab[*p] & ctype_digit) == 0) break;
545          if (*p != 0 && *p != '}')
546            {
547            c = -ESC_k;
548            break;
549            }
550        braced = TRUE;        braced = TRUE;
551        ptr++;        ptr++;
552        }        }
# Line 474  else Line 563  else
563      while ((digitab[ptr[1]] & ctype_digit) != 0)      while ((digitab[ptr[1]] & ctype_digit) != 0)
564        c = c * 10 + *(++ptr) - '0';        c = c * 10 + *(++ptr) - '0';
565    
566        if (c < 0)
567          {
568          *errorcodeptr = ERR61;
569          break;
570          }
571    
572      if (c == 0 || (braced && *(++ptr) != '}'))      if (c == 0 || (braced && *(++ptr) != '}'))
573        {        {
574        *errorcodeptr = ERR57;        *errorcodeptr = ERR57;
575        return 0;        break;
576        }        }
577    
578      if (negated)      if (negated)
# Line 485  else Line 580  else
580        if (c > bracount)        if (c > bracount)
581          {          {
582          *errorcodeptr = ERR15;          *errorcodeptr = ERR15;
583          return 0;          break;
584          }          }
585        c = bracount - (c - 1);        c = bracount - (c - 1);
586        }        }
# Line 514  else Line 609  else
609        c -= '0';        c -= '0';
610        while ((digitab[ptr[1]] & ctype_digit) != 0)        while ((digitab[ptr[1]] & ctype_digit) != 0)
611          c = c * 10 + *(++ptr) - '0';          c = c * 10 + *(++ptr) - '0';
612          if (c < 0)
613            {
614            *errorcodeptr = ERR61;
615            break;
616            }
617        if (c < 10 || c <= bracount)        if (c < 10 || c <= bracount)
618          {          {
619          c = -(ESC_REF + c);          c = -(ESC_REF + c);
# Line 609  else Line 709  else
709      if (c == 0)      if (c == 0)
710        {        {
711        *errorcodeptr = ERR2;        *errorcodeptr = ERR2;
712        return 0;        break;
713        }        }
714    
715  #ifndef EBCDIC  /* ASCII coding */  #ifndef EBCDIC  /* ASCII coding */
# Line 685  if (c == '{') Line 785  if (c == '{')
785      *negptr = TRUE;      *negptr = TRUE;
786      ptr++;      ptr++;
787      }      }
788    for (i = 0; i < sizeof(name) - 1; i++)    for (i = 0; i < (int)sizeof(name) - 1; i++)
789      {      {
790      c = *(++ptr);      c = *(++ptr);
791      if (c == 0) goto ERROR_RETURN;      if (c == 0) goto ERROR_RETURN;
# Line 714  top = _pcre_utt_size; Line 814  top = _pcre_utt_size;
814  while (bot < top)  while (bot < top)
815    {    {
816    i = (bot + top) >> 1;    i = (bot + top) >> 1;
817    c = strcmp(name, _pcre_utt[i].name);    c = strcmp(name, _pcre_utt_names + _pcre_utt[i].name_offset);
818    if (c == 0)    if (c == 0)
819      {      {
820      *dptr = _pcre_utt[i].value;      *dptr = _pcre_utt[i].value;
# Line 888  for (; *ptr != 0; ptr++) Line 988  for (; *ptr != 0; ptr++)
988      {      {
989      while (*(++ptr) != ']')      while (*(++ptr) != ']')
990        {        {
991          if (*ptr == 0) return -1;
992        if (*ptr == '\\')        if (*ptr == '\\')
993          {          {
994          if (*(++ptr) == 0) return -1;          if (*(++ptr) == 0) return -1;
# Line 915  for (; *ptr != 0; ptr++) Line 1016  for (; *ptr != 0; ptr++)
1016    /* An opening parens must now be a real metacharacter */    /* An opening parens must now be a real metacharacter */
1017    
1018    if (*ptr != '(') continue;    if (*ptr != '(') continue;
1019    if (ptr[1] != '?')    if (ptr[1] != '?' && ptr[1] != '*')
1020      {      {
1021      count++;      count++;
1022      if (name == NULL && count == lorn) return count;      if (name == NULL && count == lorn) return count;
# Line 1043  for (;;) Line 1144  for (;;)
1144    {    {
1145    int d;    int d;
1146    register int op = *cc;    register int op = *cc;
   
1147    switch (op)    switch (op)
1148      {      {
1149      case OP_CBRA:      case OP_CBRA:
# Line 1132  for (;;) Line 1232  for (;;)
1232    
1233      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1234      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1235        if (cc[3] == OP_PROP || cc[3] == OP_NOTPROP) cc += 2;
1236      cc += 4;      cc += 4;
1237      break;      break;
1238    
# Line 1240  for (;;) Line 1341  for (;;)
1341      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1342      }      }
1343    
1344    /* In UTF-8 mode, opcodes that are followed by a character may be followed by    /* Otherwise, we can get the item's length from the table, except that for
1345    a multi-byte character. The length in the table is a minimum, so we have to    repeated character types, we have to test for \p and \P, which have an extra
1346    arrange to skip the extra bytes. */    two bytes of parameters. */
1347    
1348    else    else
1349      {      {
1350        switch(c)
1351          {
1352          case OP_TYPESTAR:
1353          case OP_TYPEMINSTAR:
1354          case OP_TYPEPLUS:
1355          case OP_TYPEMINPLUS:
1356          case OP_TYPEQUERY:
1357          case OP_TYPEMINQUERY:
1358          case OP_TYPEPOSSTAR:
1359          case OP_TYPEPOSPLUS:
1360          case OP_TYPEPOSQUERY:
1361          if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1362          break;
1363    
1364          case OP_TYPEUPTO:
1365          case OP_TYPEMINUPTO:
1366          case OP_TYPEEXACT:
1367          case OP_TYPEPOSUPTO:
1368          if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
1369          break;
1370          }
1371    
1372        /* Add in the fixed length from the table */
1373    
1374      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1375    
1376      /* In UTF-8 mode, opcodes that are followed by a character may be followed by
1377      a multi-byte character. The length in the table is a minimum, so we have to
1378      arrange to skip the extra bytes. */
1379    
1380  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1381      if (utf8) switch(c)      if (utf8) switch(c)
1382        {        {
# Line 1304  for (;;) Line 1434  for (;;)
1434    
1435    if (c == OP_XCLASS) code += GET(code, 1);    if (c == OP_XCLASS) code += GET(code, 1);
1436    
1437    /* Otherwise, we get the item's length from the table. In UTF-8 mode, opcodes    /* Otherwise, we can get the item's length from the table, except that for
1438    that are followed by a character may be followed by a multi-byte character.    repeated character types, we have to test for \p and \P, which have an extra
1439    The length in the table is a minimum, so we have to arrange to skip the extra    two bytes of parameters. */
   bytes. */  
1440    
1441    else    else
1442      {      {
1443        switch(c)
1444          {
1445          case OP_TYPESTAR:
1446          case OP_TYPEMINSTAR:
1447          case OP_TYPEPLUS:
1448          case OP_TYPEMINPLUS:
1449          case OP_TYPEQUERY:
1450          case OP_TYPEMINQUERY:
1451          case OP_TYPEPOSSTAR:
1452          case OP_TYPEPOSPLUS:
1453          case OP_TYPEPOSQUERY:
1454          if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1455          break;
1456    
1457          case OP_TYPEPOSUPTO:
1458          case OP_TYPEUPTO:
1459          case OP_TYPEMINUPTO:
1460          case OP_TYPEEXACT:
1461          if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
1462          break;
1463          }
1464    
1465        /* Add in the fixed length from the table */
1466    
1467      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1468    
1469        /* In UTF-8 mode, opcodes that are followed by a character may be followed
1470        by a multi-byte character. The length in the table is a minimum, so we have
1471        to arrange to skip the extra bytes. */
1472    
1473  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1474      if (utf8) switch(c)      if (utf8) switch(c)
1475        {        {
# Line 1371  for (code = first_significant_code(code Line 1529  for (code = first_significant_code(code
1529    
1530    c = *code;    c = *code;
1531    
1532    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE)    /* Groups with zero repeats can of course be empty; skip them. */
1533    
1534      if (c == OP_BRAZERO || c == OP_BRAMINZERO)
1535        {
1536        code += _pcre_OP_lengths[c];
1537        do code += GET(code, 1); while (*code == OP_ALT);
1538        c = *code;
1539        continue;
1540        }
1541    
1542      /* For other groups, scan the branches. */
1543    
1544      if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE || c == OP_COND)
1545      {      {
1546      BOOL empty_branch;      BOOL empty_branch;
1547      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */
# Line 1387  for (code = first_significant_code(code Line 1557  for (code = first_significant_code(code
1557        }        }
1558      while (*code == OP_ALT);      while (*code == OP_ALT);
1559      if (!empty_branch) return FALSE;   /* All branches are non-empty */      if (!empty_branch) return FALSE;   /* All branches are non-empty */
1560        c = *code;
     /* Move past the KET and fudge things so that the increment in the "for"  
     above has no effect. */  
   
     c = OP_END;  
     code += 1 + LINK_SIZE - _pcre_OP_lengths[c];  
1561      continue;      continue;
1562      }      }
1563    
# Line 1400  for (code = first_significant_code(code Line 1565  for (code = first_significant_code(code
1565    
1566    switch (c)    switch (c)
1567      {      {
1568      /* Check for quantifiers after a class */      /* Check for quantifiers after a class. XCLASS is used for classes that
1569        cannot be represented just by a bit map. This includes negated single
1570        high-valued characters. The length in _pcre_OP_lengths[] is zero; the
1571        actual length is stored in the compiled code, so we must update "code"
1572        here. */
1573    
1574  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1575      case OP_XCLASS:      case OP_XCLASS:
1576      ccode = code + GET(code, 1);      ccode = code += GET(code, 1);
1577      goto CHECK_CLASS_REPEAT;      goto CHECK_CLASS_REPEAT;
