/[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 206 by ph10, Fri Aug 3 14:53:04 2007 UTC revision 218 by ph10, Thu Aug 16 10:13:23 2007 UTC
# Line 140  static const short int escapes[] = { Line 140  static const short int escapes[] = {
140  #endif  #endif
141    
142    
143    /* Table of special "verbs" like (*PRUNE) */
144    
145    typedef struct verbitem {
146      const char *name;
147      int   len;
148      int   op;
149    } verbitem;
150    
151    static verbitem verbs[] = {
152      { "ACCEPT", 6, OP_ACCEPT },
153      { "COMMIT", 6, OP_COMMIT },
154      { "F",      1, OP_FAIL },
155      { "FAIL",   4, OP_FAIL },
156      { "PRUNE",  5, OP_PRUNE },
157      { "SKIP",   4, OP_SKIP  },
158      { "THEN",   4, OP_THEN  }
159    };
160    
161    static int verbcount = sizeof(verbs)/sizeof(verbitem);
162    
163    
164  /* Tables of names of POSIX character classes and their lengths. The list is  /* Tables of names of POSIX character classes and their lengths. The list is
165  terminated by a zero length entry. The first three must be alpha, lower, upper,  terminated by a zero length entry. The first three must be alpha, lower, upper,
166  as this is assumed for handling case independence. */  as this is assumed for handling case independence. */
# Line 258  static const char *error_texts[] = { Line 279  static const char *error_texts[] = {
279    "repeating a DEFINE group is not allowed",    "repeating a DEFINE group is not allowed",
280    "inconsistent NEWLINE options",    "inconsistent NEWLINE options",
281    "\\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",
282    "(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number"    "(?+ or (?- or (?(+ or (?(- must be followed by a non-zero number",
283      "(*VERB) with an argument is not supported",
284      /* 60 */
285      "(*VERB) not recognized",
286      "number is too big"
287  };  };
288    
289    
# Line 415  Arguments: Line 440  Arguments:
440    
441  Returns:         zero or positive => a data character  Returns:         zero or positive => a data character
442                   negative => a special escape sequence                   negative => a special escape sequence
443                   on error, errorptr is set                   on error, errorcodeptr is set
444  */  */
445    
446  static int  static int
# Line 499  else Line 524  else
524      c = 0;      c = 0;
525      while ((digitab[ptr[1]] & ctype_digit) != 0)      while ((digitab[ptr[1]] & ctype_digit) != 0)
526        c = c * 10 + *(++ptr) - '0';        c = c * 10 + *(++ptr) - '0';
527    
528        if (c < 0)
529          {
530          *errorcodeptr = ERR61;
531          break;
532          }
533    
534      if (c == 0 || (braced && *(++ptr) != '}'))      if (c == 0 || (braced && *(++ptr) != '}'))
535        {        {
536        *errorcodeptr = ERR57;        *errorcodeptr = ERR57;
537        return 0;        break;
538        }        }
539    
540      if (negated)      if (negated)
# Line 511  else Line 542  else
542        if (c > bracount)        if (c > bracount)
543          {          {
544          *errorcodeptr = ERR15;          *errorcodeptr = ERR15;
545          return 0;          break;
546          }          }
547        c = bracount - (c - 1);        c = bracount - (c - 1);
548        }        }
# Line 540  else Line 571  else
571        c -= '0';        c -= '0';
572        while ((digitab[ptr[1]] & ctype_digit) != 0)        while ((digitab[ptr[1]] & ctype_digit) != 0)
573          c = c * 10 + *(++ptr) - '0';          c = c * 10 + *(++ptr) - '0';
574          if (c < 0)
575            {
576            *errorcodeptr = ERR61;
577            break;
578            }
579        if (c < 10 || c <= bracount)        if (c < 10 || c <= bracount)
580          {          {
581          c = -(ESC_REF + c);          c = -(ESC_REF + c);
# Line 635  else Line 671  else
671      if (c == 0)      if (c == 0)
672        {        {
673        *errorcodeptr = ERR2;        *errorcodeptr = ERR2;
674        return 0;        break;
675        }        }
676    
677  #ifndef EBCDIC  /* ASCII coding */  #ifndef EBCDIC  /* ASCII coding */
# Line 914  for (; *ptr != 0; ptr++) Line 950  for (; *ptr != 0; ptr++)
950      {      {
951      while (*(++ptr) != ']')      while (*(++ptr) != ']')
952        {        {
953          if (*ptr == 0) return -1;
954        if (*ptr == '\\')        if (*ptr == '\\')
955          {          {
956          if (*(++ptr) == 0) return -1;          if (*(++ptr) == 0) return -1;
# Line 941  for (; *ptr != 0; ptr++) Line 978  for (; *ptr != 0; ptr++)
978    /* An opening parens must now be a real metacharacter */    /* An opening parens must now be a real metacharacter */
979    
980    if (*ptr != '(') continue;    if (*ptr != '(') continue;
981    if (ptr[1] != '?')    if (ptr[1] != '?' && ptr[1] != '*')
982      {      {
983      count++;      count++;
