/[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 176 by ph10, Mon Jun 11 13:48:37 2007 UTC revision 265 by ph10, Wed Nov 14 11:35:48 2007 UTC
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  supporting internal functions that are not used by other modules. */  supporting internal functions that are not used by other modules. */
43    
44    
45    #ifdef HAVE_CONFIG_H
46    #include "config.h"
47    #endif
48    
49  #define NLBLOCK cd             /* Block containing newline information */  #define NLBLOCK cd             /* Block containing newline information */
50  #define PSSTART start_pattern  /* Field containing processed string start */  #define PSSTART start_pattern  /* Field containing processed string start */
51  #define PSEND   end_pattern    /* Field containing processed string end */  #define PSEND   end_pattern    /* Field containing processed string end */
52    
   
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    
# Line 58  used by pcretest. DEBUG is not defined w Line 61  used by pcretest. DEBUG is not defined w
61  #endif  #endif
62    
63    
64    /* Macro for setting individual bits in class bitmaps. */
65    
66    #define SETBIT(a,b) a[b/8] |= (1 << (b%8))
67    
68    /* Maximum length value to check against when making sure that the integer that
69    holds the compiled pattern length does not overflow. We make it a bit less than
70    INT_MAX to allow for adding in group terminating bytes, so that we don't have
71    to check them every time. */
72    
73    #define OFLOW_MAX (INT_MAX - 20)
74    
75    
76  /*************************************************  /*************************************************
77  *      Code parameters and static tables         *  *      Code parameters and static tables         *
78  *************************************************/  *************************************************/
# Line 87  static const short int escapes[] = { Line 102  static const short int escapes[] = {
102       0,      0,      0,      0,      0,      0,      0,      0,   /* 0 - 7 */       0,      0,      0,      0,      0,      0,      0,      0,   /* 0 - 7 */
103       0,      0,    ':',    ';',    '<',    '=',    '>',    '?',   /* 8 - ? */       0,      0,    ':',    ';',    '<',    '=',    '>',    '?',   /* 8 - ? */
104     '@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E,      0, -ESC_G,   /* @ - G */     '@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E,      0, -ESC_G,   /* @ - G */
105       0,      0,      0, -ESC_K,      0,      0,      0,      0,   /* H - O */  -ESC_H,      0,      0, -ESC_K,      0,      0,      0,      0,   /* H - O */
106  -ESC_P, -ESC_Q, -ESC_R, -ESC_S,      0,      0,      0, -ESC_W,   /* P - W */  -ESC_P, -ESC_Q, -ESC_R, -ESC_S,      0,      0, -ESC_V, -ESC_W,   /* P - W */
107  -ESC_X,      0, -ESC_Z,    '[',   '\\',    ']',    '^',    '_',   /* X - _ */  -ESC_X,      0, -ESC_Z,    '[',   '\\',    ']',    '^',    '_',   /* X - _ */
108     '`',      7, -ESC_b,      0, -ESC_d,  ESC_e,  ESC_f,      0,   /* ` - g */     '`',      7, -ESC_b,      0, -ESC_d,  ESC_e,  ESC_f,      0,   /* ` - g */
109       0,      0,      0, -ESC_k,      0,      0,  ESC_n,      0,   /* h - o */  -ESC_h,      0,      0, -ESC_k,      0,      0,  ESC_n,      0,   /* h - o */
110  -ESC_p,      0,  ESC_r, -ESC_s,  ESC_tee,    0,      0, -ESC_w,   /* p - w */  -ESC_p,      0,  ESC_r, -ESC_s,  ESC_tee,    0, -ESC_v, -ESC_w,   /* p - w */
111       0,      0, -ESC_z                                            /* x - z */       0,      0, -ESC_z                                            /* x - z */
112  };  };
113    
# Line 106  static const short int escapes[] = { Line 121  static const short int escapes[] = {
121  /*  70 */     0,     0,      0,       0,      0,     0,      0,      0,  /*  70 */     0,     0,      0,       0,      0,     0,      0,      0,
122  /*  78 */     0,   '`',    ':',     '#',    '@',  '\'',    '=',    '"',  /*  78 */     0,   '`',    ':',     '#',    '@',  '\'',    '=',    '"',
123  /*  80 */     0,     7, -ESC_b,       0, -ESC_d, ESC_e,  ESC_f,      0,  /*  80 */     0,     7, -ESC_b,       0, -ESC_d, ESC_e,  ESC_f,      0,
124  /*  88 */     0,     0,      0,     '{',      0,     0,      0,      0,  /*  88 */-ESC_h,     0,      0,     '{',      0,     0,      0,      0,
125  /*  90 */     0,     0, -ESC_k,     'l',      0, ESC_n,      0, -ESC_p,  /*  90 */     0,     0, -ESC_k,     'l',      0, ESC_n,      0, -ESC_p,
126  /*  98 */     0, ESC_r,      0,     '}',      0,     0,      0,      0,  /*  98 */     0, ESC_r,      0,     '}',      0,     0,      0,      0,
127  /*  A0 */     0,   '~', -ESC_s, ESC_tee,      0,     0, -ESC_w,      0,  /*  A0 */     0,   '~', -ESC_s, ESC_tee,      0,-ESC_v, -ESC_w,      0,
128  /*  A8 */     0,-ESC_z,      0,       0,      0,   '[',      0,      0,  /*  A8 */     0,-ESC_z,      0,       0,      0,   '[',      0,      0,
129  /*  B0 */     0,     0,      0,       0,      0,     0,      0,      0,  /*  B0 */     0,     0,      0,       0,      0,     0,      0,      0,
130  /*  B8 */     0,     0,      0,       0,      0,   ']',    '=',    '-',  /*  B8 */     0,     0,      0,       0,      0,   ']',    '=',    '-',
131  /*  C0 */   '{',-ESC_A, -ESC_B,  -ESC_C, -ESC_D,-ESC_E,      0, -ESC_G,  /*  C0 */   '{',-ESC_A, -ESC_B,  -ESC_C, -ESC_D,-ESC_E,      0, -ESC_G,
132  /*  C8 */     0,     0,      0,       0,      0,     0,      0,      0,  /*  C8 */-ESC_H,     0,      0,       0,      0,     0,      0,      0,
133  /*  D0 */   '}',     0,      0,       0,      0,     0,      0, -ESC_P,  /*  D0 */   '}',     0, -ESC_K,       0,      0,     0,      0, -ESC_P,
134  /*  D8 */-ESC_Q,-ESC_R,      0,       0,      0,     0,      0,      0,  /*  D8 */-ESC_Q,-ESC_R,      0,       0,      0,     0,      0,      0,
135  /*  E0 */  '\\',     0, -ESC_S,       0,      0,     0, -ESC_W, -ESC_X,  /*  E0 */  '\\',     0, -ESC_S,       0,      0,-ESC_V, -ESC_W, -ESC_X,
136  /*  E8 */     0,-ESC_Z,      0,       0,      0,     0,      0,      0,  /*  E8 */     0,-ESC_Z,      0,       0,      0,     0,      0,      0,
137  /*  F0 */     0,     0,      0,       0,      0,     0,      0,      0,  /*  F0 */     0,     0,      0,       0,      0,     0,      0,      0,
138  /*  F8 */     0,     0,      0,       0,      0,     0,      0,      0  /*  F8 */     0,     0,      0,       0,      0,     0,      0,      0