1578  #endif  #endif
1579    
# Line 1466  for (code = first_significant_code(code Line 1635  for (code = first_significant_code(code
1635      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1636      return FALSE;      return FALSE;
1637    
1638        /* These are going to continue, as they may be empty, but we have to
1639        fudge the length for the \p and \P cases. */
1640    
1641        case OP_TYPESTAR:
1642        case OP_TYPEMINSTAR:
1643        case OP_TYPEPOSSTAR:
1644        case OP_TYPEQUERY:
1645        case OP_TYPEMINQUERY:
1646        case OP_TYPEPOSQUERY:
1647        if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1648        break;
1649    
1650        /* Same for these */
1651    
1652        case OP_TYPEUPTO:
1653        case OP_TYPEMINUPTO:
1654        case OP_TYPEPOSUPTO:
1655        if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
1656        break;
1657    
1658      /* End of branch */      /* End of branch */
1659    
1660      case OP_KET:      case OP_KET:
# Line 1582  Returns:     a value representing the na Line 1771  Returns:     a value representing the na
1771  static int  static int
1772  check_posix_name(const uschar *ptr, int len)  check_posix_name(const uschar *ptr, int len)
1773  {  {
1774    const char *pn = posix_names;
1775  register int yield = 0;  register int yield = 0;
1776  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
1777    {    {
1778    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
1779      strncmp((const char *)ptr, posix_names[yield], len) == 0) return yield;      strncmp((const char *)ptr, pn, len) == 0) return yield;
1780      pn += posix_name_lengths[yield] + 1;
1781    yield++;    yield++;
1782    }    }
1783  return -1;  return -1;
# Line 1628  adjust_recurse(uschar *group, int adjust Line 1819  adjust_recurse(uschar *group, int adjust
1819    uschar *save_hwm)    uschar *save_hwm)
1820  {  {
1821  uschar *ptr = group;  uschar *ptr = group;
1822    
1823  while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL)  while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL)
1824    {    {
1825    int offset;    int offset;
# Line 1926  if (next >= 0) switch(op_code) Line 2118  if (next >= 0) switch(op_code)
2118    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2119    return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;    return next <= 127 && (cd->ctypes[next] & ctype_word) != 0;
2120    
2121      case OP_HSPACE:
2122      case OP_NOT_HSPACE:
2123      switch(next)
2124        {
2125        case 0x09:
2126        case 0x20:
2127        case 0xa0:
2128        case 0x1680:
2129        case 0x180e:
2130        case 0x2000:
2131        case 0x2001:
2132        case 0x2002:
2133        case 0x2003:
2134        case 0x2004:
2135        case 0x2005:
2136        case 0x2006:
2137        case 0x2007:
2138        case 0x2008:
2139        case 0x2009:
2140        case 0x200A:
2141        case 0x202f:
2142        case 0x205f:
2143        case 0x3000:
2144        return op_code != OP_HSPACE;
2145        default:
2146        return op_code == OP_HSPACE;
2147        }
2148    
2149      case OP_VSPACE:
2150      case OP_NOT_VSPACE:
2151      switch(next)
2152        {
2153        case 0x0a:
2154        case 0x0b:
2155        case 0x0c:
2156        case 0x0d:
2157        case 0x85:
2158        case 0x2028:
2159        case 0x2029:
2160        return op_code != OP_VSPACE;
2161        default:
2162        return op_code == OP_VSPACE;
2163        }
2164    
2165    default:    default:
2166    return FALSE;    return FALSE;
2167    }    }
# Line 1960  switch(op_code) Line 2196  switch(op_code)
2196      case ESC_W:      case ESC_W:
2197      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;      return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;
2198    
2199        case ESC_h:
2200        case ESC_H:
2201        switch(item)
2202          {
2203          case 0x09:
2204          case 0x20:
2205          case 0xa0:
2206          case 0x1680:
2207          case 0x180e:
2208          case 0x2000:
2209          case 0x2001:
2210          case 0x2002:
2211          case 0x2003:
2212          case 0x2004:
2213          case 0x2005:
2214          case 0x2006:
2215          case 0x2007:
2216          case 0x2008:
2217          case 0x2009:
2218          case 0x200A:
2219          case 0x202f:
2220          case 0x205f:
2221          case 0x3000:
2222          return -next != ESC_h;
2223          default:
2224          return -next == ESC_h;
2225          }
2226    
2227        case ESC_v:
2228        case ESC_V:
2229        switch(item)
2230          {
2231          case 0x0a:
2232          case 0x0b:
2233          case 0x0c:
2234          case 0x0d:
2235          case 0x85:
2236          case 0x2028:
2237          case 0x2029:
2238          return -next != ESC_v;
2239          default:
2240          return -next == ESC_v;
2241          }
2242    
2243      default:      default:
2244      return FALSE;      return FALSE;
2245      }      }
2246    
2247    case OP_DIGIT:    case OP_DIGIT:
2248    return next == -ESC_D || next == -ESC_s || next == -ESC_W;    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
2249             next == -ESC_h || next == -ESC_v;
2250    
2251    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
2252    return next == -ESC_d;    return next == -ESC_d;
# Line 1974  switch(op_code) Line 2255  switch(op_code)
2255    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return next == -ESC_S || next == -ESC_d || next == -ESC_w;
2256    
2257    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
2258    return next == -ESC_s;    return next == -ESC_s || next == -ESC_h || next == -ESC_v;
2259    
2260      case OP_HSPACE:
2261      return next == -ESC_S || next == -ESC_H || next == -ESC_d || next == -ESC_w;
2262    
2263      case OP_NOT_HSPACE:
2264      return next == -ESC_h;
2265    
2266      /* Can't have \S in here because VT matches \S (Perl anomaly) */
2267      case OP_VSPACE:
2268      return next == -ESC_V || next == -ESC_d || next == -ESC_w;
2269    
2270      case OP_NOT_VSPACE:
2271      return next == -ESC_v;
2272    
2273    case OP_WORDCHAR:    case OP_WORDCHAR:
2274    return next == -ESC_W || next == -ESC_s;    return next == -ESC_W || next == -ESC_s || next == -ESC_h || next == -ESC_v;
2275    
2276    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
2277    return next == -ESC_w || next == -ESC_d;    return next == -ESC_w || next == -ESC_d;
# Line 2089  req_caseopt = ((options & PCRE_CASELESS) Line 2383  req_caseopt = ((options & PCRE_CASELESS)
2383  for (;; ptr++)  for (;; ptr++)
2384    {    {
2385    BOOL negate_class;    BOOL negate_class;
2386      BOOL should_flip_negation;
2387    BOOL possessive_quantifier;    BOOL possessive_quantifier;
2388    BOOL is_quantifier;    BOOL is_quantifier;
2389    BOOL is_recurse;    BOOL is_recurse;
2390      BOOL reset_bracount;
2391    int class_charcount;    int class_charcount;
2392    int class_lastchar;    int class_lastchar;
2393    int newoptions;    int newoptions;
2394    int recno;    int recno;
2395    int refsign;    int refsign;
2396    int skipbytes;    int skipbytes;
2397    int subreqbyte;    int subreqbyte;
2398    int subfirstbyte;    int subfirstbyte;
# Line 2129  for (;; ptr++) Line 2425  for (;; ptr++)
2425      */      */
2426    
2427      if (code < last_code) code = last_code;      if (code < last_code) code = last_code;
2428    
2429        /* Paranoid check for integer overflow */
2430    
2431        if (OFLOW_MAX - *lengthptr < code - last_code)
2432          {
2433          *errorcodeptr = ERR20;
2434          goto FAILED;
2435          }
2436    
2437      *lengthptr += code - last_code;      *lengthptr += code - last_code;
2438      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));
2439    
# Line 2241  for (;; ptr++) Line 2546  for (;; ptr++)
2546      *ptrptr = ptr;      *ptrptr = ptr;
2547      if (lengthptr != NULL)      if (lengthptr != NULL)
2548        {        {
2549          if (OFLOW_MAX - *lengthptr < code - last_code)
2550            {
2551            *errorcodeptr = ERR20;
2552            goto FAILED;
2553            }
2554        *lengthptr += code - last_code;   /* To include callout length */        *lengthptr += code - last_code;   /* To include callout length */
2555        DPRINTF((">> end branch\n"));        DPRINTF((">> end branch\n"));
2556        }        }
# Line 2303  for (;; ptr++) Line 2613  for (;; ptr++)
2613        goto FAILED;        goto FAILED;
2614        }        }
2615    
2616      /* 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,
2617        if the first few characters (either before or after ^) are \Q\E or \E we
2618        skip them too. This makes for compatibility with Perl. */
2619    
2620      if ((c = *(++ptr)) == '^')      negate_class = FALSE;
2621        for (;;)
2622        {        {
       negate_class = TRUE;  
2623        c = *(++ptr);        c = *(++ptr);
2624          if (c == '\\')
2625            {
2626            if (ptr[1] == 'E') ptr++;
2627              else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3;
2628                else break;
2629            }
2630          else if (!negate_class && c == '^')
2631            negate_class = TRUE;
2632          else break;
2633        }        }
2634      else  
2635        {      /* If a class contains a negative special such as \S, we need to flip the
2636        negate_class = FALSE;      negation flag at the end, so that support for characters > 255 works
2637        }      correctly (they are all included in the class). */
2638    
2639        should_flip_negation = FALSE;
2640    
2641      /* 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
2642      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 2453  for (;; ptr++) Line 2776  for (;; ptr++)
2776        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
2777        case. Inside a class (and only there) it is treated as backspace.        case. Inside a class (and only there) it is treated as backspace.