984      if (name == NULL && count == lorn) return count;      if (name == NULL && count == lorn) return count;
# Line 1069  for (;;) Line 1106  for (;;)
1106    {    {
1107    int d;    int d;
1108    register int op = *cc;    register int op = *cc;
   
1109    switch (op)    switch (op)
1110      {      {
1111      case OP_CBRA:      case OP_CBRA:
# Line 1158  for (;;) Line 1194  for (;;)
1194    
1195      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1196      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1197        if (cc[3] == OP_PROP || cc[3] == OP_NOTPROP) cc += 2;
1198      cc += 4;      cc += 4;
1199      break;      break;
1200    
# Line 1266  for (;;) Line 1303  for (;;)
1303      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1304      }      }
1305    
1306    /* 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
1307    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
1308    arrange to skip the extra bytes. */    two bytes of parameters. */
1309    
1310    else    else
1311      {      {
1312        switch(c)
1313          {
1314          case OP_TYPESTAR:
1315          case OP_TYPEMINSTAR:
1316          case OP_TYPEPLUS:
1317          case OP_TYPEMINPLUS:
1318          case OP_TYPEQUERY:
1319          case OP_TYPEMINQUERY:
1320          case OP_TYPEUPTO:
1321          case OP_TYPEMINUPTO:
1322          case OP_TYPEEXACT:
1323          case OP_TYPEPOSSTAR:
1324          case OP_TYPEPOSPLUS:
1325          case OP_TYPEPOSQUERY:
1326          case OP_TYPEPOSUPTO:
1327          if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1328          break;
1329          }
1330    
1331        /* Add in the fixed length from the table */
1332    
1333      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1334    
1335      /* In UTF-8 mode, opcodes that are followed by a character may be followed by
1336      a multi-byte character. The length in the table is a minimum, so we have to
1337      arrange to skip the extra bytes. */
1338    
1339  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1340      if (utf8) switch(c)      if (utf8) switch(c)
1341        {        {
# Line 1323  for (;;) Line 1386  for (;;)
1386    register int c = *code;    register int c = *code;
1387    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
1388    if (c == OP_RECURSE) return code;    if (c == OP_RECURSE) return code;
1389    
1390    /* XCLASS is used for classes that cannot be represented just by a bit    /* XCLASS is used for classes that cannot be represented just by a bit
1391    map. This includes negated single high-valued characters. The length in    map. This includes negated single high-valued characters. The length in
1392    the table is zero; the actual length is stored in the compiled code. */    the table is zero; the actual length is stored in the compiled code. */
1393    
1394    if (c == OP_XCLASS) code += GET(code, 1);    if (c == OP_XCLASS) code += GET(code, 1);
1395    
1396    /* 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
1397    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
1398    The length in the table is a minimum, so we have to arrange to skip the extra    two bytes of parameters. */
   bytes. */  
1399    
1400    else    else
1401      {      {
1402        switch(c)
1403          {
1404          case OP_TYPESTAR:
1405          case OP_TYPEMINSTAR:
1406          case OP_TYPEPLUS:
1407          case OP_TYPEMINPLUS:
1408          case OP_TYPEQUERY:
1409          case OP_TYPEMINQUERY:
1410          case OP_TYPEUPTO:
1411          case OP_TYPEMINUPTO:
1412          case OP_TYPEEXACT:
1413          case OP_TYPEPOSSTAR:
1414          case OP_TYPEPOSPLUS:
1415          case OP_TYPEPOSQUERY:
1416          case OP_TYPEPOSUPTO:
1417          if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2;
1418          break;
1419          }
1420    
1421        /* Add in the fixed length from the table */
1422    
1423      code += _pcre_OP_lengths[c];      code += _pcre_OP_lengths[c];
1424    
1425        /* In UTF-8 mode, opcodes that are followed by a character may be followed
1426        by a multi-byte character. The length in the table is a minimum, so we have
1427        to arrange to skip the extra bytes. */
1428    
1429  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1430      if (utf8) switch(c)      if (utf8) switch(c)
1431        {        {
# Line 1433  for (code = first_significant_code(code Line 1521  for (code = first_significant_code(code
1521    
1522    switch (c)    switch (c)
1523      {      {
1524      /* Check for quantifiers after a class */      /* Check for quantifiers after a class. XCLASS is used for classes that
1525        cannot be represented just by a bit map. This includes negated single
1526        high-valued characters. The length in _pcre_OP_lengths[] is zero; the
1527        actual length is stored in the compiled code, so we must update "code"
1528        here. */
1529    