# Line 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 a braced name or an optionally braced non-zero number",    "\\g is not followed by a braced name or an optionally braced non-zero number\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, BOOL, int,    compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,
434      int *, 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 485  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 496  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 525  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 620  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 696  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 725  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 899  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 926  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 1054  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 1143  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 1251  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 1315  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 1394  for (code = first_significant_code(code Line 1541  for (code = first_significant_code(code
1541    
1542    /* For other groups, scan the branches. */    /* For other groups, scan the branches. */
1543    
1544    if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE)    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 1418  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 1484  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 1600  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 1646  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 1944  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 1978  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 1992  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 2107  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;    BOOL reset_bracount;
2391    int class_charcount;    int class_charcount;
2392    int class_lastchar;    int class_lastchar;
2393    int newoptions;    int newoptions;
# Line 2148  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 2260  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 2322  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 2406  for (;; ptr++) Line 2710  for (;; ptr++)
2710          if (*ptr == '^')          if (*ptr == '^')
2711            {            {
2712            local_negate = TRUE;            local_negate = TRUE;
2713              should_flip_negation = TRUE;  /* Note negative special */
2714            ptr++;            ptr++;
2715            }            }
2716    
# Line 2472  for (;; ptr++) Line 2777  for (;; ptr++)
2777        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
2778        case. Inside a class (and only there) it is treated as backspace.        case. Inside a class (and only there) it is treated as backspace.
2779        Elsewhere it marks a word boundary. Other escapes have preset maps ready        Elsewhere it marks a word boundary. Other escapes have preset maps ready
2780        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
2781        character in them, so set class_charcount bigger than one. */        character in them, so set class_charcount bigger than one. */
2782    
2783        if (c == '\\')        if (c == '\\')
# Line 2492  for (;; ptr++) Line 2797  for (;; ptr++)
2797            else inescq = TRUE;            else inescq = TRUE;
2798            continue;            continue;
2799            }            }
2800            else if (-c == ESC_E) continue;  /* Ignore orphan \E */
2801    
2802          if (c < 0)          if (c < 0)
2803            {            {
# Line 2507  for (;; ptr++) Line 2813  for (;; ptr++)
2813              continue;              continue;
2814    
2815              case ESC_D:              case ESC_D:
2816                should_flip_negation = TRUE;
2817              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit];
2818              continue;              continue;
2819    
# Line 2515  for (;; ptr++) Line 2822  for (;; ptr++)
2822              continue;              continue;
2823    
2824              case ESC_W:              case ESC_W:
2825                should_flip_negation = TRUE;
2826              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];
2827              continue;              continue;
2828    
# Line 2524  for (;; ptr++) Line 2832  for (;; ptr++)
2832              continue;              continue;
2833    
2834              case ESC_S:              case ESC_S:
2835                should_flip_negation = TRUE;
2836              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];
2837              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */
2838              continue;              continue;
# Line 2540  for (;; ptr++) Line 2849  for (;; ptr++)
2849            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||            else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||
2850                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;                     c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;
2851    
2852              /* We need to deal with \H, \h, \V, and \v in both phases because
2853              they use extra memory. */
2854    
2855              if (-c == ESC_h)
2856                {
2857                SETBIT(classbits, 0x09); /* VT */
2858                SETBIT(classbits, 0x20); /* SPACE */
2859                SETBIT(classbits, 0xa0); /* NSBP */
2860    #ifdef SUPPORT_UTF8
2861                if (utf8)
2862                  {
2863                  class_utf8 = TRUE;
2864                  *class_utf8data++ = XCL_SINGLE;
2865                  class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data);
2866                  *class_utf8data++ = XCL_SINGLE;
2867                  class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data);
2868                  *class_utf8data++ = XCL_RANGE;
2869                  class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data);
2870                  class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data);
2871                  *class_utf8data++ = XCL_SINGLE;
2872                  class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data);
2873                  *class_utf8data++ = XCL_SINGLE;
2874                  class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data);