2778        Elsewhere it marks a word boundary. Other escapes have preset maps ready        Elsewhere it marks a word boundary. Other escapes have preset maps ready
2779        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
2780        character in them, so set class_charcount bigger than one. */        character in them, so set class_charcount bigger than one. */
2781    
2782        if (c == '\\')        if (c == '\\')
# Line 2473  for (;; ptr++) Line 2796  for (;; ptr++)
2796            else inescq = TRUE;            else inescq = TRUE;
2797            continue;            continue;
2798            }            }
2799            else if (-c == ESC_E) continue;  /* Ignore orphan \E */
2800    
2801          if (c < 0)          if (c < 0)
2802            {            {
# Line 2488  for (;; ptr++) Line 2812  for (;; ptr++)
2812              continue;              continue;
2813    
2814              case ESC_D:              case ESC_D:
2815                should_flip_negation = TRUE;
2816              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit];
2817              continue;              continue;
2818    
# Line 2496  for (;; ptr++) Line 2821  for (;; ptr++)
2821              continue;              continue;
2822    
2823              case ESC_W:              case ESC_W:
2824                should_flip_negation = TRUE;
2825              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];
2826              continue;              continue;
2827    
# Line 2505  for (;; ptr++) Line 2831  for (;; ptr++)
2831              continue;              continue;
2832    
2833              case ESC_S:              case ESC_S:
2834                should_flip_negation = TRUE;
2835              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];
2836              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */
2837              continue;              continue;
# Line 2521  for (;; ptr++) Line 2848  for (;; ptr++)
2848            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||
2849                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;
2850    
2851              /* We need to deal with \H, \h, \V, and \v in both phases because
2852              they use extra memory. */
2853    
2854              if (-c == ESC_h)
2855                {
2856                SETBIT(classbits, 0x09); /* VT */
2857                SETBIT(classbits, 0x20); /* SPACE */
2858                SETBIT(classbits, 0xa0); /* NSBP */
2859    #ifdef SUPPORT_UTF8
2860                if (utf8)
2861                  {
2862                  class_utf8 = TRUE;
2863                  *class_utf8data++ = XCL_SINGLE;
2864                  class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);
2865                  *class_utf8data++ = XCL_SINGLE;
2866                  class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);
2867                  *class_utf8data++ = XCL_RANGE;
2868                  class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);
2869                  class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);
2870                  *class_utf8data++ = XCL_SINGLE;
2871                  class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);
2872                  *class_utf8data++ = XCL_SINGLE;
2873                  class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);
2874                  *class_utf8data++ = XCL_SINGLE;
2875                  class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);
2876                  }
2877    #endif
2878                continue;
2879                }
2880    
2881              if (-c == ESC_H)
2882                {
2883                for (c = 0; c < 32; c++)
2884                  {
2885                  int x = 0xff;
2886                  switch (c)
2887                    {
2888                    case 0x09/8: x ^= 1 << (0x09%8); break;
2889                    case 0x20/8: x ^= 1 << (0x20%8); break;
2890                    case 0xa0/8: x ^= 1 << (0xa0%8); break;
2891                    default: break;
2892                    }
2893                  classbits[c] |= x;
2894                  }
2895    
2896    #ifdef SUPPORT_UTF8
2897                if (utf8)
2898                  {
2899                  class_utf8 = TRUE;
2900                  *class_utf8data++ = XCL_RANGE;
2901                  class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2902                  class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);
2903                  *class_utf8data++ = XCL_RANGE;
2904                  class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);
2905                  class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);
2906                  *class_utf8data++ = XCL_RANGE;
2907                  class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);
2908                  class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);
2909                  *class_utf8data++ = XCL_RANGE;
2910                  class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);
2911                  class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);
2912                  *class_utf8data++ = XCL_RANGE;
2913                  class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);
2914                  class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);
2915                  *class_utf8data++ = XCL_RANGE;
2916                  class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);
2917                  class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);
2918                  *class_utf8data++ = XCL_RANGE;
2919                  class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);
2920                  class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2921                  }
2922    #endif
2923                continue;
2924                }
2925    
2926              if (-c == ESC_v)
2927                {
2928                SETBIT(classbits, 0x0a); /* LF */
2929                SETBIT(classbits, 0x0b); /* VT */
2930                SETBIT(classbits, 0x0c); /* FF */
2931                SETBIT(classbits, 0x0d); /* CR */
2932                SETBIT(classbits, 0x85); /* NEL */
2933    #ifdef SUPPORT_UTF8
2934                if (utf8)
2935                  {
2936                  class_utf8 = TRUE;
2937                  *class_utf8data++ = XCL_RANGE;
2938                  class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);
2939                  class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2940                  }
2941    #endif
2942                continue;
2943                }
2944    
2945              if (-c == ESC_V)
2946                {
2947                for (c = 0; c < 32; c++)
2948                  {
2949                  int x = 0xff;
2950                  switch (c)
2951                    {
2952                    case 0x0a/8: x ^= 1 << (0x0a%8);
2953                                 x ^= 1 << (0x0b%8);
2954                                 x ^= 1 << (0x0c%8);
2955                                 x ^= 1 << (0x0d%8);
2956                                 break;
2957                    case 0x85/8: x ^= 1 << (0x85%8); break;
2958                    default: break;
2959                    }
2960                  classbits[c] |= x;
2961                  }
2962    
2963    #ifdef SUPPORT_UTF8
2964                if (utf8)
2965                  {
2966                  class_utf8 = TRUE;
2967                  *class_utf8data++ = XCL_RANGE;
2968                  class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2969                  class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);
2970                  *class_utf8data++ = XCL_RANGE;
2971                  class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2972                  class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2973                  }
2974    #endif
2975                continue;
2976                }
2977    
2978            /* We need to deal with \P and \p in both phases. */            /* We need to deal with \P and \p in both phases. */
2979    
2980  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2572  for (;; ptr++) Line 3026  for (;; ptr++)
3026    
3027        oldptr = ptr;        oldptr = ptr;
3028    
3029          /* Remember \r or \n */
3030    
3031          if (c == '\r' || c == '\n') cd->external_flags |= PCRE_HASCRORLF;
3032    
3033          /* Check for range */
3034    
3035        if (!inescq && ptr[1] == '-')        if (!inescq && ptr[1] == '-')
3036          {          {
3037          int d;          int d;
# Line 2639  for (;; ptr++) Line 3099  for (;; ptr++)
3099    
3100          if (d == c) goto LONE_SINGLE_CHARACTER;  /* A few lines below */          if (d == c) goto LONE_SINGLE_CHARACTER;  /* A few lines below */
3101    
3102            /* Remember \r or \n */
3103    
3104            if (d == '\r' || d == '\n') cd->external_flags |= PCRE_HASCRORLF;
3105    
3106          /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless          /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless
3107          matching, we have to use an XCLASS with extra data items. Caseless          matching, we have to use an XCLASS with extra data items. Caseless