1530  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1531      case OP_XCLASS:      case OP_XCLASS:
1532      ccode = code + GET(code, 1);      ccode = code += GET(code, 1);
1533      goto CHECK_CLASS_REPEAT;      goto CHECK_CLASS_REPEAT;
1534  #endif  #endif
1535    
# Line 2454  for (;; ptr++) Line 2546  for (;; ptr++)
2546        }        }
2547    
2548      /* If the first character is '^', set the negation flag and skip it. Also,      /* If the first character is '^', set the negation flag and skip it. Also,
2549      if the first few characters (either before or after ^) are \Q\E or \E we      if the first few characters (either before or after ^) are \Q\E or \E we
2550      skip them too. This makes for compatibility with Perl. */      skip them too. This makes for compatibility with Perl. */
2551    
2552      negate_class = FALSE;      negate_class = FALSE;
2553      for (;;)      for (;;)
2554        {        {
2555        c = *(++ptr);        c = *(++ptr);
2556        if (c == '\\')        if (c == '\\')
2557          {          {
2558          if (ptr[1] == 'E') ptr++;          if (ptr[1] == 'E') ptr++;
2559            else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3;            else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3;
2560              else break;              else break;
2561          }          }
2562        else if (!negate_class && c == '^')        else if (!negate_class && c == '^')
2563          negate_class = TRUE;          negate_class = TRUE;
2564        else break;        else break;
2565        }        }
2566    
2567      /* 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
2568      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 2630  for (;; ptr++) Line 2722  for (;; ptr++)
2722            else inescq = TRUE;            else inescq = TRUE;
2723            continue;            continue;
2724            }            }
2725            else if (-c == ESC_E) continue;  /* Ignore orphan \E */
2726    
2727          if (c < 0)          if (c < 0)
2728            {            {
# Line 3075  for (;; ptr++) Line 3168  for (;; ptr++)
3168        *errorcodeptr = ERR6;        *errorcodeptr = ERR6;
3169        goto FAILED;        goto FAILED;
3170        }        }
3171    
3172      /* 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
3173      less than 256. In non-UTF-8 mode we can always optimize. In UTF-8 mode, we      less than 256. In non-UTF-8 mode we can always optimize. In UTF-8 mode, we
3174      can optimize the negative case only if there were no characters >= 128      can optimize the negative case only if there were no characters >= 128
# Line 3860  for (;; ptr++) Line 3953  for (;; ptr++)
3953      /* ===================================================================*/      /* ===================================================================*/
3954      /* Start of nested parenthesized sub-expression, or comment or lookahead or      /* Start of nested parenthesized sub-expression, or comment or lookahead or
3955      lookbehind or option setting or condition or all the other extended      lookbehind or option setting or condition or all the other extended
3956      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. */  
3957    
3958      case '(':      case '(':
3959      newoptions = options;      newoptions = options;
# Line 3871  for (;; ptr++) Line 3962  for (;; ptr++)
3962      save_hwm = cd->hwm;      save_hwm = cd->hwm;
3963      reset_bracount = FALSE;      reset_bracount = FALSE;
3964    
3965      if (*(++ptr) == '?')      /* First deal with various "verbs" that can be introduced by '*'. */
3966    
3967        if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0)
3968          {
3969          int i, namelen;
3970          const uschar *name = ++ptr;
3971          previous = NULL;
3972          while ((cd->ctypes[*++ptr] & ctype_letter) != 0);
3973          if (*ptr == ':')
3974            {
3975            *errorcodeptr = ERR59;   /* Not supported */
3976            goto FAILED;
3977            }
3978          if (*ptr != ')')
3979            {
3980            *errorcodeptr = ERR60;
3981            goto FAILED;
3982            }
3983          namelen = ptr - name;
3984          for (i = 0; i < verbcount; i++)
3985            {
3986            if (namelen == verbs[i].len &&
3987                strncmp((char *)name, verbs[i].name, namelen) == 0)
3988              {
3989              *code = verbs[i].op;
3990              if (*code++ == OP_ACCEPT) cd->had_accept = TRUE;
3991              break;
3992              }
3993            }
3994          if (i < verbcount) continue;
3995          *errorcodeptr = ERR60;
3996          goto FAILED;
3997          }
3998    
3999        /* Deal with the extended parentheses; all are introduced by '?', and the
4000        appearance of any of them means that this is not a capturing group. */
4001    
4002        else if (*ptr == '?')