2875                  *class_utf8data++ = XCL_SINGLE;
2876                  class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data);
2877                  }
2878    #endif
2879                continue;
2880                }
2881    
2882              if (-c == ESC_H)
2883                {
2884                for (c = 0; c < 32; c++)
2885                  {
2886                  int x = 0xff;
2887                  switch (c)
2888                    {
2889                    case 0x09/8: x ^= 1 << (0x09%8); break;
2890                    case 0x20/8: x ^= 1 << (0x20%8); break;
2891                    case 0xa0/8: x ^= 1 << (0xa0%8); break;
2892                    default: break;
2893                    }
2894                  classbits[c] |= x;
2895                  }
2896    
2897    #ifdef SUPPORT_UTF8
2898                if (utf8)
2899                  {
2900                  class_utf8 = TRUE;
2901                  *class_utf8data++ = XCL_RANGE;
2902                  class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2903                  class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data);
2904                  *class_utf8data++ = XCL_RANGE;
2905                  class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data);
2906                  class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data);
2907                  *class_utf8data++ = XCL_RANGE;
2908                  class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data);
2909                  class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data);
2910                  *class_utf8data++ = XCL_RANGE;
2911                  class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data);
2912                  class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data);
2913                  *class_utf8data++ = XCL_RANGE;
2914                  class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data);
2915                  class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data);
2916                  *class_utf8data++ = XCL_RANGE;
2917                  class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data);
2918                  class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data);
2919                  *class_utf8data++ = XCL_RANGE;
2920                  class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data);
2921                  class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2922                  }
2923    #endif
2924                continue;
2925                }
2926    
2927              if (-c == ESC_v)
2928                {
2929                SETBIT(classbits, 0x0a); /* LF */
2930                SETBIT(classbits, 0x0b); /* VT */
2931                SETBIT(classbits, 0x0c); /* FF */
2932                SETBIT(classbits, 0x0d); /* CR */
2933                SETBIT(classbits, 0x85); /* NEL */
2934    #ifdef SUPPORT_UTF8
2935                if (utf8)
2936                  {
2937                  class_utf8 = TRUE;
2938                  *class_utf8data++ = XCL_RANGE;
2939                  class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data);
2940                  class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2941                  }
2942    #endif
2943                continue;
2944                }
2945    
2946              if (-c == ESC_V)
2947                {
2948                for (c = 0; c < 32; c++)
2949                  {
2950                  int x = 0xff;
2951                  switch (c)
2952                    {
2953                    case 0x0a/8: x ^= 1 << (0x0a%8);
2954                                 x ^= 1 << (0x0b%8);
2955                                 x ^= 1 << (0x0c%8);
2956                                 x ^= 1 << (0x0d%8);
2957                                 break;
2958                    case 0x85/8: x ^= 1 << (0x85%8); break;
2959                    default: break;
2960                    }
2961                  classbits[c] |= x;
2962                  }
2963    
2964    #ifdef SUPPORT_UTF8
2965                if (utf8)
2966                  {
2967                  class_utf8 = TRUE;
2968                  *class_utf8data++ = XCL_RANGE;
2969                  class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data);
2970                  class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data);
2971                  *class_utf8data++ = XCL_RANGE;
2972                  class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data);
2973                  class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data);
2974                  }
2975    #endif
2976                continue;
2977                }
2978    
2979            /* We need to deal with \P and \p in both phases. */            /* We need to deal with \P and \p in both phases. */
2980    
2981  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2591  for (;; ptr++) Line 3027  for (;; ptr++)
3027    
3028        oldptr = ptr;        oldptr = ptr;
3029    
3030          /* Remember \r or \n */
3031    
3032          if (c == '\r' || c == '\n') cd->external_flags |= PCRE_HASCRORLF;
3033    
3034          /* Check for range */
3035    
3036        if (!inescq && ptr[1] == '-')        if (!inescq && ptr[1] == '-')
3037          {          {
3038          int d;          int d;
# Line 2658  for (;; ptr++) Line 3100  for (;; ptr++)
3100    
3101          if (d == c) goto LONE_SINGLE_CHARACTER;  /* A few lines below */          if (d == c) goto LONE_SINGLE_CHARACTER;  /* A few lines below */
3102    
3103            /* Remember \r or \n */
3104    
3105            if (d == '\r' || d == '\n') cd->external_flags |= PCRE_HASCRORLF;
3106    
3107          /* 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
3108          matching, we have to use an XCLASS with extra data items. Caseless          matching, we have to use an XCLASS with extra data items. Caseless
3109          matching for characters > 127 is available only if UCP support is          matching for characters > 127 is available only if UCP support is
# Line 2680  for (;; ptr++) Line 3126  for (;; ptr++)
3126              unsigned int origd = d;              unsigned int origd = d;
3127              while (get_othercase_range(&cc, origd, &occ, &ocd))              while (get_othercase_range(&cc, origd, &occ, &ocd))
3128                {                {