3108          matching for characters > 127 is available only if UCP support is          matching for characters > 127 is available only if UCP support is
# Line 2661  for (;; ptr++) Line 3125  for (;; ptr++)
3125              unsigned int origd = d;              unsigned int origd = d;
3126              while (get_othercase_range(&cc, origd, &occ, &ocd))              while (get_othercase_range(&cc, origd, &occ, &ocd))
3127                {                {
3128                if (occ >= c && ocd <= d) continue;  /* Skip embedded ranges */                if (occ >= (unsigned int)c &&
3129                      ocd <= (unsigned int)d)
3130                    continue;                          /* Skip embedded ranges */
3131    
3132                if (occ < c  && ocd >= c - 1)        /* Extend the basic range */                if (occ < (unsigned int)c  &&
3133                      ocd >= (unsigned int)c - 1)      /* Extend the basic range */
3134                  {                                  /* if there is overlap,   */                  {                                  /* if there is overlap,   */
3135                  c = occ;                           /* noting that if occ < c */                  c = occ;                           /* noting that if occ < c */
3136                  continue;                          /* we can't have ocd > d  */                  continue;                          /* we can't have ocd > d  */
3137                  }                                  /* because a subrange is  */                  }                                  /* because a subrange is  */
3138                if (ocd > d && occ <= d + 1)         /* always shorter than    */                if (ocd > (unsigned int)d &&
3139                      occ <= (unsigned int)d + 1)      /* always shorter than    */
3140                  {                                  /* the basic range.       */                  {                                  /* the basic range.       */
3141                  d = ocd;                  d = ocd;
3142                  continue;                  continue;
# Line 2788  for (;; ptr++) Line 3256  for (;; ptr++)
3256        goto FAILED;        goto FAILED;
3257        }        }
3258    
3259    
3260    /* This code has been disabled because it would mean that \s counts as
3261    an explicit \r or \n reference, and that's not really what is wanted. Now
3262    we set the flag only if there is a literal "\r" or "\n" in the class. */
3263    
3264    #if 0
3265        /* Remember whether \r or \n are in this class */
3266    
3267        if (negate_class)
3268          {
3269          if ((classbits[1] & 0x24) != 0x24) cd->external_flags |= PCRE_HASCRORLF;
3270          }
3271        else
3272          {
3273          if ((classbits[1] & 0x24) != 0) cd->external_flags |= PCRE_HASCRORLF;
3274          }
3275    #endif
3276    
3277    
3278      /* 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
3279      less than 256. In non-UTF-8 mode we can always optimize. In UTF-8 mode, we      less than 256. As long as there were no characters >= 128 and there was no
3280      can optimize the negative case only if there were no characters >= 128      use of \p or \P, in other words, no use of any XCLASS features, we can
3281      because OP_NOT and the related opcodes like OP_NOTSTAR operate on      optimize.
3282      single-bytes only. This is an historical hangover. Maybe one day we can  
3283      tidy these opcodes to handle multi-byte characters.      In UTF-8 mode, we can optimize the negative case only if there were no
3284        characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR
3285        operate on single-bytes only. This is an historical hangover. Maybe one day
3286        we can tidy these opcodes to handle multi-byte characters.
3287    
3288      The optimization throws away the bit map. We turn the item into a      The optimization throws away the bit map. We turn the item into a
3289      1-character OP_CHAR[NC] if it's positive, or OP_NOT if it's negative. Note      1-character OP_CHAR[NC] if it's positive, or OP_NOT if it's negative. Note
# Line 2803  for (;; ptr++) Line 3293  for (;; ptr++)
3293      reqbyte, save the previous value for reinstating. */      reqbyte, save the previous value for reinstating. */
3294    
3295  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3296      if (class_charcount == 1 &&      if (class_charcount == 1 && !class_utf8 &&
3297            (!utf8 ||        (!utf8 || !negate_class || class_lastchar < 128))
           (!class_utf8 && (!negate_class || class_lastchar < 128))))  
   
3298  #else  #else
3299      if (class_charcount == 1)      if (class_charcount == 1)
3300  #endif  #endif
# Line 2849  for (;; ptr++) Line 3337  for (;; ptr++)
3337      zeroreqbyte = reqbyte;      zeroreqbyte = reqbyte;
3338    
3339      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
3340      extended class, with its own opcode. If there are no characters < 256,      extended class, with its own opcode, unless there was a negated special
3341      we can omit the bitmap in the actual compiled code. */      such as \S in the class, because in that case all characters > 255 are in
3342        the class, so any that were explicitly given as well can be ignored. If
3343        (when there are explicit characters > 255 that must be listed) there are no
3344        characters < 256, we can omit the bitmap in the actual compiled code. */
3345    
3346  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3347      if (class_utf8)      if (class_utf8 && !should_flip_negation)
3348        {        {
3349        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */
3350        *code++ = OP_XCLASS;        *code++ = OP_XCLASS;
# Line 2879  for (;; ptr++) Line 3370  for (;; ptr++)
3370        }        }
3371  #endif  #endif
3372    
3373      /* If there are no characters > 255, negate the 32-byte map if necessary,      /* If there are no characters > 255, set the opcode to OP_CLASS or
3374      and copy it into the code vector. If this is the first thing in the branch,      OP_NCLASS, depending on whether the whole class was negated and whether
3375      there can be no first char setting, whatever the repeat count. Any reqbyte      there were negative specials such as \S in the class. Then copy the 32-byte
3376      setting must remain unchanged after any kind of repeat. */      map into the code vector, negating it if necessary. */
3377    
3378        *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;
3379      if (negate_class)      if (negate_class)
3380        {        {
       *code++ = OP_NCLASS;  
3381        if (lengthptr == NULL)    /* Save time in the pre-compile phase */        if (lengthptr == NULL)    /* Save time in the pre-compile phase */
3382          for (c = 0; c < 32; c++) code[c] = ~classbits[c];          for (c = 0; c < 32; c++) code[c] = ~classbits[c];
3383        }        }
3384      else      else
3385        {        {
       *code++ = OP_CLASS;  
3386        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
3387        }        }
3388      code += 32;      code += 32;
# Line 3078  for (;; ptr++) Line 3568  for (;; ptr++)
3568        /* All real repeats make it impossible to handle partial matching (maybe        /* All real repeats make it impossible to handle partial matching (maybe
3569        one day we will be able to remove this restriction). */        one day we will be able to remove this restriction). */
3570    
3571        if (repeat_max != 1) cd->nopartial = TRUE;        if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL;
3572    
3573        /* Combine the op_type with the repeat_type */        /* Combine the op_type with the repeat_type */
3574    
# Line 3228  for (;; ptr++) Line 3718  for (;; ptr++)
3718        /* All real repeats make it impossible to handle partial matching (maybe        /* All real repeats make it impossible to handle partial matching (maybe
3719        one day we will be able to remove this restriction). */        one day we will be able to remove this restriction). */
3720    
3721        if (repeat_max != 1) cd->nopartial = TRUE;        if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL;
3722    
3723        if (repeat_min == 0 && repeat_max == -1)        if (repeat_min == 0 && repeat_max == -1)
3724          *code++ = OP_CRSTAR + repeat_type;          *code++ = OP_CRSTAR + repeat_type;
# Line 3264  for (;; ptr++) Line 3754  for (;; ptr++)
3754          goto FAILED;          goto FAILED;
3755          }          }
3756    
       /* This is a paranoid check to stop integer overflow later on */  
   
       if (len > MAX_DUPLENGTH)  
         {  
         *errorcodeptr = ERR50;  
         goto FAILED;  
         }  
   
3757        /* 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
3758        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
3759        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 3360  for (;; ptr++) Line 3842  for (;; ptr++)
3842          if (repeat_min > 1)          if (repeat_min > 1)
3843            {            {
3844            /* 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