4003        {        {
4004        int i, set, unset, namelen;        int i, set, unset, namelen;
4005        int *optset;        int *optset;
# Line 4113  for (;; ptr++) Line 4241  for (;; ptr++)
4241    
4242          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
4243          case '!':                 /* Negative lookahead */          case '!':                 /* Negative lookahead */
         bravalue = OP_ASSERT_NOT;  
4244          ptr++;          ptr++;
4245            if (*ptr == ')')          /* Optimize (?!) */
4246              {
4247              *code++ = OP_FAIL;
4248              previous = NULL;
4249              continue;
4250              }
4251            bravalue = OP_ASSERT_NOT;
4252          break;          break;
4253    
4254    
# Line 4663  for (;; ptr++) Line 4797  for (;; ptr++)
4797        goto FAILED;        goto FAILED;
4798        }        }
4799    
4800      /* 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,
4801      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
4802      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
4803      a quantifier. */      duplicated by a quantifier.*/
4804    
4805      if (lengthptr != NULL)      if (lengthptr != NULL)
4806        {        {
# Line 4676  for (;; ptr++) Line 4810  for (;; ptr++)
4810          goto FAILED;          goto FAILED;
4811          }          }
4812        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;
4813        code++;        *code++ = OP_BRA;
4814        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
4815        *code++ = OP_KET;        *code++ = OP_KET;
4816        PUTINC(code, 0, 1 + LINK_SIZE);        PUTINC(code, 0, 1 + LINK_SIZE);
4817          break;    /* No need to waste time with special character handling */
4818        }        }
4819    
4820      /* Otherwise update the main code pointer to the end of the group. */      /* Otherwise update the main code pointer to the end of the group. */
4821    
4822      else code = tempcode;      code = tempcode;
4823    
4824      /* For a DEFINE group, required and first character settings are not      /* For a DEFINE group, required and first character settings are not
4825      relevant. */      relevant. */
# Line 5706  cd->start_code = codestart; Line 5841  cd->start_code = codestart;
5841  cd->hwm = cworkspace;  cd->hwm = cworkspace;
5842  cd->req_varyopt = 0;  cd->req_varyopt = 0;
5843  cd->nopartial = FALSE;  cd->nopartial = FALSE;
5844    cd->had_accept = FALSE;
5845    
5846  /* Set up a starting, non-extracting bracket, then compile the expression. On  /* Set up a starting, non-extracting bracket, then compile the expression. On
5847  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
# Line 5720  re->top_bracket = cd->bracount; Line 5856  re->top_bracket = cd->bracount;
5856  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
5857    
5858  if (cd->nopartial) re->options |= PCRE_NOPARTIAL;  if (cd->nopartial) re->options |= PCRE_NOPARTIAL;
5859    if (cd->had_accept) reqbyte = -1;   /* Must disable after (*ACCEPT) */
5860    
5861  /* 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. */
5862    
Legend:
Removed from v.206  
changed lines
  Added in v.218
webmaster@exim.org
 ViewVC Help
Powered by ViewVC 1.1.12