3129                if (occ >= (unsigned int)c &&                if (occ >= (unsigned int)c &&
3130                    ocd <= (unsigned int)d)                    ocd <= (unsigned int)d)
3131                  continue;                          /* Skip embedded ranges */                  continue;                          /* Skip embedded ranges */
3132    
3133                if (occ < (unsigned int)c  &&                if (occ < (unsigned int)c  &&
3134                    ocd >= (unsigned int)c - 1)      /* Extend the basic range */                    ocd >= (unsigned int)c - 1)      /* Extend the basic range */
3135                  {                                  /* if there is overlap,   */                  {                                  /* if there is overlap,   */
3136                  c = occ;                           /* noting that if occ < c */                  c = occ;                           /* noting that if occ < c */
3137                  continue;                          /* we can't have ocd > d  */                  continue;                          /* we can't have ocd > d  */
3138                  }                                  /* because a subrange is  */                  }                                  /* because a subrange is  */
3139                if (ocd > (unsigned int)d &&                if (ocd > (unsigned int)d &&
3140                    occ <= (unsigned int)d + 1)      /* always shorter than    */                    occ <= (unsigned int)d + 1)      /* always shorter than    */
3141                  {                                  /* the basic range.       */                  {                                  /* the basic range.       */
3142                  d = ocd;                  d = ocd;
# Line 2811  for (;; ptr++) Line 3257  for (;; ptr++)
3257        goto FAILED;        goto FAILED;
3258        }        }
3259    
3260    
3261    /* This code has been disabled because it would mean that \s counts as
3262    an explicit \r or \n reference, and that's not really what is wanted. Now
3263    we set the flag only if there is a literal "\r" or "\n" in the class. */
3264    
3265    #if 0
3266        /* Remember whether \r or \n are in this class */
3267    
3268        if (negate_class)
3269          {
3270          if ((classbits[1] & 0x24) != 0x24) cd->external_flags |= PCRE_HASCRORLF;
3271          }
3272        else
3273          {
3274          if ((classbits[1] & 0x24) != 0) cd->external_flags |= PCRE_HASCRORLF;
3275          }
3276    #endif
3277    
3278    
3279      /* 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
3280      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
3281      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
3282      because OP_NOT and the related opcodes like OP_NOTSTAR operate on      optimize.
3283      single-bytes only. This is an historical hangover. Maybe one day we can  
3284      tidy these opcodes to handle multi-byte characters.      In UTF-8 mode, we can optimize the negative case only if there were no
3285        characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR
3286        operate on single-bytes only. This is an historical hangover. Maybe one day
3287        we can tidy these opcodes to handle multi-byte characters.
3288    
3289      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
3290      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 2826  for (;; ptr++) Line 3294  for (;; ptr++)
3294      reqbyte, save the previous value for reinstating. */      reqbyte, save the previous value for reinstating. */
3295    
3296  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3297      if (class_charcount == 1 &&      if (class_charcount == 1 && !class_utf8 &&
3298            (!utf8 ||        (!utf8 || !negate_class || class_lastchar < 128))
           (!class_utf8 && (!negate_class || class_lastchar < 128))))  
   
3299  #else  #else
3300      if (class_charcount == 1)      if (class_charcount == 1)
3301  #endif  #endif
# Line 2872  for (;; ptr++) Line 3338  for (;; ptr++)
3338      zeroreqbyte = reqbyte;      zeroreqbyte = reqbyte;
3339    
3340      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
3341      extended class, with its own opcode. If there are no characters < 256,      extended class, with its own opcode, unless there was a negated special
3342      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
3343        the class, so any that were explicitly given as well can be ignored. If
3344        (when there are explicit characters > 255 that must be listed) there are no
3345        characters < 256, we can omit the bitmap in the actual compiled code. */
3346    
3347  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3348      if (class_utf8)      if (class_utf8 && !should_flip_negation)
3349        {        {
3350        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */
3351        *code++ = OP_XCLASS;        *code++ = OP_XCLASS;
# Line 2902  for (;; ptr++) Line 3371  for (;; ptr++)
3371        }        }
3372  #endif  #endif
3373    
3374      /* 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
3375      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
3376      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
3377      setting must remain unchanged after any kind of repeat. */      map into the code vector, negating it if necessary. */
3378    
3379        *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;
3380      if (negate_class)      if (negate_class)
3381        {        {
       *code++ = OP_NCLASS;  
3382        if (lengthptr == NULL)    /* Save time in the pre-compile phase */        if (lengthptr == NULL)    /* Save time in the pre-compile phase */
3383          for (c = 0; c < 32; c++) code[c] = ~classbits[c];          for (c = 0; c < 32; c++) code[c] = ~classbits[c];
3384        }        }
3385      else      else
3386        {        {
       *code++ = OP_CLASS;  
3387        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
3388        }        }
3389      code += 32;      code += 32;
# Line 3101  for (;; ptr++) Line 3569  for (;; ptr++)
3569        /* All real repeats make it impossible to handle partial matching (maybe        /* All real repeats make it impossible to handle partial matching (maybe
3570        one day we will be able to remove this restriction). */        one day we will be able to remove this restriction). */