3845            just adjust the length as if we had. */            just adjust the length as if we had. Do some paranoid checks for
3846              potential integer overflow. */
3847    
3848            if (lengthptr != NULL)            if (lengthptr != NULL)
3849              *lengthptr += (repeat_min - 1)*length_prevgroup;              {
3850                int delta = (repeat_min - 1)*length_prevgroup;
3851                if ((double)(repeat_min - 1)*(double)length_prevgroup >
3852                                                                (double)INT_MAX ||
3853                    OFLOW_MAX - *lengthptr < delta)
3854                  {
3855                  *errorcodeptr = ERR20;
3856                  goto FAILED;
3857                  }
3858                *lengthptr += delta;
3859                }
3860    
3861            /* This is compiling for real */            /* This is compiling for real */
3862    
# Line 3401  for (;; ptr++) Line 3894  for (;; ptr++)
3894          /* 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
3895          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
3896          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
3897          add 2 + 2*LINKSIZE to allow for the nesting that occurs. */          add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some
3898            paranoid checks to avoid integer overflow. */
3899    
3900          if (lengthptr != NULL && repeat_max > 0)          if (lengthptr != NULL && repeat_max > 0)
3901            *lengthptr += repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -            {
3902              2 - 2*LINK_SIZE;  /* Last one doesn't nest */            int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -
3903                          2 - 2*LINK_SIZE;   /* Last one doesn't nest */
3904              if ((double)repeat_max *
3905                    (double)(length_prevgroup + 1 + 2 + 2*LINK_SIZE)
3906                      > (double)INT_MAX ||
3907                  OFLOW_MAX - *lengthptr < delta)
3908                {
3909                *errorcodeptr = ERR20;
3910                goto FAILED;
3911                }
3912              *lengthptr += delta;
3913              }
3914    
3915          /* This is compiling for real */          /* This is compiling for real */
3916    
# Line 3557  for (;; ptr++) Line 4062  for (;; ptr++)
4062      /* ===================================================================*/      /* ===================================================================*/
4063      /* Start of nested parenthesized sub-expression, or comment or lookahead or      /* Start of nested parenthesized sub-expression, or comment or lookahead or
4064      lookbehind or option setting or condition or all the other extended      lookbehind or option setting or condition or all the other extended
4065      parenthesis forms. First deal with the specials; all are introduced by ?,      parenthesis forms.  */
     and the appearance of any of them means that this is not a capturing  
     group. */  
4066    
4067      case '(':      case '(':
4068      newoptions = options;      newoptions = options;
4069      skipbytes = 0;      skipbytes = 0;
4070      bravalue = OP_CBRA;      bravalue = OP_CBRA;
4071      save_hwm = cd->hwm;      save_hwm = cd->hwm;
4072        reset_bracount = FALSE;
4073    
4074        /* First deal with various "verbs" that can be introduced by '*'. */
4075    
4076        if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0)
4077          {
4078          int i, namelen;
4079          const char *vn = verbnames;
4080          const uschar *name = ++ptr;
4081          previous = NULL;
4082          while ((cd->ctypes[*++ptr] & ctype_letter) != 0);
4083          if (*ptr == ':')
4084            {
4085            *errorcodeptr = ERR59;   /* Not supported */
4086            goto FAILED;
4087            }
4088          if (*ptr != ')')
4089            {
4090            *errorcodeptr = ERR60;
4091            goto FAILED;
4092            }
4093          namelen = ptr - name;
4094          for (i = 0; i < verbcount; i++)
4095            {
4096            if (namelen == verbs[i].len &&
4097                strncmp((char *)name, vn, namelen) == 0)
4098              {
4099              *code = verbs[i].op;
4100              if (*code++ == OP_ACCEPT) cd->had_accept = TRUE;
4101              break;
4102              }
4103            vn += verbs[i].len + 1;
4104            }
4105          if (i < verbcount) continue;
4106          *errorcodeptr = ERR60;
4107          goto FAILED;
4108          }
4109    
4110        /* Deal with the extended parentheses; all are introduced by '?', and the
4111        appearance of any of them means that this is not a capturing group. */
4112    
4113      if (*(++ptr) == '?')      else if (*ptr == '?')
4114        {        {
4115        int i, set, unset, namelen;        int i, set, unset, namelen;
4116        int *optset;        int *optset;
# Line 3588  for (;; ptr++) Line 4131  for (;; ptr++)
4131    
4132    
4133          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4134            case '|':                 /* Reset capture count for each branch */
4135            reset_bracount = TRUE;
4136            /* Fall through */
4137    
4138            /* ------------------------------------------------------------ */
4139          case ':':                 /* Non-capturing bracket */          case ':':                 /* Non-capturing bracket */
4140          bravalue = OP_BRA;          bravalue = OP_BRA;
4141          ptr++;          ptr++;
# Line 3623  for (;; ptr++) Line 4171  for (;; ptr++)
4171    
4172          code[1+LINK_SIZE] = OP_CREF;          code[1+LINK_SIZE] = OP_CREF;
4173          skipbytes = 3;          skipbytes = 3;
4174          refsign = -1;          refsign = -1;
4175    
4176          /* Check for a test for recursion in a named group. */          /* Check for a test for recursion in a named group. */
4177    
# Line 3647  for (;; ptr++) Line 4195  for (;; ptr++)
4195            terminator = '\'';            terminator = '\'';
4196            ptr++;            ptr++;
4197            }            }
4198          else          else
4199            {            {
4200            terminator = 0;            terminator = 0;
4201            if (ptr[1] == '-' || ptr[1] == '+') refsign = *(++ptr);            if (ptr[1] == '-' || ptr[1] == '+') refsign = *(++ptr);
4202            }            }
4203    
4204          /* We now expect to read a name; any thing else is an error */          /* We now expect to read a name; any thing else is an error */
4205    
# Line 3689  for (;; ptr++) Line 4237  for (;; ptr++)
4237          /* In the real compile we do the work of looking for the actual          /* In the real compile we do the work of looking for the actual
4238          reference. If the string started with "+" or "-" we require the rest to          reference. If the string started with "+" or "-" we require the rest to
4239          be digits, in which case recno will be set. */          be digits, in which case recno will be set. */
4240    
4241          if (refsign > 0)          if (refsign > 0)
4242            {            {
4243            if (recno <= 0)            if (recno <= 0)
4244              {              {
4245              *errorcodeptr = ERR58;              *errorcodeptr = ERR58;
4246              goto FAILED;              goto FAILED;
4247              }              }
4248            if (refsign == '-')            if (refsign == '-')
4249              {              {
4250              recno = cd->bracount - recno + 1;              recno = cd->bracount - recno + 1;
4251              if (recno <= 0)              if (recno <= 0)
4252                {                {
4253                *errorcodeptr = ERR15;                *errorcodeptr = ERR15;
4254                goto FAILED;                goto FAILED;
4255                }                }
4256              }              }
4257            else recno += cd->bracount;            else recno += cd->bracount;
4258            PUT2(code, 2+LINK_SIZE, recno);            PUT2(code, 2+LINK_SIZE, recno);
4259            break;            break;
4260            }            }
4261    
4262          /* Otherwise (did not start with "+" or "-"), start by looking for the          /* Otherwise (did not start with "+" or "-"), start by looking for the
4263          name. */          name. */
4264    
4265          slot = cd->name_table;          slot = cd->name_table;
4266          for (i = 0; i < cd->names_found; i++)          for (i = 0; i < cd->names_found; i++)
4267            {            {
# Line 3804  for (;; ptr++) Line 4352  for (;; ptr++)
4352    
4353          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4354          case '!':                 /* Negative lookahead */          case '!':                 /* Negative lookahead */
         bravalue = OP_ASSERT_NOT;  
4355          ptr++;          ptr++;
4356            if (*ptr == ')')          /* Optimize (?!) */
4357              {
4358              *code++ = OP_FAIL;
4359              previous = NULL;
4360              continue;
4361              }
4362            bravalue = OP_ASSERT_NOT;