3571    
3572        if (repeat_max != 1) cd->nopartial = TRUE;        if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL;
3573    
3574        /* Combine the op_type with the repeat_type */        /* Combine the op_type with the repeat_type */
3575    
# Line 3251  for (;; ptr++) Line 3719  for (;; ptr++)
3719        /* All real repeats make it impossible to handle partial matching (maybe        /* All real repeats make it impossible to handle partial matching (maybe
3720        one day we will be able to remove this restriction). */        one day we will be able to remove this restriction). */
3721    
3722        if (repeat_max != 1) cd->nopartial = TRUE;        if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL;
3723    
3724        if (repeat_min == 0 && repeat_max == -1)        if (repeat_min == 0 && repeat_max == -1)
3725          *code++ = OP_CRSTAR + repeat_type;          *code++ = OP_CRSTAR + repeat_type;
# Line 3287  for (;; ptr++) Line 3755  for (;; ptr++)
3755          goto FAILED;          goto FAILED;
3756          }          }
3757    
       /* This is a paranoid check to stop integer overflow later on */  
   
       if (len > MAX_DUPLENGTH)  
         {  
         *errorcodeptr = ERR50;  
         goto FAILED;  
         }  
   
3758        /* 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
3759        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
3760        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 3383  for (;; ptr++) Line 3843  for (;; ptr++)
3843          if (repeat_min > 1)          if (repeat_min > 1)
3844            {            {
3845            /* 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
3846            just adjust the length as if we had. */            just adjust the length as if we had. Do some paranoid checks for
3847              potential integer overflow. */
3848    
3849            if (lengthptr != NULL)            if (lengthptr != NULL)
3850              *lengthptr += (repeat_min - 1)*length_prevgroup;              {
3851                int delta = (repeat_min - 1)*length_prevgroup;
3852                if ((double)(repeat_min - 1)*(double)length_prevgroup >
3853                                                                (double)INT_MAX ||
3854                    OFLOW_MAX - *lengthptr < delta)
3855                  {
3856                  *errorcodeptr = ERR20;
3857                  goto FAILED;
3858                  }
3859                *lengthptr += delta;
3860                }
3861    
3862            /* This is compiling for real */            /* This is compiling for real */
3863    
# Line 3424  for (;; ptr++) Line 3895  for (;; ptr++)
3895          /* 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
3896          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
3897          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
3898          add 2 + 2*LINKSIZE to allow for the nesting that occurs. */          add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some
3899            paranoid checks to avoid integer overflow. */
3900    
3901          if (lengthptr != NULL && repeat_max > 0)          if (lengthptr != NULL && repeat_max > 0)
3902            *lengthptr += repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -            {
3903              2 - 2*LINK_SIZE;  /* Last one doesn't nest */            int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) -
3904                          2 - 2*LINK_SIZE;   /* Last one doesn't nest */
3905              if ((double)repeat_max *
3906                    (double)(length_prevgroup + 1 + 2 + 2*LINK_SIZE)
3907                      > (double)INT_MAX ||
3908                  OFLOW_MAX - *lengthptr < delta)
3909                {
3910                *errorcodeptr = ERR20;
3911                goto FAILED;
3912                }
3913              *lengthptr += delta;
3914              }
3915    
3916          /* This is compiling for real */          /* This is compiling for real */
3917    
# Line 3580  for (;; ptr++) Line 4063  for (;; ptr++)
4063      /* ===================================================================*/      /* ===================================================================*/
4064      /* Start of nested parenthesized sub-expression, or comment or lookahead or      /* Start of nested parenthesized sub-expression, or comment or lookahead or
4065      lookbehind or option setting or condition or all the other extended      lookbehind or option setting or condition or all the other extended
4066      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. */  
4067    
4068      case '(':      case '(':
4069      newoptions = options;      newoptions = options;
4070      skipbytes = 0;      skipbytes = 0;
4071      bravalue = OP_CBRA;      bravalue = OP_CBRA;
4072      save_hwm = cd->hwm;      save_hwm = cd->hwm;
4073      reset_bracount = FALSE;      reset_bracount = FALSE;
4074    
4075      if (*(++ptr) == '?')      /* First deal with various "verbs" that can be introduced by '*'. */
4076    
4077        if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0)
4078          {
4079          int i, namelen;
4080          const char *vn = verbnames;
4081          const uschar *name = ++ptr;
4082          previous = NULL;
4083          while ((cd->ctypes[*++ptr] & ctype_letter) != 0);
4084          if (*ptr == ':')
4085            {
4086            *errorcodeptr = ERR59;   /* Not supported */
4087            goto FAILED;
4088            }
4089          if (*ptr != ')')
4090            {
4091            *errorcodeptr = ERR60;
4092            goto FAILED;
4093            }
4094          namelen = ptr - name;
4095          for (i = 0; i < verbcount; i++)
4096            {
4097            if (namelen == verbs[i].len &&
4098                strncmp((char *)name, vn, namelen) == 0)
4099              {
4100              *code = verbs[i].op;
4101              if (*code++ == OP_ACCEPT) cd->had_accept = TRUE;
4102              break;
4103              }
4104            vn += verbs[i].len + 1;
4105            }
4106          if (i < verbcount) continue;
4107          *errorcodeptr = ERR60;
4108          goto FAILED;
4109          }
4110    
4111        /* Deal with the extended parentheses; all are introduced by '?', and the
4112        appearance of any of them means that this is not a capturing group. */
4113    
4114        else if (*ptr == '?')
4115        {        {
4116        int i, set, unset, namelen;        int i, set, unset, namelen;
4117        int *optset;        int *optset;
# Line 3614  for (;; ptr++) Line 4134  for (;; ptr++)
4134          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4135          case '|':                 /* Reset capture count for each branch */          case '|':                 /* Reset capture count for each branch */
4136          reset_bracount = TRUE;          reset_bracount = TRUE;
4137          /* Fall through */          /* Fall through */
4138    
4139          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4140          case ':':                 /* Non-capturing bracket */          case ':':                 /* Non-capturing bracket */
# Line 3833  for (;; ptr++) Line 4353  for (;; ptr++)
4353    
4354          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4355          case '!':                 /* Negative lookahead */          case '!':                 /* Negative lookahead */
         bravalue = OP_ASSERT_NOT;  
4356          ptr++;          ptr++;
4357            if (*ptr == ')')          /* Optimize (?!) */
4358              {
4359              *code++ = OP_FAIL;
4360              previous = NULL;
4361              continue;
4362              }
4363            bravalue = OP_ASSERT_NOT;
4364          break;          break;
4365    
4366    
# Line 4192  for (;; ptr++) Line 4718  for (;; ptr++)
4718    
4719              case 'J':    /* Record that it changed in the external options */              case 'J':    /* Record that it changed in the external options */
4720              *optset |= PCRE_DUPNAMES;              *optset |= PCRE_DUPNAMES;
4721              cd->external_options |= PCRE_JCHANGED;              cd->external_flags |= PCRE_JCHANGED;
4722              break;              break;
4723    
4724              case 'i': *optset |= PCRE_CASELESS; break;              case 'i': *optset |= PCRE_CASELESS; break;
# Line 4315  for (;; ptr++) Line 4841  for (;; ptr++)
4841           errorcodeptr,                 /* Where to put an error message */           errorcodeptr,                 /* Where to put an error message */
4842           (bravalue == OP_ASSERTBACK ||           (bravalue == OP_ASSERTBACK ||
4843            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
4844           reset_bracount,               /* True if (?| group */           reset_bracount,               /* True if (?| group */
4845           skipbytes,                    /* Skip over bracket number */           skipbytes,                    /* Skip over bracket number */
4846           &subfirstbyte,                /* For possible first char */           &subfirstbyte,                /* For possible first char */
4847           &subreqbyte,                  /* For possible last char */           &subreqbyte,                  /* For possible last char */
# Line 4383  for (;; ptr++) Line 4909  for (;; ptr++)
4909        goto FAILED;        goto FAILED;
4910        }        }
4911    
4912      /* 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,
4913      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
4914      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
4915      a quantifier. */      duplicated by a quantifier.*/
4916    
4917      if (lengthptr != NULL)      if (lengthptr != NULL)
4918        {        {
4919          if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE)
4920            {
4921            *errorcodeptr = ERR20;
4922            goto FAILED;
4923            }
4924        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;
4925        code++;        *code++ = OP_BRA;
4926        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
4927        *code++ = OP_KET;        *code++ = OP_KET;
4928        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
4929          break;    /* No need to waste time with special character handling */
4930        }        }
4931    
4932      /* Otherwise update the main code pointer to the end of the group. */      /* Otherwise update the main code pointer to the end of the group. */
4933    
4934      else code = tempcode;      code = tempcode;
4935    
4936      /* For a DEFINE group, required and first character settings are not      /* For a DEFINE group, required and first character settings are not
4937      relevant. */      relevant. */
# Line 4603  for (;; ptr++) Line 5135  for (;; ptr++)
5135      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;
5136      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
5137    
5138        /* Remember if \r or \n were seen */
5139    
5140        if (mcbuffer[0] == '\r' || mcbuffer[0] == '\n')
5141          cd->external_flags |= PCRE_HASCRORLF;
5142    
5143      /* Set the first and required bytes appropriately. If no previous first      /* Set the first and required bytes appropriately. If no previous first
5144      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.
5145      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 4675  Arguments: Line 5212  Arguments:
5212    ptrptr         -> the address of the current pattern pointer    ptrptr         -> the address of the current pattern pointer
5213    errorcodeptr   -> pointer to error code variable    errorcodeptr   -> pointer to error code variable
5214    lookbehind     TRUE if this is a lookbehind assertion    lookbehind     TRUE if this is a lookbehind assertion
5215    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
5216    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
5217    firstbyteptr   place to put the first required character, or a negative number    firstbyteptr   place to put the first required character, or a negative number
5218    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 4689  Returns:         TRUE on success Line 5226  Returns:         TRUE on success
5226    
5227  static BOOL  static BOOL
5228  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,  compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,
5229    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
5230    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,    int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,
5231    int *lengthptr)    int *lengthptr)
5232  {  {
5233  const uschar *ptr = *ptrptr;  const uschar *ptr = *ptrptr;
# Line 4735  orig_bracount = max_bracount = cd->braco Line 5272  orig_bracount = max_bracount = cd->braco
5272  for (;;)  for (;;)