4363          break;          break;
4364    
4365    
# Line 4038  for (;; ptr++) Line 4592  for (;; ptr++)
4592            const uschar *called;            const uschar *called;
4593    
4594            if ((refsign = *ptr) == '+') ptr++;            if ((refsign = *ptr) == '+') ptr++;
4595            else if (refsign == '-')            else if (refsign == '-')
4596              {              {
4597              if ((digitab[ptr[1]] & ctype_digit) == 0)              if ((digitab[ptr[1]] & ctype_digit) == 0)
4598                goto OTHER_CHAR_AFTER_QUERY;                goto OTHER_CHAR_AFTER_QUERY;
4599              ptr++;              ptr++;
4600              }              }
4601    
4602            recno = 0;            recno = 0;
4603            while((digitab[*ptr] & ctype_digit) != 0)            while((digitab[*ptr] & ctype_digit) != 0)
4604              recno = recno * 10 + *ptr++ - '0';              recno = recno * 10 + *ptr++ - '0';
# Line 4054  for (;; ptr++) Line 4608  for (;; ptr++)
4608              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
4609              goto FAILED;              goto FAILED;
4610              }              }
4611    
4612            if (refsign == '-')            if (refsign == '-')
4613              {              {
4614              if (recno == 0)              if (recno == 0)
4615                {                {
4616                *errorcodeptr = ERR58;                *errorcodeptr = ERR58;
4617                goto FAILED;                goto FAILED;
4618                }                }
4619              recno = cd->bracount - recno + 1;              recno = cd->bracount - recno + 1;
4620              if (recno <= 0)              if (recno <= 0)
4621                {                {
4622                *errorcodeptr = ERR15;                *errorcodeptr = ERR15;
4623                goto FAILED;                goto FAILED;
4624                }                }
4625              }              }
4626            else if (refsign == '+')            else if (refsign == '+')
4627              {              {
# Line 4075  for (;; ptr++) Line 4629  for (;; ptr++)
4629                {                {
4630                *errorcodeptr = ERR58;                *errorcodeptr = ERR58;
4631                goto FAILED;                goto FAILED;
4632                }                }
4633              recno += cd->bracount;              recno += cd->bracount;
4634              }              }
4635    
4636            /* Come here from code above that handles a named recursion */            /* Come here from code above that handles a named recursion */
4637    
# Line 4151  for (;; ptr++) Line 4705  for (;; ptr++)
4705    
4706          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4707          default:              /* Other characters: check option setting */          default:              /* Other characters: check option setting */
4708          OTHER_CHAR_AFTER_QUERY:          OTHER_CHAR_AFTER_QUERY:
4709          set = unset = 0;          set = unset = 0;
4710          optset = &set;          optset = &set;
4711    
# Line 4163  for (;; ptr++) Line 4717  for (;; ptr++)
4717    
4718              case 'J':    /* Record that it changed in the external options */              case 'J':    /* Record that it changed in the external options */
4719              *optset |= PCRE_DUPNAMES;              *optset |= PCRE_DUPNAMES;
4720              cd->external_options |= PCRE_JCHANGED;              cd->external_flags |= PCRE_JCHANGED;
4721              break;              break;
4722    
4723              case 'i': *optset |= PCRE_CASELESS; break;              case 'i': *optset |= PCRE_CASELESS; break;
# Line 4286  for (;; ptr++) Line 4840  for (;; ptr++)
4840           errorcodeptr,                 /* Where to put an error message */           errorcodeptr,                 /* Where to put an error message */
4841           (bravalue == OP_ASSERTBACK ||           (bravalue == OP_ASSERTBACK ||
4842            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
4843             reset_bracount,               /* True if (?| group */
4844           skipbytes,                    /* Skip over bracket number */           skipbytes,                    /* Skip over bracket number */
4845           &subfirstbyte,                /* For possible first char */           &subfirstbyte,                /* For possible first char */
4846           &subreqbyte,                  /* For possible last char */           &subreqbyte,                  /* For possible last char */
# Line 4302  for (;; ptr++) Line 4857  for (;; ptr++)
4857      is on the bracket. */      is on the bracket. */
4858    
4859      /* If this is a conditional bracket, check that there are no more than      /* If this is a conditional bracket, check that there are no more than
4860      two branches in the group, or just one if it's a DEFINE group. */      two branches in the group, or just one if it's a DEFINE group. We do this
4861        in the real compile phase, not in the pre-pass, where the whole group may
4862        not be available. */
4863    
4864      if (bravalue == OP_COND)      if (bravalue == OP_COND && lengthptr == NULL)
4865        {        {
4866        uschar *tc = code;        uschar *tc = code;
4867        int condcount = 0;        int condcount = 0;
# Line 4351  for (;; ptr++) Line 4908  for (;; ptr++)
4908        goto FAILED;        goto FAILED;
4909        }        }
4910    
4911      /* In the pre-compile phase, update the length by the length of the nested      /* In the pre-compile phase, update the length by the length of the group,
4912      group, less the brackets at either end. Then reduce the compiled code to      less the brackets at either end. Then reduce the compiled code to just a
4913      just the brackets so that it doesn't use much memory if it is duplicated by      set of non-capturing brackets so that it doesn't use much memory if it is
4914      a quantifier. */      duplicated by a quantifier.*/
4915    
4916      if (lengthptr != NULL)      if (lengthptr != NULL)
4917        {        {
4918          if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE)
4919            {
4920            *errorcodeptr = ERR20;
4921            goto FAILED;
4922            }
4923        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;
4924        code++;        *code++ = OP_BRA;
4925        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
4926        *code++ = OP_KET;        *code++ = OP_KET;
4927        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
4928          break;    /* No need to waste time with special character handling */
4929        }        }
4930    
4931      /* Otherwise update the main code pointer to the end of the group. */      /* Otherwise update the main code pointer to the end of the group. */
4932    
4933      else code = tempcode;      code = tempcode;
4934    
4935      /* For a DEFINE group, required and first character settings are not      /* For a DEFINE group, required and first character settings are not
4936      relevant. */      relevant. */
# Line 4464  for (;; ptr++) Line 5027  for (;; ptr++)
5027        zerofirstbyte = firstbyte;        zerofirstbyte = firstbyte;
5028        zeroreqbyte = reqbyte;        zeroreqbyte = reqbyte;
5029    
5030        /* \k<name> or \k'name' is a back reference by name (Perl syntax) */        /* \k<name> or \k'name' is a back reference by name (Perl syntax).
5031          We also support \k{name} (.NET syntax) */
5032    
5033        if (-c == ESC_k && (ptr[1] == '<' || ptr[1] == '\''))        if (-c == ESC_k && (ptr[1] == '<' || ptr[1] == '\'' || ptr[1] == '{'))
5034          {          {
5035          is_recurse = FALSE;          is_recurse = FALSE;
5036          terminator = (*(++ptr) == '<')? '>' : '\'';          terminator = (*(++ptr) == '<')? '>' : (*ptr == '\'')? '\'' : '}';
5037          goto NAMED_REF_OR_RECURSE;          goto NAMED_REF_OR_RECURSE;
5038          }          }
5039    
# Line 4570  for (;; ptr++) Line 5134  for (;; ptr++)
5134      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;
5135      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
5136    
5137        /* Remember if \r or \n were seen */
5138    
5139        if (mcbuffer[0] == '\r' || mcbuffer[0] == '\n')
5140          cd->external_flags |= PCRE_HASCRORLF;
5141    
5142      /* Set the first and required bytes appropriately. If no previous first      /* Set the first and required bytes appropriately. If no previous first
5143      byte, set it from this character, but revert to none on a zero repeat.      byte, set it from this character, but revert to none on a zero repeat.
5144      Otherwise, leave the firstbyte value alone, and don't change it on a zero      Otherwise, leave the firstbyte value alone, and don't change it on a zero
# Line 4635  This function is used during the pre-com Line 5204  This function is used during the pre-com
5204  out the amount of memory needed, as well as during the real compile phase. The  out the amount of memory needed, as well as during the real compile phase. The
5205  value of lengthptr distinguishes the two phases.  value of lengthptr distinguishes the two phases.