5273    {    {
5274    /* For a (?| group, reset the capturing bracket count so that each branch    /* For a (?| group, reset the capturing bracket count so that each branch
5275    uses the same numbers. */    uses the same numbers. */
5276    
5277    if (reset_bracount) cd->bracount = orig_bracount;    if (reset_bracount) cd->bracount = orig_bracount;
5278    
5279    /* Handle a change of ims options at the start of the branch */    /* Handle a change of ims options at the start of the branch */
5280    
5281    if ((options & PCRE_IMS) != oldims)    if ((options & PCRE_IMS) != oldims)
# Line 4767  for (;;) Line 5304  for (;;)
5304      *ptrptr = ptr;      *ptrptr = ptr;
5305      return FALSE;      return FALSE;
5306      }      }
5307    
5308    /* Keep the highest bracket count in case (?| was used and some branch    /* Keep the highest bracket count in case (?| was used and some branch
5309    has fewer than the rest. */    has fewer than the rest. */
5310    
5311    if (cd->bracount > max_bracount) max_bracount = cd->bracount;    if (cd->bracount > max_bracount) max_bracount = cd->bracount;
5312    
5313    /* In the real compile phase, there is some post-processing to be done. */    /* In the real compile phase, there is some post-processing to be done. */
# Line 4874  for (;;) Line 5411  for (;;)
5411        *code++ = oldims;        *code++ = oldims;
5412        length += 2;        length += 2;
5413        }        }
5414    
5415      /* Retain the highest bracket number, in case resetting was used. */      /* Retain the highest bracket number, in case resetting was used. */
5416    
5417      cd->bracount = max_bracount;      cd->bracount = max_bracount;
5418    
5419      /* Set values to pass back */      /* Set values to pass back */
# Line 4885  for (;;) Line 5422  for (;;)
5422      *ptrptr = ptr;      *ptrptr = ptr;
5423      *firstbyteptr = firstbyte;      *firstbyteptr = firstbyte;
5424      *reqbyteptr = reqbyte;      *reqbyteptr = reqbyte;
5425      if (lengthptr != NULL) *lengthptr += length;      if (lengthptr != NULL)
5426          {
5427          if (OFLOW_MAX - *lengthptr < length)
5428            {
5429            *errorcodeptr = ERR20;
5430            return FALSE;
5431            }
5432          *lengthptr += length;
5433          }
5434      return TRUE;      return TRUE;
5435      }      }
5436    
# Line 5194  real_pcre *re; Line 5739  real_pcre *re;
5739  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
5740  int firstbyte, reqbyte, newline;  int firstbyte, reqbyte, newline;
5741  int errorcode = 0;  int errorcode = 0;
5742    int skipatstart = 0;
5743  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
5744  BOOL utf8;  BOOL utf8;
5745  #endif  #endif
# Line 5272  cd->fcc = tables + fcc_offset; Line 5818  cd->fcc = tables + fcc_offset;
5818  cd->cbits = tables + cbits_offset;  cd->cbits = tables + cbits_offset;
5819  cd->ctypes = tables + ctypes_offset;  cd->ctypes = tables + ctypes_offset;
5820    
5821    /* Check for global one-time settings at the start of the pattern, and remember
5822    the offset for later. */
5823    
5824    while (ptr[skipatstart] == '(' && ptr[skipatstart+1] == '*')
5825      {
5826      int newnl = 0;
5827      int newbsr = 0;
5828    
5829      if (strncmp((char *)(ptr+skipatstart+2), "CR)", 3) == 0)
5830        { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }
5831      else if (strncmp((char *)(ptr+skipatstart+2), "LF)", 3)  == 0)
5832        { skipatstart += 5; newnl = PCRE_NEWLINE_LF; }
5833      else if (strncmp((char *)(ptr+skipatstart+2), "CRLF)", 5)  == 0)
5834        { skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; }
5835      else if (strncmp((char *)(ptr+skipatstart+2), "ANY)", 4) == 0)
5836        { skipatstart += 6; newnl = PCRE_NEWLINE_ANY; }
5837      else if (strncmp((char *)(ptr+skipatstart+2), "ANYCRLF)", 8)  == 0)
5838        { skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; }
5839    
5840      else if (strncmp((char *)(ptr+skipatstart+2), "BSR_ANYCRLF)", 12) == 0)
5841        { skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; }
5842      else if (strncmp((char *)(ptr+skipatstart+2), "BSR_UNICODE)", 12) == 0)
5843        { skipatstart += 14; newbsr = PCRE_BSR_UNICODE; }
5844    
5845      if (newnl != 0)
5846        options = (options & ~PCRE_NEWLINE_BITS) | newnl;
5847      else if (newbsr != 0)
5848        options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr;
5849      else break;
5850      }
5851    
5852    /* Check validity of \R options. */
5853    
5854    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
5855      {
5856      case 0:
5857      case PCRE_BSR_ANYCRLF:
5858      case PCRE_BSR_UNICODE:
5859      break;
5860      default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN;
5861      }
5862    
5863  /* Handle different types of newline. The three bits give seven cases. The  /* Handle different types of newline. The three bits give seven cases. The
5864  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
5865  "anycrlf". */  "anycrlf". */
5866    
5867  switch (options & (PCRE_NEWLINE_CRLF | PCRE_NEWLINE_ANY))  switch (options & PCRE_NEWLINE_BITS)
5868    {    {
5869    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Build-time default */
5870    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
5871    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = '\n'; break;
5872    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
# Line 5341  cd->hwm = cworkspace; Line 5929  cd->hwm = cworkspace;
5929  cd->start_pattern = (const uschar *)pattern;  cd->start_pattern = (const uschar *)pattern;
5930  cd->end_pattern = (const uschar *)(pattern + strlen(pattern));  cd->end_pattern = (const uschar *)(pattern + strlen(pattern));
5931  cd->req_varyopt = 0;  cd->req_varyopt = 0;
 cd->nopartial = FALSE;  
5932  cd->external_options = options;  cd->external_options = options;
5933    cd->external_flags = 0;
5934    
5935  /* 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
5936  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 5350  been put into the cd block so that they Line 5938  been put into the cd block so that they
5938  found within the regex right at the beginning. Bringing initial option settings  found within the regex right at the beginning. Bringing initial option settings
5939  outside can help speed up starting point checks. */  outside can help speed up starting point checks. */
5940    
5941    ptr += skipatstart;
5942  code = cworkspace;  code = cworkspace;
5943  *code = OP_BRA;  *code = OP_BRA;
5944  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,  (void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS,
5945    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,    &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd,
5946    &length);    &length);
5947  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
5948    
# Line 5380  if (re == NULL) Line 5969  if (re == NULL)
5969    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
5970    }    }
5971    
5972  /* 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
5973  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
5974  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
5975  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
5976    pointers. */
5977    
5978  re->magic_number = MAGIC_NUMBER;  re->magic_number = MAGIC_NUMBER;
5979  re->size = size;  re->size = size;
5980  re->options = cd->external_options;  re->options = cd->external_options;
5981    re->flags = cd->external_flags;
5982  re->dummy1 = 0;  re->dummy1 = 0;
5983  re->first_byte = 0;  re->first_byte = 0;
5984  re->req_byte = 0;  re->req_byte = 0;
# Line 5412  codestart = cd->name_table + re->name_en Line 6003  codestart = cd->name_table + re->name_en
6003  cd->start_code = codestart;  cd->start_code = codestart;
6004  cd->hwm = cworkspace;  cd->hwm = cworkspace;
6005  cd->req_varyopt = 0;  cd->req_varyopt = 0;
6006  cd->nopartial = FALSE;  cd->had_accept = FALSE;
6007    
6008  /* Set up a starting, non-extracting bracket, then compile the expression. On  /* Set up a starting, non-extracting bracket, then compile the expression. On
6009  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
6010  of the function here. */  of the function here. */
6011    
6012  ptr = (const uschar *)pattern;  ptr = (const uschar *)pattern + skipatstart;
6013  code = (uschar *)codestart;  code = (uschar *)codestart;
6014  *code = OP_BRA;  *code = OP_BRA;
6015  (void)compile_regex(re->options, re->options & PCRE_IMS, &code, &ptr,  (void)compile_regex(re->options, re->options & PCRE_IMS, &code, &ptr,
6016    &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL);    &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL);
6017  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
6018  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
6019    re->flags = cd->external_flags;
6020    
6021  if (cd->nopartial) re->options |= PCRE_NOPARTIAL;  if (cd->had_accept) reqbyte = -1;   /* Must disable after (*ACCEPT) */
6022    
6023  /* 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. */
6024    
# Line 5468  if (errorcode != 0) Line 6060  if (errorcode != 0)
6060    PCRE_EARLY_ERROR_RETURN:    PCRE_EARLY_ERROR_RETURN:
6061    *erroroffset = ptr - (const uschar *)pattern;    *erroroffset = ptr - (const uschar *)pattern;
6062    PCRE_EARLY_ERROR_RETURN2:    PCRE_EARLY_ERROR_RETURN2:
6063    *errorptr = error_texts[errorcode];    *errorptr = find_error_text(errorcode);
6064    if (errorcodeptr != NULL) *errorcodeptr = errorcode;    if (errorcodeptr != NULL) *errorcodeptr = errorcode;
6065    return NULL;    return NULL;
6066    }    }
# Line 5497  if ((re->options & PCRE_ANCHORED) == 0) Line 6089  if ((re->options & PCRE_ANCHORED) == 0)
6089        int ch = firstbyte & 255;        int ch = firstbyte & 255;
6090        re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&        re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&
6091           cd->fcc[ch] == ch)? ch : firstbyte;           cd->fcc[ch] == ch)? ch : firstbyte;
6092        re->options |= PCRE_FIRSTSET;        re->flags |= PCRE_FIRSTSET;
6093        }        }
6094      else if (is_startline(codestart, 0, cd->backref_map))      else if (is_startline(codestart, 0, cd->backref_map))
6095        re->options |= PCRE_STARTLINE;        re->flags |= PCRE_STARTLINE;
6096      }      }
6097    }    }
6098    
# Line 5514  if (reqbyte >= 0 && Line 6106  if (reqbyte >= 0 &&
6106    int ch = reqbyte & 255;    int ch = reqbyte & 255;
6107    re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&    re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&
6108      cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;      cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;
6109    re->options |= PCRE_REQCHSET;    re->flags |= PCRE_REQCHSET;
6110    }    }
6111    
6112  /* 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 5525  case when building a production library. Line 6117  case when building a production library.
6117  printf("Length = %d top_bracket = %d top_backref = %d\n",  printf("Length = %d top_bracket = %d top_backref = %d\n",
6118    length, re->top_bracket, re->top_backref);    length, re->top_bracket, re->top_backref);
6119    
6120  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 " : "");  
   }  
6121    
6122  if ((re->options & PCRE_FIRSTSET) != 0)  if ((re->flags & PCRE_FIRSTSET) != 0)
6123    {    {
6124    int ch = re->first_byte & 255;    int ch = re->first_byte & 255;
6125    const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?    const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?
# Line 5548  if ((re->options & PCRE_FIRSTSET) != 0) Line 6128  if ((re->options & PCRE_FIRSTSET) != 0)
6128      else printf("First char = \\x%02x%s\n", ch, caseless);      else printf("First char = \\x%02x%s\n", ch, caseless);
6129    }    }
6130    
6131  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6132    {    {
6133    int ch = re->req_byte & 255;    int ch = re->req_byte & 255;
6134    const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?    const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?
# Line 5565  was compiled can be seen. */ Line 6145  was compiled can be seen. */
6145  if (code - codestart > length)  if (code - codestart > length)
6146    {    {
6147    (pcre_free)(re);    (pcre_free)(re);
6148    *errorptr = error_texts[ERR23];    *errorptr = find_error_text(ERR23);
6149    *erroroffset = ptr - (uschar *)pattern;    *erroroffset = ptr - (uschar *)pattern;
6150    if (errorcodeptr != NULL) *errorcodeptr = ERR23;    if (errorcodeptr != NULL) *errorcodeptr = ERR23;
6151    return NULL;    return NULL;

Legend:
Removed from v.176  
changed lines
  Added in v.265

  ViewVC Help
Powered by ViewVC 1.1.5