5206    
5207  Argument:  Arguments:
5208    options        option bits, including any changes for this subpattern    options        option bits, including any changes for this subpattern
5209    oldims         previous settings of ims option bits    oldims         previous settings of ims option bits
5210    codeptr        -> the address of the current code pointer    codeptr        -> the address of the current code pointer
5211    ptrptr         -> the address of the current pattern pointer    ptrptr         -> the address of the current pattern pointer
5212    errorcodeptr   -> pointer to error code variable    errorcodeptr   -> pointer to error code variable
5213    lookbehind     TRUE if this is a lookbehind assertion    lookbehind     TRUE if this is a lookbehind assertion
5214      reset_bracount TRUE to reset the count for each branch
5215    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
5216    firstbyteptr   place to put the first required character, or a negative number    firstbyteptr   place to put the first required character, or a negative number
5217    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 4655  Returns:         TRUE on success Line 5225  Returns:         TRUE on success
5225    
5226  static BOOL  static BOOL
5227  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,
5228    int *errorcodeptr, BOOL lookbehind, int skipbytes, int *firstbyteptr,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
5229    int *reqbyteptr, branch_chain *bcptr, compile_data *cd, int *lengthptr)    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,
5230      int *lengthptr)
5231  {  {
5232  const uschar *ptr = *ptrptr;  const uschar *ptr = *ptrptr;
5233  uschar *code = *codeptr;  uschar *code = *codeptr;
# Line 4666  uschar *reverse_count = NULL; Line 5237  uschar *reverse_count = NULL;
5237  int firstbyte, reqbyte;  int firstbyte, reqbyte;
5238  int branchfirstbyte, branchreqbyte;  int branchfirstbyte, branchreqbyte;
5239  int length;  int length;
5240    int orig_bracount;
5241    int max_bracount;
5242  branch_chain bc;  branch_chain bc;
5243    
5244  bc.outer = bcptr;  bc.outer = bcptr;
# Line 4694  code += 1 + LINK_SIZE + skipbytes; Line 5267  code += 1 + LINK_SIZE + skipbytes;
5267    
5268  /* Loop for each alternative branch */  /* Loop for each alternative branch */
5269    
5270    orig_bracount = max_bracount = cd->bracount;
5271  for (;;)  for (;;)
5272    {    {
5273      /* For a (?| group, reset the capturing bracket count so that each branch
5274      uses the same numbers. */
5275    
5276      if (reset_bracount) cd->bracount = orig_bracount;
5277    
5278    /* Handle a change of ims options at the start of the branch */    /* Handle a change of ims options at the start of the branch */
5279    
5280    if ((options & PCRE_IMS) != oldims)    if ((options & PCRE_IMS) != oldims)
# Line 4725  for (;;) Line 5304  for (;;)
5304      return FALSE;      return FALSE;
5305      }      }
5306    
5307      /* Keep the highest bracket count in case (?| was used and some branch
5308      has fewer than the rest. */
5309    
5310      if (cd->bracount > max_bracount) max_bracount = cd->bracount;
5311    
5312    /* 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. */
5313    
5314    if (lengthptr == NULL)    if (lengthptr == NULL)
# Line 4788  for (;;) Line 5372  for (;;)
5372        }        }
5373      }      }
5374    
5375    /* Reached end of expression, either ')' or end of pattern. Go back through    /* Reached end of expression, either ')' or end of pattern. In the real
5376    the alternative branches and reverse the chain of offsets, with the field in    compile phase, go back through the alternative branches and reverse the chain
5377    the BRA item now becoming an offset to the first alternative. If there are    of offsets, with the field in the BRA item now becoming an offset to the
5378    no alternatives, it points to the end of the group. The length in the    first alternative. If there are no alternatives, it points to the end of the
5379    terminating ket is always the length of the whole bracketed item. If any of    group. The length in the terminating ket is always the length of the whole
5380    the ims options were changed inside the group, compile a resetting op-code    bracketed item. If any of the ims options were changed inside the group,
5381    following, except at the very end of the pattern. Return leaving the pointer    compile a resetting op-code following, except at the very end of the pattern.
5382    at the terminating char. */    Return leaving the pointer at the terminating char. */
5383    
5384    if (*ptr != '|')    if (*ptr != '|')
5385      {      {
5386      int branch_length = code - last_branch;      if (lengthptr == NULL)
     do  
5387        {        {
5388        int prev_length = GET(last_branch, 1);        int branch_length = code - last_branch;
5389        PUT(last_branch, 1, branch_length);        do
5390        branch_length = prev_length;          {
5391        last_branch -= branch_length;          int prev_length = GET(last_branch, 1);
5392            PUT(last_branch, 1, branch_length);
5393            branch_length = prev_length;
5394            last_branch -= branch_length;
5395            }
5396          while (branch_length > 0);
5397        }        }
     while (branch_length > 0);  
5398    
5399      /* Fill in the ket */      /* Fill in the ket */
5400    
# Line 4824  for (;;) Line 5411  for (;;)
5411        length += 2;        length += 2;
5412        }        }
5413    
5414        /* Retain the highest bracket number, in case resetting was used. */
5415    
5416        cd->bracount = max_bracount;
5417    
5418      /* Set values to pass back */      /* Set values to pass back */
5419    
5420      *codeptr = code;      *codeptr = code;
5421      *ptrptr = ptr;      *ptrptr = ptr;
5422      *firstbyteptr = firstbyte;      *firstbyteptr = firstbyte;
5423      *reqbyteptr = reqbyte;      *reqbyteptr = reqbyte;
5424      if (lengthptr != NULL) *lengthptr += length;      if (lengthptr != NULL)
5425          {
5426          if (OFLOW_MAX - *lengthptr < length)
5427            {
5428            *errorcodeptr = ERR20;
5429            return FALSE;
5430            }
5431          *lengthptr += length;
5432          }
5433      return TRUE;      return TRUE;
5434      }      }
5435    
5436    /* Another branch follows; insert an "or" node. Its length field points back    /* Another branch follows. In the pre-compile phase, we can move the code
5437      pointer back to where it was for the start of the first branch. (That is,
5438      pretend that each branch is the only one.)
5439    
5440      In the real compile phase, insert an ALT node. Its length field points back
5441    to the previous branch while the bracket remains open. At the end the chain    to the previous branch while the bracket remains open. At the end the chain
5442    is reversed. It's done like this so that the start of the bracket has a    is reversed. It's done like this so that the start of the bracket has a
5443    zero offset until it is closed, making it possible to detect recursion. */    zero offset until it is closed, making it possible to detect recursion. */
5444    
5445    *code = OP_ALT;    if (lengthptr != NULL)
5446    PUT(code, 1, code - last_branch);      {
5447    bc.current = last_branch = code;      code = *codeptr + 1 + LINK_SIZE + skipbytes;
5448    code += 1 + LINK_SIZE;      length += 1 + LINK_SIZE;
5449        }
5450      else
5451        {
5452        *code = OP_ALT;
5453        PUT(code, 1, code - last_branch);
5454        bc.current = last_branch = code;
5455        code += 1 + LINK_SIZE;
5456        }
5457    
5458    ptr++;    ptr++;
   length += 1 + LINK_SIZE;  
5459    }    }
5460  /* Control never reaches here */  /* Control never reaches here */
5461  }  }
# Line 5127  real_pcre *re; Line 5738  real_pcre *re;
5738  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
5739  int firstbyte, reqbyte, newline;  int firstbyte, reqbyte, newline;
5740  int errorcode = 0;  int errorcode = 0;
5741    int skipatstart = 0;
5742  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
5743  BOOL utf8;  BOOL utf8;
5744  #endif  #endif
# Line 5205  cd->fcc = tables + fcc_offset; Line 5817  cd->fcc = tables + fcc_offset;
5817  cd->cbits = tables + cbits_offset;  cd->cbits = tables + cbits_offset;
5818  cd->ctypes = tables + ctypes_offset;  cd->ctypes = tables + ctypes_offset;
5819    
5820    /* Check for global one-time settings at the start of the pattern, and remember
5821    the offset for later. */
5822    
5823    while (ptr[skipatstart] == '(' && ptr[skipatstart+1] == '*')
5824      {
5825      int newnl = 0;
5826      int newbsr = 0;
5827    
5828      if (strncmp((char *)(ptr+skipatstart+2), "CR)", 3) == 0)
5829        { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }
5830      else if (strncmp((char *)(ptr+skipatstart+2), "LF)", 3)  == 0)
5831        { skipatstart += 5; newnl = PCRE_NEWLINE_LF; }
5832      else if (strncmp((char *)(ptr+skipatstart+2), "CRLF)", 5)  == 0)
5833        { skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; }
5834      else if (strncmp((char *)(ptr+skipatstart+2), "ANY)", 4) == 0)
5835        { skipatstart += 6; newnl = PCRE_NEWLINE_ANY; }
5836      else if (strncmp((char *)(ptr+skipatstart+2), "ANYCRLF)", 8)  == 0)
5837        { skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; }
5838    
5839      else if (strncmp((char *)(ptr+skipatstart+2), "BSR_ANYCRLF)", 12) == 0)
5840        { skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; }
5841      else if (strncmp((char *)(ptr+skipatstart+2), "BSR_UNICODE)", 12) == 0)
5842        { skipatstart += 14; newbsr = PCRE_BSR_UNICODE; }
5843    
5844      if (newnl != 0)
5845        options = (options & ~PCRE_NEWLINE_BITS) | newnl;
5846      else if (newbsr != 0)
5847        options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr;
5848      else break;
5849      }
5850    
5851    /* Check validity of \R options. */
5852    
5853    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
5854      {
5855      case 0:
5856      case PCRE_BSR_ANYCRLF:
5857      case PCRE_BSR_UNICODE:
5858      break;
5859      default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN;
5860      }
5861    
5862  /* Handle different types of newline. The three bits give seven cases. The  /* Handle different types of newline. The three bits give seven cases. The
5863  current code allows for fixed one- or two-byte sequences, plus "any" and  current code allows for fixed one- or two-byte sequences, plus "any" and
5864  "anycrlf". */  "anycrlf". */
5865    
5866  switch (options & (PCRE_NEWLINE_CRLF | PCRE_NEWLINE_ANY))  switch (options & PCRE_NEWLINE_BITS)
5867    {    {
5868    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Build-time default */
5869    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
5870    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = '\n'; break;
5871    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
# Line 5274  cd->hwm = cworkspace; Line 5928  cd->hwm = cworkspace;
5928  cd->start_pattern = (const uschar *)pattern;  cd->start_pattern = (const uschar *)pattern;
5929  cd->end_pattern = (const uschar *)(pattern + strlen(pattern));  cd->end_pattern = (const uschar *)(pattern + strlen(pattern));
5930  cd->req_varyopt = 0;  cd->req_varyopt = 0;
 cd->nopartial = FALSE;  
5931  cd->external_options = options;  cd->external_options = options;
5932    cd->external_flags = 0;
5933    
5934  /* Now do the pre-compile. On error, errorcode will be set non-zero, so we  /* Now do the pre-compile. On error, errorcode will be set non-zero, so we
5935  don't need to look at the result of the function here. The initial options have  don't need to look at the result of the function here. The initial options have
# Line 5283  been put into the cd block so that they Line 5937  been put into the cd block so that they
5937  found within the regex right at the beginning. Bringing initial option settings  found within the regex right at the beginning. Bringing initial option settings
5938  outside can help speed up starting point checks. */  outside can help speed up starting point checks. */
5939    
5940    ptr += skipatstart;
5941  code = cworkspace;  code = cworkspace;
5942  *code = OP_BRA;  *code = OP_BRA;
5943  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,
5944    &code, &ptr, &errorcode, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, &length);    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,
5945      &length);
5946  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
5947    
5948  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 5312  if (re == NULL) Line 5968  if (re == NULL)
5968    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
5969    }    }
5970    
5971  /* Put in the magic number, and save the sizes, initial options, and character  /* Put in the magic number, and save the sizes, initial options, internal
5972  table pointer. NULL is used for the default character tables. The nullpad field  flags, and character table pointer. NULL is used for the default character
5973  is at the end; it's there to help in the case when a regex compiled on a system  tables. The nullpad field is at the end; it's there to help in the case when a
5974  with 4-byte pointers is run on another with 8-byte pointers. */  regex compiled on a system with 4-byte pointers is run on another with 8-byte
5975    pointers. */
5976    
5977  re->magic_number = MAGIC_NUMBER;  re->magic_number = MAGIC_NUMBER;
5978  re->size = size;  re->size = size;
5979  re->options = cd->external_options;  re->options = cd->external_options;
5980    re->flags = cd->external_flags;
5981  re->dummy1 = 0;  re->dummy1 = 0;
5982  re->first_byte = 0;  re->first_byte = 0;
5983  re->req_byte = 0;  re->req_byte = 0;
# Line 5344  codestart = cd->name_table + re->name_en Line 6002  codestart = cd->name_table + re->name_en
6002  cd->start_code = codestart;  cd->start_code = codestart;
6003  cd->hwm = cworkspace;  cd->hwm = cworkspace;
6004  cd->req_varyopt = 0;  cd->req_varyopt = 0;
6005  cd->nopartial = FALSE;  cd->had_accept = FALSE;
6006    
6007  /* Set up a starting, non-extracting bracket, then compile the expression. On  /* Set up a starting, non-extracting bracket, then compile the expression. On
6008  error, errorcode will be set non-zero, so we don't need to look at the result  error, errorcode will be set non-zero, so we don't need to look at the result
6009  of the function here. */  of the function here. */
6010    
6011  ptr = (const uschar *)pattern;  ptr = (const uschar *)pattern + skipatstart;
6012  code = (uschar *)codestart;  code = (uschar *)codestart;
6013  *code = OP_BRA;  *code = OP_BRA;
6014  (void)compile_regex(re->options, re->options & PCRE_IMS, &code, &ptr,  (void)compile_regex(re->options, re->options & PCRE_IMS, &code, &ptr,
6015    &errorcode, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL);    &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL);
6016  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
6017  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
6018    re->flags = cd->external_flags;
6019    
6020  if (cd->nopartial) re->options |= PCRE_NOPARTIAL;  if (cd->had_accept) reqbyte = -1;   /* Must disable after (*ACCEPT) */
6021    
6022  /* If not reached end of pattern on success, there's an excess bracket. */  /* If not reached end of pattern on success, there's an excess bracket. */
6023    
# Line 5400  if (errorcode != 0) Line 6059  if (errorcode != 0)
6059    PCRE_EARLY_ERROR_RETURN:    PCRE_EARLY_ERROR_RETURN:
6060    *erroroffset = ptr - (const uschar *)pattern;    *erroroffset = ptr - (const uschar *)pattern;
6061    PCRE_EARLY_ERROR_RETURN2:    PCRE_EARLY_ERROR_RETURN2:
6062    *errorptr = error_texts[errorcode];    *errorptr = find_error_text(errorcode);
6063    if (errorcodeptr != NULL) *errorcodeptr = errorcode;    if (errorcodeptr != NULL) *errorcodeptr = errorcode;
6064    return NULL;    return NULL;
6065    }    }
# Line 5429  if ((re->options & PCRE_ANCHORED) == 0) Line 6088  if ((re->options & PCRE_ANCHORED) == 0)
6088        int ch = firstbyte & 255;        int ch = firstbyte & 255;
6089        re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&        re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&
6090           cd->fcc[ch] == ch)? ch : firstbyte;           cd->fcc[ch] == ch)? ch : firstbyte;
6091        re->options |= PCRE_FIRSTSET;        re->flags |= PCRE_FIRSTSET;
6092        }        }
6093      else if (is_startline(codestart, 0, cd->backref_map))      else if (is_startline(codestart, 0, cd->backref_map))
6094        re->options |= PCRE_STARTLINE;        re->flags |= PCRE_STARTLINE;
6095      }      }
6096    }    }
6097    
# Line 5446  if (reqbyte >= 0 && Line 6105  if (reqbyte >= 0 &&
6105    int ch = reqbyte & 255;    int ch = reqbyte & 255;
6106    re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&    re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&
6107      cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;      cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;
6108    re->options |= PCRE_REQCHSET;    re->flags |= PCRE_REQCHSET;
6109    }    }
6110    
6111  /* Print out the compiled data if debugging is enabled. This is never the  /* Print out the compiled data if debugging is enabled. This is never the
# Line 5457  case when building a production library. Line 6116  case when building a production library.
6116  printf("Length = %d top_bracket = %d top_backref = %d\n",  printf("Length = %d top_bracket = %d top_backref = %d\n",
6117    length, re->top_bracket, re->top_backref);    length, re->top_bracket, re->top_backref);
6118    
6119  if (re->options != 0)  printf("Options=%08x\n", re->options);
   {  
   printf("%s%s%s%s%s%s%s%s%s\n",  
     ((re->options & PCRE_NOPARTIAL) != 0)? "nopartial " : "",  
     ((re->options & PCRE_ANCHORED) != 0)? "anchored " : "",  
     ((re->options & PCRE_CASELESS) != 0)? "caseless " : "",  
     ((re->options & PCRE_EXTENDED) != 0)? "extended " : "",  
     ((re->options & PCRE_MULTILINE) != 0)? "multiline " : "",  
     ((re->options & PCRE_DOTALL) != 0)? "dotall " : "",  
     ((re->options & PCRE_DOLLAR_ENDONLY) != 0)? "endonly " : "",  
     ((re->options & PCRE_EXTRA) != 0)? "extra " : "",  
     ((re->options & PCRE_UNGREEDY) != 0)? "ungreedy " : "");  
   }  
6120    
6121  if ((re->options & PCRE_FIRSTSET) != 0)  if ((re->flags & PCRE_FIRSTSET) != 0)
6122    {    {
6123    int ch = re->first_byte & 255;    int ch = re->first_byte & 255;
6124    const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?    const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?
# Line 5480  if ((re->options & PCRE_FIRSTSET) != 0) Line 6127  if ((re->options & PCRE_FIRSTSET) != 0)
6127      else printf("First char = \\x%02x%s\n", ch, caseless);      else printf("First char = \\x%02x%s\n", ch, caseless);
6128    }    }
6129    
6130  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6131    {    {
6132    int ch = re->req_byte & 255;    int ch = re->req_byte & 255;
6133    const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?    const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?
# Line 5497  was compiled can be seen. */ Line 6144  was compiled can be seen. */
6144  if (code - codestart > length)  if (code - codestart > length)
6145    {    {
6146    (pcre_free)(re);    (pcre_free)(re);
6147    *errorptr = error_texts[ERR23];    *errorptr = find_error_text(ERR23);
6148    *erroroffset = ptr - (uschar *)pattern;    *erroroffset = ptr - (uschar *)pattern;
6149    if (errorcodeptr != NULL) *errorcodeptr = ERR23;    if (errorcodeptr != NULL) *errorcodeptr = ERR23;
6150    return NULL;    return NULL;

Legend:
Removed from v.167  
changed lines
  Added in v.264

  ViewVC Help
Powered by ViewVC 1.1.5